Datasets:

---

maxiw

/

hf-posts

like 18

[ { "type": "text", "value": "We noticed that all the open-source models and datasets from ", "raw": "We noticed that all the open-source models and datasets from ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/WizardLM", "href": "https://huggingface.co/WizardLM", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " in their personal Hugging Face account & in the Microsoft Hugging Face organization (", "raw": " in their personal Hugging Face account & in the Microsoft Hugging Face organization (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/microsoft", "href": "https://huggingface.co/microsoft", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ") have been made private by the author, which will lead some demos to fail (these models were collectively downloaded over a hundred thousand times a month). ", "raw": ") have been made private by the author, which will lead some demos to fail (these models were collectively downloaded over a hundred thousand times a month). ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This is the explanation that ", "raw": "This is the explanation that ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@WizardLM", "href": null, "resource": null, "url": null, "code": null, "user": "WizardLM", "label": null, "lang": null }, { "type": "text", "value": " communicated a few hours ago: ", "raw": " communicated a few hours ago: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/posts/WizardLM/329547800484476#661e0d17bca1a6038b60503e", "href": "https://huggingface.co/posts/WizardLM/329547800484476#661e0d17bca1a6038b60503e", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We apologize for the inconvenience & are trying to get in touch with the author & Microsoft in order to try to find a good resolution for community members. Let us know if you have any questions!", "raw": "We apologize for the inconvenience & are trying to get in touch with the author & Microsoft in order to try to find a good resolution for community members. Let us know if you have any questions!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/627b9f3f4d0858f0034efbb9/2Qnattrzv6qvqiZVVfV5x.png", "fullname": "WizardLM", "name": "WizardLM", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3191 } ]

[ { "reaction": "👀", "users": [ "osanseviero", "fdaudens", "ijohn07", "pierrci", "BrigitteTousi", "lazarustda", "samusenps", "clefourrier", "lunarflu", "Diz187", "guilherme-am", "m-ric", "acamilogg88" ], "count": 13 }, { "reaction": "😔", "users": [ "osanseviero", "pierrci", "Mehrdadgpt", "lunarflu", "fynnkroeger", "m-ric" ], "count": 6 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/654db98e1b922725af7ec5ec/gChOe5aQjFnQxnksbD2AY.jpeg", "fullname": "Diz", "name": "Diz187", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "Learn Your Reference Model for Real Good Alignment", "raw": "Learn Your Reference Model for Real Good Alignment", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.09656", "href": null, "resource": { "type": "paper", "id": "2404.09656", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.09656", "code": null, "user": null, "label": "Learn Your Reference Model for Real Good Alignment (2404.09656)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The complexity of the alignment problem stems from the fact that existing methods are unstable. Researchers continuously invent various tricks to address this shortcoming. For instance, in the fundamental Reinforcement Learning From Human Feedback (RLHF) technique of Language Model alignment, in addition to reward maximization, the Kullback-Leibler divergence between the trainable policy and the SFT policy is minimized. This addition prevents the model from being overfitted to the Reward Model (RM) and generating texts that are out-of-domain for the RM. The Direct Preference Optimization (DPO) method reformulates the optimization task of RLHF and eliminates the Reward Model while tacitly maintaining the requirement for the policy to be close to the SFT policy. In our paper, we argue that this implicit limitation in the DPO method leads to sub-optimal results. We propose a new method called Trust Region DPO (TR-DPO), which updates the reference policy during training. With such a straightforward update, we demonstrate the effectiveness of TR-DPO against DPO on the Anthropic HH and TLDR datasets. We show that TR-DPO outperforms DPO by up to 19%, measured by automatic evaluation with GPT-4. The new alignment approach that we propose allows us to improve the quality of models across several parameters at once, such as coherence, correctness, level of detail, helpfulness, and harmlessness.", "raw": "The complexity of the alignment problem stems from the fact that existing methods are unstable. Researchers continuously invent various tricks to address this shortcoming. For instance, in the fundamental Reinforcement Learning From Human Feedback (RLHF) technique of Language Model alignment, in addition to reward maximization, the Kullback-Leibler divergence between the trainable policy and the SFT policy is minimized. This addition prevents the model from being overfitted to the Reward Model (RM) and generating texts that are out-of-domain for the RM. The Direct Preference Optimization (DPO) method reformulates the optimization task of RLHF and eliminates the Reward Model while tacitly maintaining the requirement for the policy to be close to the SFT policy. In our paper, we argue that this implicit limitation in the DPO method leads to sub-optimal results. We propose a new method called Trust Region DPO (TR-DPO), which updates the reference policy during training. With such a straightforward update, we demonstrate the effectiveness of TR-DPO against DPO on the Anthropic HH and TLDR datasets. We show that TR-DPO outperforms DPO by up to 19%, measured by automatic evaluation with GPT-4. The new alignment approach that we propose allows us to improve the quality of models across several parameters at once, such as coherence, correctness, level of detail, helpfulness, and harmlessness.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60f1abe7544c2adfd699860c/W70kKgLAmdyZtynGwSy9S.png" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "IsmaelMousa", "lunarflu" ], "count": 3 } ]

[ { "avatarUrl": "/avatars/b44641ffaea8336110f5d443228559bb.svg", "fullname": "Jeff Nyzio", "name": "TheOneTrueNiz", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false } ]

[ { "type": "text", "value": "🔍 Today's pick in Interpretability & Analysis of LMs: ", "raw": "🔍 Today's pick in Interpretability & Analysis of LMs: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.07004", "href": null, "resource": { "type": "paper", "id": "2404.07004", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.07004", "code": null, "user": null, "label": "LM Transparency Tool: Interactive Tool for Analyzing Transformer\n Language Models (2404.07004)", "lang": null }, { "type": "text", "value": " by ", "raw": " by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@igortufanov", "href": null, "resource": null, "url": null, "code": null, "user": "igortufanov", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@mahnerak", "href": null, "resource": null, "url": null, "code": null, "user": "mahnerak", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@javifer", "href": null, "resource": null, "url": null, "code": null, "user": "javifer", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@lena-voita", "href": null, "resource": null, "url": null, "code": null, "user": "lena-voita", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The LLM transparency toolkit is an open source toolkit and visual interface to efficiently identify component circuits in LMs responsible for their predictions, using the Information Flow Routes approach (", "raw": "The LLM transparency toolkit is an open source toolkit and visual interface to efficiently identify component circuits in LMs responsible for their predictions, using the Information Flow Routes approach (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2403.00824", "href": null, "resource": { "type": "paper", "id": "2403.00824", "discussionNum": null }, "url": "https://huggingface.co/papers/2403.00824", "code": null, "user": null, "label": "Information Flow Routes: Automatically Interpreting Language Models at\n Scale (2403.00824)", "lang": null }, { "type": "text", "value": ").", "raw": ").", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The tool enables fine-grained customization, highlighting the importance of individual FFN neurons and attention heads. Moreover, vocabulary projections computed using the logit lens approach are provided to examine intermediate predictions of the residual stream, and tokens promoted by specific component updates.", "raw": "The tool enables fine-grained customization, highlighting the importance of individual FFN neurons and attention heads. Moreover, vocabulary projections computed using the logit lens approach are provided to examine intermediate predictions of the residual stream, and tokens promoted by specific component updates.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💻 Code: ", "raw": "💻 Code: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/facebookresearch/llm-transparency-tool", "href": "https://github.com/facebookresearch/llm-transparency-tool", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Demo: ", "raw": "🚀 Demo: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/facebook/llm-transparency-tool-demo", "href": null, "resource": { "type": "space", "id": "facebook/llm-transparency-tool-demo", "discussionNum": null }, "url": "https://huggingface.co/spaces/facebook/llm-transparency-tool-demo", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 All daily picks: ", "raw": "🔍 All daily picks: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-ofc-lms-65ae3339949c5675d25de2f9", "href": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-ofc-lms-65ae3339949c5675d25de2f9", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/a73ZUcuqFt5HrqE46g8SN.png" } ]

[ { "avatarUrl": "/avatars/bcc12268567ee767306d328458603c75.svg", "fullname": "Igor Tufanov", "name": "igortufanov", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null }, { "avatarUrl": "/avatars/bcc94a31fab7486ca9d018245a289fb0.svg", "fullname": "Javier Ferrando", "name": "javifer", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 5 }, { "avatarUrl": "/avatars/d5e02ca2e88bb9b17f92703927a34df4.svg", "fullname": "Elena Voita", "name": "lena-voita", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/63a2fb1cdca41424cb8956bd/9BDkkHXPifN-_czXBAecF.png", "fullname": "Karen Hambardzumyan", "name": "mahnerak", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 11 } ]

[ { "reaction": "🔥", "users": [ "radames", "osanseviero", "lunarflu", "Yunorga" ], "count": 4 }, { "reaction": "➕", "users": [ "samusenps", "lunarflu" ], "count": 2 } ]

[]

[ { "type": "text", "value": "FiftyOne Datasets <> Hugging Face Hub Integration!", "raw": "FiftyOne Datasets <> Hugging Face Hub Integration!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "As of yesterday's release of FiftyOne ", "raw": "As of yesterday's release of FiftyOne ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "inline_code", "value": null, "raw": "`0.23.8`", "href": null, "resource": null, "url": null, "code": "0.23.8", "user": null, "label": null, "lang": null }, { "type": "text", "value": ", the FiftyOne open source library for dataset curation and visualization is now integrated with the Hugging Face Hub!", "raw": ", the FiftyOne open source library for dataset curation and visualization is now integrated with the Hugging Face Hub!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "You can now load Parquet datasets from the hub and have them converted directly into FiftyOne datasets. To load MNIST, for example:", "raw": "You can now load Parquet datasets from the hub and have them converted directly into FiftyOne datasets. To load MNIST, for example:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "code_fence", "value": null, "raw": "```bash\npip install -U fiftyone\n```", "href": null, "resource": null, "url": null, "code": "pip install -U fiftyone", "user": null, "label": null, "lang": "bash" }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "code_fence", "value": null, "raw": "```py\nimport fiftyone as fo\nimport fiftyone.utils.huggingface as fouh\n\ndataset = fouh.load_from_hub(\n \"mnist\",\n format=\"ParquetFilesDataset\",\n classification_fields=\"label\",\n)\nsession = fo.launch_app(dataset)\n```", "href": null, "resource": null, "url": null, "code": "import fiftyone as fo\nimport fiftyone.utils.huggingface as fouh\n\ndataset = fouh.load_from_hub(\n \"mnist\",\n format=\"ParquetFilesDataset\",\n classification_fields=\"label\",\n)\nsession = fo.launch_app(dataset)", "user": null, "label": null, "lang": "py" }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "You can also load FiftyOne datasets directly from the hub. Here's how you load the first 1000 samples from the VisDrone dataset:", "raw": "You can also load FiftyOne datasets directly from the hub. Here's how you load the first 1000 samples from the VisDrone dataset:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "code_fence", "value": null, "raw": "```py\nimport fiftyone as fo\nimport fiftyone.utils.huggingface as fouh\n\ndataset = fouh.load_from_hub(\"jamarks/VisDrone2019-DET\", max_samples=1000)\n\n# Launch the App\nsession = fo.launch_app(dataset)\n```", "href": null, "resource": null, "url": null, "code": "import fiftyone as fo\nimport fiftyone.utils.huggingface as fouh\n\ndataset = fouh.load_from_hub(\"jamarks/VisDrone2019-DET\", max_samples=1000)\n\n# Launch the App\nsession = fo.launch_app(dataset)", "user": null, "label": null, "lang": "py" }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "And tying it all together, you can push your FiftyOne datasets directly to the hub:", "raw": "And tying it all together, you can push your FiftyOne datasets directly to the hub:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "code_fence", "value": null, "raw": "```py\nimport fiftyone.zoo as foz\nimport fiftyone.utils.huggingface as fouh\n\ndataset = foz.load_zoo_dataset(\"quickstart\")\nfouh.push_to_hub(dataset, \"my-dataset\")\n```", "href": null, "resource": null, "url": null, "code": "import fiftyone.zoo as foz\nimport fiftyone.utils.huggingface as fouh\n\ndataset = foz.load_zoo_dataset(\"quickstart\")\nfouh.push_to_hub(dataset, \"my-dataset\")", "user": null, "label": null, "lang": "py" }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Major thanks to ", "raw": "Major thanks to ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@tomaarsen", "href": null, "resource": null, "url": null, "code": null, "user": "tomaarsen", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@davanstrien", "href": null, "resource": null, "url": null, "code": null, "user": "davanstrien", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@severo", "href": null, "resource": null, "url": null, "code": null, "user": "severo", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@osanseviero", "href": null, "resource": null, "url": null, "code": null, "user": "osanseviero", "label": null, "lang": null }, { "type": "text", "value": " and ", "raw": " and ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@julien-c", "href": null, "resource": null, "url": null, "code": null, "user": "julien-c", "label": null, "lang": null }, { "type": "text", "value": " for helping to make this happen!!!", "raw": " for helping to make this happen!!!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Full documentation and details here: ", "raw": "Full documentation and details here: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://docs.voxel51.com/integrations/huggingface.html#huggingface-hub", "href": "https://docs.voxel51.com/integrations/huggingface.html#huggingface-hub", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/63127e2495407887cb79c5ea/TrF_YBYHSvjVM7lCwL3h2.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1627505688463-60107b385ac3e86b3ea4fc34.jpeg", "fullname": "Daniel van Strien", "name": "davanstrien", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 410 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/5dd96eb166059660ed1ee413/NQtzmrDdbG0H8qkZvRyGk.jpeg", "fullname": "Julien Chaumond", "name": "julien-c", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 1580 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6032802e1f993496bc14d9e3/w6hr-DEQot4VVkoyRIBiy.png", "fullname": "Omar Sanseviero", "name": "osanseviero", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2868 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/60a76b174e24361791fe822d/inEvYwrd4z0xvRQN3ikdE.jpeg", "fullname": "Sylvain Lesage", "name": "severo", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 129 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6317233cc92fd6fee317e030/cJHSvvimr1kqgQfHOjO5n.png", "fullname": "Tom Aarsen", "name": "tomaarsen", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1060 } ]

[ { "reaction": "🔥", "users": [ "severo", "brianmoore", "davanstrien", "ehofesmann", "jguerrero", "harpreetsahota", "victor", "clem", "merve", "osanseviero", "lunarflu", "julien-c", "dgural", "not-lain", "jamarks", "shamy777", "Maralk" ], "count": 17 }, { "reaction": "🤗", "users": [ "brianmoore", "davanstrien", "ehofesmann", "jguerrero", "harpreetsahota", "jamarks", "victor", "clem", "osanseviero", "lunarflu", "julien-c", "dgural", "not-lain" ], "count": 13 }, { "reaction": "🚀", "users": [ "brianmoore", "davanstrien", "ehofesmann", "jguerrero", "harpreetsahota", "victor", "lunarflu", "julien-c", "dgural", "not-lain", "jamarks" ], "count": 11 }, { "reaction": "🤯", "users": [ "ehofesmann", "jguerrero", "harpreetsahota", "lunarflu", "julien-c", "dgural", "not-lain", "jamarks" ], "count": 8 }, { "reaction": "❤️", "users": [ "samusenps", "lunarflu", "dgural", "not-lain", "jamarks" ], "count": 5 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6317233cc92fd6fee317e030/cJHSvvimr1kqgQfHOjO5n.png", "fullname": "Tom Aarsen", "name": "tomaarsen", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1060, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/630904f2c038bf42d56d9d11/S8mYgFpPSHYOiifBnNfwG.jpeg", "fullname": "Harpreet Sahota", "name": "harpreetsahota", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 52, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/5dd96eb166059660ed1ee413/NQtzmrDdbG0H8qkZvRyGk.jpeg", "fullname": "Julien Chaumond", "name": "julien-c", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 1580, "isFollowing": false } ]

[ { "type": "text", "value": "Demo for IDEFICS-8B demo is out! ", "raw": "Demo for IDEFICS-8B demo is out! ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/HuggingFaceM4/idefics-8b", "href": null, "resource": { "type": "space", "id": "HuggingFaceM4/idefics-8b", "discussionNum": null }, "url": "https://huggingface.co/spaces/HuggingFaceM4/idefics-8b", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This checkpoint is not optimized to chat, but rather works very well for various tasks, incl visual question answering and document tasks 💬📑", "raw": "This checkpoint is not optimized to chat, but rather works very well for various tasks, incl visual question answering and document tasks 💬📑", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Chatty one is coming soon!", "raw": "Chatty one is coming soon!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6141a88b3a0ec78603c9e784/3qPbJyCY_UonpS0ncxGDU.png" } ]

[]

[ { "reaction": "🔥", "users": [ "radames", "kramp", "clem", "Redox3275", "osanseviero", "CristianJD", "lunarflu", "VictorSanh" ], "count": 8 }, { "reaction": "❤️", "users": [ "lunarflu" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Super RAGs in Mistral 8x7B-v1 -", "raw": "Super RAGs in Mistral 8x7B-v1 -", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The recent paper on arXiv introduces Super Retrieval-Augmented Generation (Super RAGs), a groundbreaking approach to improve Large Language Models (LLMs) by integrating external knowledge sources. This integration into the Mistral 8x7B v1 LLM has shown notable improvements in accuracy, speed, and user satisfaction.", "raw": "The recent paper on arXiv introduces Super Retrieval-Augmented Generation (Super RAGs), a groundbreaking approach to improve Large Language Models (LLMs) by integrating external knowledge sources. This integration into the Mistral 8x7B v1 LLM has shown notable improvements in accuracy, speed, and user satisfaction.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "What are your thoughts on the potential of Super RAGs to transform the future of AI and LLMs?", "raw": "What are your thoughts on the potential of Super RAGs to transform the future of AI and LLMs?", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.08940", "href": null, "resource": { "type": "paper", "id": "2404.08940", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.08940", "code": null, "user": null, "label": "Introducing Super RAGs in Mistral 8x7B-v1 (2404.08940)", "lang": null } ]

[]

[]

[ { "reaction": "🤗", "users": [ "victor", "clem", "merve", "lunarflu", "not-lain", "Jason-Lu" ], "count": 6 }, { "reaction": "❤️", "users": [ "raghav2543", "clem", "merve", "lunarflu", "not-lain" ], "count": 5 }, { "reaction": "🔥", "users": [ "raghav2543", "clem", "lunarflu", "not-lain", "Jason-Lu" ], "count": 5 }, { "reaction": "🧠", "users": [ "raghav2543", "clem", "lunarflu", "not-lain" ], "count": 4 } ]

[]

[ { "type": "text", "value": "Just had the former chief public health officer of the Netherlands🇳🇱 review a huggingface AI-doctor I made via a simple orpo-zephyr-8x22b-GPT and they think it's really good. ", "raw": "Just had the former chief public health officer of the Netherlands🇳🇱 review a huggingface AI-doctor I made via a simple orpo-zephyr-8x22b-GPT and they think it's really good. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://hf.co/chat/assistant/661d77310e3aea9ae571e43c", "href": "https://hf.co/chat/assistant/661d77310e3aea9ae571e43c", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🔥", "users": [ "pierrci", "pankaj-munde", "mbilalkhaan", "kramp", "jpodivin", "m-ric", "clem", "mdadnan", "lunarflu", "Violette", "BrigitteTousi", "jbilcke-hf", "dillfrescott", "kellykellz", "akyarborough", "andrewgleave", "surajp", "bbooth", "dillen-hub" ], "count": 19 } ]

[ { "avatarUrl": "/avatars/26b513cf205a182ca5d32d664402377e.svg", "fullname": "Jean Francois Gutierrez", "name": "FranZuzz", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/d5a5b64bfff4408a653eba9bc417b43a.svg", "fullname": "Adnan Afzal", "name": "mdadnan", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6379683a81c1783a4a2ddba8/Ru0X82AlVgkJ4TOv28TyH.jpeg", "fullname": "nisten", "name": "nisten", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 174, "isFollowing": false } ]

[ { "type": "text", "value": "We release Idefics2-8B, a foundation vision language model with SOTA results for its size on many benchmarks.", "raw": "We release Idefics2-8B, a foundation vision language model with SOTA results for its size on many benchmarks.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "For Idefics2, we adopted a simple architecture:", "raw": "For Idefics2, we adopted a simple architecture:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "-Images are fed to a vision encoder, then to a modality projection to match the input dimension of the LLM, and finally to a perceiver resampler for efficient pooling.", "raw": "-Images are fed to a vision encoder, then to a modality projection to match the input dimension of the LLM, and finally to a perceiver resampler for efficient pooling.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "-Interleaved image-text data are then passed to the LLM.", "raw": "-Interleaved image-text data are then passed to the LLM.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "During the pre-training:", "raw": "During the pre-training:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "-The modality projection and perceiver resampler weights are newly initialized.", "raw": "-The modality projection and perceiver resampler weights are newly initialized.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "-We start with pre-trained models for the vision encoder and the LLM, and continue the training with LoRA.", "raw": "-We start with pre-trained models for the vision encoder and the LLM, and continue the training with LoRA.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "-In total, we see 1.5T images!", "raw": "-In total, we see 1.5T images!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We pre-train on 3 types of data, all publicly available:", "raw": "We pre-train on 3 types of data, all publicly available:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "-Interleaved image-text documents: our dataset OBELICS ", "raw": "-Interleaved image-text documents: our dataset OBELICS ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/HuggingFaceM4/OBELICS", "href": null, "resource": { "type": "dataset", "id": "HuggingFaceM4/OBELICS", "discussionNum": null }, "url": "https://huggingface.co/datasets/HuggingFaceM4/OBELICS", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "-Image caption pairs: only synthetic captions!", "raw": "-Image caption pairs: only synthetic captions!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "-PDF documents: IDL and PDFA", "raw": "-PDF documents: IDL and PDFA", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We kept the aspect ratio of the images with the Patch n' Pack strategy, with a resolution of up to 980x980.", "raw": "We kept the aspect ratio of the images with the Patch n' Pack strategy, with a resolution of up to 980x980.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "At inference, it's also more efficient for lower-resolution images.", "raw": "At inference, it's also more efficient for lower-resolution images.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "For the SFT, we build The Cauldron, a collection of 50 high-quality datasets in the user/assistant format.", "raw": "For the SFT, we build The Cauldron, a collection of 50 high-quality datasets in the user/assistant format.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "It is a ready-to-use dataset for the fine-tuning of any VLM.", "raw": "It is a ready-to-use dataset for the fine-tuning of any VLM.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/HuggingFaceM4/the_cauldron", "href": null, "resource": { "type": "dataset", "id": "HuggingFaceM4/the_cauldron", "discussionNum": null }, "url": "https://huggingface.co/datasets/HuggingFaceM4/the_cauldron", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Most current models, like LLaVA-NeXT, encode images with an excessive number of tokens, like 2880.", "raw": "Most current models, like LLaVA-NeXT, encode images with an excessive number of tokens, like 2880.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Instead, we put a focus on being efficient at inference by training on a mix of images encoded with 64 tokens, and 320 tokens.", "raw": "Instead, we put a focus on being efficient at inference by training on a mix of images encoded with 64 tokens, and 320 tokens.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The result is that we perform favorably compared to the best models in our size class, while being efficient at inference.", "raw": "The result is that we perform favorably compared to the best models in our size class, while being efficient at inference.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6177322d37f32ecb1e2d4cdf/O6uV2_qM0FpnaG669pwhF.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6177322d37f32ecb1e2d4cdf/dpROrEyRmIrn7i4gmDL8y.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6177322d37f32ecb1e2d4cdf/qK891rSiF_P8u3BoQ245O.png" } ]

[]

[ { "reaction": "🤯", "users": [ "VictorSanh", "victor", "danaaubakirova", "sbarman25", "m-ric", "clem", "lunarflu", "Violette", "fdaudens", "sbrandeis" ], "count": 10 }, { "reaction": "🚀", "users": [ "m-ric", "clem", "samusenps", "Sylvestre", "Molbap", "NePe", "lunarflu", "fynnkroeger", "Violette" ], "count": 9 }, { "reaction": "🧠", "users": [ "VictorSanh", "m-ric", "clem", "lunarflu", "Violette" ], "count": 5 }, { "reaction": "🤝", "users": [ "WenchuanYang", "clem", "samusenps", "lunarflu", "Violette" ], "count": 5 }, { "reaction": "👍", "users": [ "Violette" ], "count": 1 } ]

[]

[ { "type": "text", "value": "New open multimodal model in town: Idefics2! ", "raw": "New open multimodal model in town: Idefics2! ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💪 Strong 8B-parameters model: often on par with open 30B counterparts. ", "raw": "💪 Strong 8B-parameters model: often on par with open 30B counterparts. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔓Open license: Apache 2.0. ", "raw": "🔓Open license: Apache 2.0. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Strong improvement over Idefics1: +12 points on VQAv2, +30 points on TextVQA while having 10x fewer parameters. ", "raw": "🚀 Strong improvement over Idefics1: +12 points on VQAv2, +30 points on TextVQA while having 10x fewer parameters. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📚 Better data: boosting OCR capabilities with 6TB of documents to transcribe, and improving QA capabilities on charts/figures/diagrams. ", "raw": "📚 Better data: boosting OCR capabilities with 6TB of documents to transcribe, and improving QA capabilities on charts/figures/diagrams. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🕵️‍♀️ Transparent training data: inspect and build upon all the data (10s of TB of data) we trained on. ", "raw": "🕵️‍♀️ Transparent training data: inspect and build upon all the data (10s of TB of data) we trained on. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔲 More natural image processing: Incorporating strategies to treat images in their native resolution and native aspect ratio. ", "raw": "🔲 More natural image processing: Incorporating strategies to treat images in their native resolution and native aspect ratio. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📸 High-resolution images: image resolutions up to 980 x 980 and integrating strategies that allow to trade computational efficiency for performance. ", "raw": "📸 High-resolution images: image resolutions up to 980 x 980 and integrating strategies that allow to trade computational efficiency for performance. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "😎 2 checkpoints: Releasing both base checkpoint and instruction fine-tuned checkpoint. Chat version to come.", "raw": "😎 2 checkpoints: Releasing both base checkpoint and instruction fine-tuned checkpoint. Chat version to come.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Ressources: ", "raw": "Ressources: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/HuggingFaceM4/idefics2-661d1971b7c50831dd3ce0fe", "href": null, "resource": { "type": "collection", "id": "HuggingFaceM4/idefics2-661d1971b7c50831dd3ce0fe", "discussionNum": null }, "url": "https://huggingface.co/collections/HuggingFaceM4/idefics2-661d1971b7c50831dd3ce0fe", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Blogpost: ", "raw": "Blogpost: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/idefics2", "href": "https://huggingface.co/blog/idefics2", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🚀", "users": [ "VictorSanh", "rwightman", "monsoon-nlp", "oceansweep", "pierrci", "ajibawa-2023", "osanseviero", "Ayushnangia", "andrewrreed", "m-ric", "clem", "Molbap", "lunarflu", "fynnkroeger", "marcsun13", "raincandy-u", "jlzhou" ], "count": 17 }, { "reaction": "🔥", "users": [ "VictorSanh", "clem", "oceansweep", "pierrci", "samusenps", "osanseviero", "andrewrreed", "m-ric", "Molbap", "lunarflu" ], "count": 10 }, { "reaction": "❤️", "users": [ "VictorSanh", "clem", "samusenps", "andrewrreed", "m-ric", "lunarflu" ], "count": 6 } ]

[]

[ { "type": "text", "value": "🔥🔥🔥 Introducing WizardLM-2!", "raw": "🔥🔥🔥 Introducing WizardLM-2!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📙Release Blog: ", "raw": "📙Release Blog: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://wizardlm.github.io/WizardLM2", "href": "https://wizardlm.github.io/WizardLM2", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "✅Model Weights: ", "raw": "✅Model Weights: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/microsoft/wizardlm-661d403f71e6c8257dbd598a", "href": null, "resource": { "type": "collection", "id": "microsoft/wizardlm-661d403f71e6c8257dbd598a", "discussionNum": null }, "url": "https://huggingface.co/collections/microsoft/wizardlm-661d403f71e6c8257dbd598a", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🐦Twitter: ", "raw": "🐦Twitter: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://twitter.com/WizardLM_AI/status/1779899325868589372", "href": "https://twitter.com/WizardLM_AI/status/1779899325868589372", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We introduce and opensource WizardLM-2, our next generation state-of-the-art large language models, which have improved performance on complex chat, multilingual, reasoning and agent. New family includes three cutting-edge models: WizardLM-2 8x22B, WizardLM-2 70B, and WizardLM-2 7B.", "raw": "We introduce and opensource WizardLM-2, our next generation state-of-the-art large language models, which have improved performance on complex chat, multilingual, reasoning and agent. New family includes three cutting-edge models: WizardLM-2 8x22B, WizardLM-2 70B, and WizardLM-2 7B.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "WizardLM-2 8x22B is our most advanced model, and the best opensource LLM in our internal evaluation on highly complex tasks. WizardLM-2 70B reaches top-tier reasoning capabilities and is the first choice in the same size. WizardLM-2 7B is the fastest and achieves comparable performance with existing 10x larger opensource leading models.", "raw": "WizardLM-2 8x22B is our most advanced model, and the best opensource LLM in our internal evaluation on highly complex tasks. WizardLM-2 70B reaches top-tier reasoning capabilities and is the first choice in the same size. WizardLM-2 7B is the fastest and achieves comparable performance with existing 10x larger opensource leading models.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤗 WizardLM 2 Capacities:", "raw": "🤗 WizardLM 2 Capacities:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "1. MT-Bench (Figure-1)", "raw": "1. MT-Bench (Figure-1)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The WizardLM-2 8x22B even demonstrates highly competitive performance compared to the most advanced proprietary works such as GPT-4-Trubo and Glaude-3. Meanwhile, WizardLM-2 7B and WizardLM-2 70B are all the top-performing models among the other leading baselines at 7B to 70B model scales.", "raw": "The WizardLM-2 8x22B even demonstrates highly competitive performance compared to the most advanced proprietary works such as GPT-4-Trubo and Glaude-3. Meanwhile, WizardLM-2 7B and WizardLM-2 70B are all the top-performing models among the other leading baselines at 7B to 70B model scales.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "2. Human Preferences Evaluation (Figure 2)", "raw": "2. Human Preferences Evaluation (Figure 2)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Through this human preferences evaluation, WizardLM-2's capabilities are very close to the cutting-edge proprietary models such as GPT-4-1106-preview, and significantly ahead of all the other open source models.", "raw": "Through this human preferences evaluation, WizardLM-2's capabilities are very close to the cutting-edge proprietary models such as GPT-4-1106-preview, and significantly ahead of all the other open source models.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍Method Overview:", "raw": "🔍Method Overview:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "As the natural world's human-generated data becomes increasingly exhausted through LLM training, we believe that: the data carefully created by AI and the model step-by-step supervised by AI will be the sole path towards more powerful AI.", "raw": "As the natural world's human-generated data becomes increasingly exhausted through LLM training, we believe that: the data carefully created by AI and the model step-by-step supervised by AI will be the sole path towards more powerful AI.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "In the past one year, we built a fully AI powered synthetic training system. (As shown in the Figure 3). ", "raw": "In the past one year, we built a fully AI powered synthetic training system. (As shown in the Figure 3). ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/627b9f3f4d0858f0034efbb9/_uYEW3gST1aPlWx5qOPIo.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/627b9f3f4d0858f0034efbb9/_wBregVEvOpA8xqQGUUcH.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/627b9f3f4d0858f0034efbb9/Fph9BTXej8_1DkfnvKj0O.png" } ]

[]

[ { "reaction": "🚀", "users": [ "WizardLM", "tangmen", "victor", "vaclavkosar", "s0me-0ne", "oceansweep", "Bruno", "xianbao", "haipeng1", "syslot", "Lin-Rexter", "Chaklam", "JosMax", "veeragoni", "CocoSun", "ChuckMcSneed", "osanseviero", "julien-c", "Ramikan-BR", "niclaslindqvist", "radames", "taher30", "jinkazamapro07", "m-ric", "jayjay5230", "Jaward", "latent-variable", "jameshwade", "Revs64", "JSCPPProgrammer", "cppowboy", "ShaneTian", "jlzhou", "alesthehuman", "bardusco-tau", "mahirzukic2", "noobmaster29", "Emigmo", "KvrParaskevi", "Aressions" ], "count": 40 }, { "reaction": "🔥", "users": [ "haipeng1", "Lin-Rexter", "osanseviero", "julien-c", "radames", "kramp", "latent-variable", "djester", "alesthehuman", "bardusco-tau", "mahirzukic2", "noobmaster29", "c4wrd", "wucl", "usholanb", "bs82", "mitra-mir" ], "count": 17 }, { "reaction": "🤗", "users": [ "sparkycollier", "haipeng1", "Lin-Rexter", "YueJie", "osanseviero", "julien-c", "ragibhasan", "radames", "latent-variable", "sugatoray", "alesthehuman", "mahirzukic2", "noobmaster29" ], "count": 13 }, { "reaction": "👍", "users": [ "KyleLyu", "djester", "alesthehuman", "pharmaia", "noobmaster29", "tobitege", "StudentWen", "VDBLOI2024" ], "count": 8 }, { "reaction": "❤️", "users": [ "Tenbo", "fysp", "latent-variable", "alesthehuman", "noobmaster29" ], "count": 5 }, { "reaction": "👀", "users": [ "bs82" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/627b9f3f4d0858f0034efbb9/2Qnattrzv6qvqiZVVfV5x.png", "fullname": "WizardLM", "name": "WizardLM", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3191, "isFollowing": false }, { "avatarUrl": "/avatars/58b3b0475d38435ac147a285fa313d4a.svg", "fullname": "Logan ", "name": "rakataprime", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 4, "isFollowing": false }, { "avatarUrl": "/avatars/3f9d8b01fbf255171b5e55bb751b1290.svg", "fullname": "yuhannaa", "name": "yuhannaa", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false }, { "avatarUrl": "/avatars/ea4398745974d781ae9dc0e95b12cabe.svg", "fullname": "Joseph", "name": "Joseph717171", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 22, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/630ebe6881bf616c051f19f9/u_8_QRMOjmXCKOBct_8DH.png", "fullname": "Arthur Colle", "name": "arthurcolle", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/640261025caf6d21d67d01ba/c8p-8DXo6Kz9eg7YvR0Nd.png", "fullname": "Shaun Prince", "name": "Suparious", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 27, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1677134945205-62f32eab52ad88c930bb3f3b.png", "fullname": "Asankhaya Sharma", "name": "codelion", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 46, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/5dd96eb166059660ed1ee413/NQtzmrDdbG0H8qkZvRyGk.jpeg", "fullname": "Julien Chaumond", "name": "julien-c", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 1580, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1670448346835-6391029cb73a0f8c02a7947f.png", "fullname": "Hailey Collet", "name": "HaileyStorm", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 6, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64e6364e362fc663a567d560/4t6Yh8B1x_pW6J5rXl_Uw.jpeg", "fullname": "Kanta Hayashi", "name": "misdelivery", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 5, "isFollowing": false }, { "avatarUrl": "/avatars/55f471b03d0ee29d9ccca22c7b9c2fb1.svg", "fullname": "Jay W", "name": "JayDoubleu", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/h-yM9HQrosNku5oXP7Iuu.jpeg", "fullname": "Jackson Hall", "name": "jacksth22", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/1d35c9c4bada277b160abf06093a5f16.svg", "fullname": "Cameron Rafaela", "name": "Gr8-CLR", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/d694ccd2c580220a0a860331278e083e.svg", "fullname": "sdfsdf", "name": "fdgdgadbvsaz", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/62716bc83623fb51324c8539/nbifF_3W3CbjTU0T2hfS9.jpeg", "fullname": "Awal Premi", "name": "Awal", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1678284286114-noauth.png", "fullname": "Milan Fornůsek", "name": "Nendos", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/7cc7d587a851f3a54fc8883df0fe5ddb.svg", "fullname": "Chris Swanson", "name": "chrisjswanson", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1640356718818-61c47e9c71a107e9d80e33e3.jpeg", "fullname": "Henky!!", "name": "Henk717", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 82, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64c75f7f5671d42e0afa86f0/3-DSbC90ehSCnzBcRYxPk.jpeg", "fullname": "HyeonwooKim", "name": "choco9966", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 7, "isFollowing": false }, { "avatarUrl": "/avatars/ef847bdde7b4a6e07b874da7c59f287b.svg", "fullname": "Ben Burns", "name": "benjamincburns", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/c148a32d2e7deeeea6db252b90973ea0.svg", "fullname": "AaRon", "name": "AARon99", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3, "isFollowing": false }, { "avatarUrl": "/avatars/89e38fb5e7f294165561e246720ca014.svg", "fullname": "S Srihari", "name": "Sriharimcw", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/4f651e1eec3b6b33aaf5644ea99ebb6b.svg", "fullname": "Tom K.", "name": "ToKrCZ", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2, "isFollowing": false }, { "avatarUrl": "/avatars/e57287ea18a74ceb62f4eca1323de1e4.svg", "fullname": "Alec Lazarescu", "name": "aleclaza", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/4VOzArmrRaX_DUTxGmm59.jpeg", "fullname": "Charles McSneed", "name": "ChuckMcSneed", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 57, "isFollowing": false }, { "avatarUrl": "/avatars/02b7c8cf48dcbd52aab9ffae3b8addb2.svg", "fullname": "Pizza", "name": "Giuseppe1971", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/5hlzrS1M4OY_ve12S7Vvg.jpeg", "fullname": "Merie Eze", "name": "Merie08", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "Just tested Argilla's new data annotation feature - it's a game changer for AI project quality. ", "raw": "Just tested Argilla's new data annotation feature - it's a game changer for AI project quality. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Upload CSVs, work with published datasets, or improve existing ones directly on HuggingFace Hub. Setup took < 2 minutes, no code needed (see example below where I selected a dataset to classify tweets in categories).", "raw": "Upload CSVs, work with published datasets, or improve existing ones directly on HuggingFace Hub. Setup took < 2 minutes, no code needed (see example below where I selected a dataset to classify tweets in categories).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Real world impact: Missing in Chicago won a Pulitzer using a similar approach - 200 volunteers labeled police misconduct files to train their model. That's the power of good data annotation.", "raw": "Real world impact: Missing in Chicago won a Pulitzer using a similar approach - 200 volunteers labeled police misconduct files to train their model. That's the power of good data annotation.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Three immediate use cases I see:", "raw": "Three immediate use cases I see:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Build collaborative training sets with your community (surprisingly underused in AI journalism)", "raw": "- Build collaborative training sets with your community (surprisingly underused in AI journalism)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Turn your website chatbot logs into high-quality fine-tuning data", "raw": "- Turn your website chatbot logs into high-quality fine-tuning data", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Compare generated vs published content (great for SEO headlines)", "raw": "- Compare generated vs published content (great for SEO headlines)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Works for solo projects or teams up to 100 people. All integrated with HuggingFace Hub for immediate model training.", "raw": "Works for solo projects or teams up to 100 people. All integrated with HuggingFace Hub for immediate model training.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Interesting to see tools like this making data quality more accessible. Data quality is the hidden driver of AI success that we don't talk about enough.", "raw": "Interesting to see tools like this making data quality more accessible. Data quality is the hidden driver of AI success that we don't talk about enough.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Check out the blogpost: ", "raw": "- Check out the blogpost: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/argilla-ui-hub", "href": "https://huggingface.co/blog/argilla-ui-hub", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- And the quickstart guide: ", "raw": "- And the quickstart guide: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://docs.argilla.io/latest/getting_started/quickstart/", "href": "https://docs.argilla.io/latest/getting_started/quickstart/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/647f36a8454af0237bd49574/lXUDPMW88hbFTCIpkylDT.mp4" } ]

[]

[ { "reaction": "➕", "users": [ "pradnyab", "alielfilali01", "OmbelineM" ], "count": 3 }, { "reaction": "👀", "users": [ "John6666", "djuna", "Rsln" ], "count": 3 }, { "reaction": "❤️", "users": [ "alielfilali01", "julietbaby" ], "count": 2 }, { "reaction": "🔥", "users": [ "abdullah" ], "count": 1 } ]

[]

[ { "type": "text", "value": "𝗛𝘂𝗻𝘆𝘂𝗮𝗻-𝗟𝗮𝗿𝗴𝗲 𝗷𝘂𝘀𝘁 𝗿𝗲𝗹𝗲𝗮𝘀𝗲𝗱 𝗯𝘆 𝗧𝗲𝗻𝗰𝗲𝗻𝘁: 𝗟𝗮𝗿𝗴𝗲𝘀𝘁 𝗲𝘃𝗲𝗿 𝗼𝗽𝗲𝗻 𝗠𝗼𝗘 𝗟𝗟𝗠, 𝗼𝗻𝗹𝘆 𝟱𝟮𝗕 𝗮𝗰𝘁𝗶𝘃𝗲 𝗽𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿𝘀 𝗯𝘂𝘁 𝗯𝗲𝗮𝘁𝘀 𝗟𝗟𝗮𝗠𝗔 𝟯.𝟭-𝟰𝟬𝟱𝗕 𝗼𝗻 𝗺𝗼𝘀𝘁 𝗮𝗰𝗮𝗱𝗲𝗺𝗶𝗰 𝗯𝗲𝗻𝗰𝗵𝗺𝗮𝗿𝗸𝘀 🚀", "raw": "𝗛𝘂𝗻𝘆𝘂𝗮𝗻-𝗟𝗮𝗿𝗴𝗲 𝗷𝘂𝘀𝘁 𝗿𝗲𝗹𝗲𝗮𝘀𝗲𝗱 𝗯𝘆 𝗧𝗲𝗻𝗰𝗲𝗻𝘁: 𝗟𝗮𝗿𝗴𝗲𝘀𝘁 𝗲𝘃𝗲𝗿 𝗼𝗽𝗲𝗻 𝗠𝗼𝗘 𝗟𝗟𝗠, 𝗼𝗻𝗹𝘆 𝟱𝟮𝗕 𝗮𝗰𝘁𝗶𝘃𝗲 𝗽𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿𝘀 𝗯𝘂𝘁 𝗯𝗲𝗮𝘁𝘀 𝗟𝗟𝗮𝗠𝗔 𝟯.𝟭-𝟰𝟬𝟱𝗕 𝗼𝗻 𝗺𝗼𝘀𝘁 𝗮𝗰𝗮𝗱𝗲𝗺𝗶𝗰 𝗯𝗲𝗻𝗰𝗵𝗺𝗮𝗿𝗸𝘀 🚀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "⚡ Mixture of Experts (MoE) architecture: 389 B parameters in total, but only 52B are activated for any input", "raw": "⚡ Mixture of Experts (MoE) architecture: 389 B parameters in total, but only 52B are activated for any input", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🧪 Trained on 7T tokens, including 1.5T tokens of synthetic data", "raw": "🧪 Trained on 7T tokens, including 1.5T tokens of synthetic data", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🏗️ Architecture : Novel \"recycle routing\" prevents token dropping when experts are overrloaded", "raw": "🏗️ Architecture : Novel \"recycle routing\" prevents token dropping when experts are overrloaded", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📊 Great benchmark results: Surpasses Llama-3-405B-Instruct in most benchmarks although it has 8x fewer active parameters", "raw": "📊 Great benchmark results: Surpasses Llama-3-405B-Instruct in most benchmarks although it has 8x fewer active parameters", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "‣ Impressive perf on MATH: 77.4", "raw": "‣ Impressive perf on MATH: 77.4", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🐋 Large context length: up to 256K tokens", "raw": "🐋 Large context length: up to 256K tokens", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔒 License:", "raw": "🔒 License:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "‣ Commercial use allowed, except if your products have >100M monthly active users", "raw": "‣ Commercial use allowed, except if your products have >100M monthly active users", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "‣ No access in the EU", "raw": "‣ No access in the EU", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤗 Model weights available on HF!", "raw": "🤗 Model weights available on HF!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Read the full paper here 👉 ", "raw": "Read the full paper here 👉 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2411.02265", "href": null, "resource": { "type": "paper", "id": "2411.02265", "discussionNum": null }, "url": "https://huggingface.co/papers/2411.02265", "code": null, "user": null, "label": "Hunyuan-Large: An Open-Source MoE Model with 52 Billion Activated\n Parameters by Tencent (2411.02265)", "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/63d10d4e8eaa4831005e92b5/3LPlzA5TVmgNvE-J3kg5R.png" } ]

[]

[ { "reaction": "🔥", "users": [ "Hooo5", "YaTharThShaRma999", "John6666", "Chief-Inspector" ], "count": 4 }, { "reaction": "🚀", "users": [ "Hooo5", "fdaudens", "YaTharThShaRma999", "prithivMLmods" ], "count": 4 }, { "reaction": "👍", "users": [ "Rsln" ], "count": 1 } ]

[]

[ { "type": "text", "value": "🎓 Introducing Kompy.info Uzbek Educational Dataset - ", "raw": "🎓 Introducing Kompy.info Uzbek Educational Dataset - ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/nyuuzyou/kompy", "href": null, "resource": { "type": "dataset", "id": "nyuuzyou/kompy", "discussionNum": null }, "url": "https://huggingface.co/datasets/nyuuzyou/kompy", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Dataset highlights:", "raw": "Dataset highlights:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- 584,648 pages of educational content extracted from kompy.info, a comprehensive educational resource website", "raw": "- 584,648 pages of educational content extracted from kompy.info, a comprehensive educational resource website", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Content exclusively in Uzbek language, focusing on technical and scientific topics", "raw": "- Content exclusively in Uzbek language, focusing on technical and scientific topics", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Each entry contains: URL, page title, and extracted main text content", "raw": "- Each entry contains: URL, page title, and extracted main text content", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Data extracted using trafilatura HTML extraction tool", "raw": "- Data extracted using trafilatura HTML extraction tool", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Covers a wide range of academic and educational materials", "raw": "- Covers a wide range of academic and educational materials", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Released to the public domain under Creative Commons Zero (CC0) license", "raw": "- Released to the public domain under Creative Commons Zero (CC0) license", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The dataset presents a valuable resource for natural language processing tasks in the Uzbek language, particularly in educational and technical domains. It can be used for text classification, topic modeling, and content analysis of educational materials. The large-scale collection of Uzbek-language academic content makes it especially useful for developing educational technology applications and studying pedagogical approaches in Uzbek-language instruction. The dataset's monolingual nature provides a focused corpus for understanding technical and scientific terminology in Uzbek educational contexts.", "raw": "The dataset presents a valuable resource for natural language processing tasks in the Uzbek language, particularly in educational and technical domains. It can be used for text classification, topic modeling, and content analysis of educational materials. The large-scale collection of Uzbek-language academic content makes it especially useful for developing educational technology applications and studying pedagogical approaches in Uzbek-language instruction. The dataset's monolingual nature provides a focused corpus for understanding technical and scientific terminology in Uzbek educational contexts.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "👀", "users": [ "John6666", "djuna" ], "count": 2 } ]

[]

[ { "type": "text", "value": "🚀 Excited to introduce a new member of the OS-Copilot family: OS-Atlas - an open-sourced foundational action model for GUI agents", "raw": "🚀 Excited to introduce a new member of the OS-Copilot family: OS-Atlas - an open-sourced foundational action model for GUI agents", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📘 Paper: ", "raw": "📘 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2410.23218", "href": null, "resource": { "type": "paper", "id": "2410.23218", "discussionNum": null }, "url": "https://huggingface.co/papers/2410.23218", "code": null, "user": null, "label": "OS-ATLAS: A Foundation Action Model for Generalist GUI Agents (2410.23218)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔗 Website: ", "raw": "🔗 Website: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://osatlas.github.io", "href": "https://osatlas.github.io", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "😇 TL;DR: OS-Atlas offers:", "raw": "😇 TL;DR: OS-Atlas offers:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "1. State-of-the-Art GUI Grounding: Helps GUI agents accurately locate GUI elements.", "raw": "1. State-of-the-Art GUI Grounding: Helps GUI agents accurately locate GUI elements.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "2. Strong OOD Performance and Cross-platform Compatibility: Excels in out-of-domain agentic tasks across MacOS, Windows, Linux, Android, and Web. ", "raw": "2. Strong OOD Performance and Cross-platform Compatibility: Excels in out-of-domain agentic tasks across MacOS, Windows, Linux, Android, and Web. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3. Complete Infrastructure for GUI Data Synthesis: ", "raw": "3. Complete Infrastructure for GUI Data Synthesis: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "You can easily build your own OS agent upon it!", "raw": "You can easily build your own OS agent upon it!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🚀", "users": [ "Symbol-LLM", "John6666", "MH0386" ], "count": 3 }, { "reaction": "🔥", "users": [ "Symbol-LLM", "topaasi" ], "count": 2 } ]

[]

[ { "type": "text", "value": "🚀 New feature of the Comparator of the 🤗 Open LLM Leaderboard: now compare models with their base versions & derivatives (finetunes, adapters, etc.). Perfect for tracking how adjustments affect performance & seeing innovations in action. Dive deeper into the leaderboard!", "raw": "🚀 New feature of the Comparator of the 🤗 Open LLM Leaderboard: now compare models with their base versions & derivatives (finetunes, adapters, etc.). Perfect for tracking how adjustments affect performance & seeing innovations in action. Dive deeper into the leaderboard!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🛠️ Here's how to use it:", "raw": "🛠️ Here's how to use it:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "1. Select your model from the leaderboard.", "raw": "1. Select your model from the leaderboard.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "2. Load its model tree.", "raw": "2. Load its model tree.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3. Choose any base & derived models (adapters, finetunes, merges, quantizations) for comparison.", "raw": "3. Choose any base & derived models (adapters, finetunes, merges, quantizations) for comparison.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "4. Press Load.", "raw": "4. Press Load.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "See side-by-side performance metrics instantly!", "raw": "See side-by-side performance metrics instantly!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Ready to dive in? 🏆 Try the 🤗 Open LLM Leaderboard Comparator now! See how models stack up against their base versions and derivatives to understand fine-tuning and other adjustments. Easier model analysis for better insights! Check it out here: ", "raw": "Ready to dive in? 🏆 Try the 🤗 Open LLM Leaderboard Comparator now! See how models stack up against their base versions and derivatives to understand fine-tuning and other adjustments. Easier model analysis for better insights! Check it out here: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/open-llm-leaderboard/comparator", "href": null, "resource": { "type": "space", "id": "open-llm-leaderboard/comparator", "discussionNum": null }, "url": "https://huggingface.co/spaces/open-llm-leaderboard/comparator", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " 🌐", "raw": " 🌐", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5fbfd09ee366524fe8e97cd3/qOVAuQMQ4EcpTf6TyfvEU.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5fbfd09ee366524fe8e97cd3/KdBSQXL0l8mGZKWyW4dhl.png" } ]

[]

[ { "reaction": "👀", "users": [ "John6666", "djuna" ], "count": 2 } ]

[]

[ { "type": "resource", "value": null, "raw": "https://huggingface.co/Alfitaria/Q25-1.5B-VeoLu", "href": null, "resource": { "type": "model", "id": "Alfitaria/Q25-1.5B-VeoLu", "discussionNum": null }, "url": "https://huggingface.co/Alfitaria/Q25-1.5B-VeoLu", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Q2.5-1.5-VeoLu is a 1.5 billion parameter General Purpose Creative model trained on Qwen2.5-1.5B-Instruct. Intended mostly as an educational process for myself, Veo Lu nevertheless manages to be usable most of the time, while also being light enough to potentially run on a smartphone.", "raw": "Q2.5-1.5-VeoLu is a 1.5 billion parameter General Purpose Creative model trained on Qwen2.5-1.5B-Instruct. Intended mostly as an educational process for myself, Veo Lu nevertheless manages to be usable most of the time, while also being light enough to potentially run on a smartphone.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6685d39f64da708c0f553c5d/zGH2GfvXTV_2fzJYsb3ro.png" } ]

[]

[ { "reaction": "👀", "users": [ "John6666", "bannerlord", "djuna", "AuriAetherwiing", "EloyOn", "Rsln" ], "count": 6 } ]

[]

[ { "type": "text", "value": "Hi HugginfgFacers!🤗", "raw": "Hi HugginfgFacers!🤗", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "If you're into biomedical sciences, you will know the pain that, sometimes, searching PubMed can be🙇‍♀️", "raw": "If you're into biomedical sciences, you will know the pain that, sometimes, searching PubMed can be🙇‍♀️", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "For these purposes, I built a bot that scrapes PubMed for you, starting from the exact title of a publication or key word search - all beautifully rendered through Gradio✅", "raw": "For these purposes, I built a bot that scrapes PubMed for you, starting from the exact title of a publication or key word search - all beautifully rendered through Gradio✅", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Find it here: ", "raw": "Find it here: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/as-cle-bert/BioMedicalPapersBot", "href": null, "resource": { "type": "space", "id": "as-cle-bert/BioMedicalPapersBot", "discussionNum": null }, "url": "https://huggingface.co/spaces/as-cle-bert/BioMedicalPapersBot", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "And here's the GitHub repository🐱: ", "raw": "And here's the GitHub repository🐱: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/AstraBert/BioMedicalPapersBot", "href": "https://github.com/AstraBert/BioMedicalPapersBot", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "It's also available as a Docker image!🐳", "raw": "It's also available as a Docker image!🐳", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "code_fence", "value": null, "raw": "```\ndocker pull ghcr.io/astrabert/biomedicalpapersbot:main\n```", "href": null, "resource": null, "url": null, "code": "docker pull ghcr.io/astrabert/biomedicalpapersbot:main", "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Best of luck with your research!", "raw": "Best of luck with your research!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "PS: in the very near future some AI summarization features will be included!", "raw": "PS: in the very near future some AI summarization features will be included!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "👍", "users": [ "ZyAi", "eskayML" ], "count": 2 }, { "reaction": "👀", "users": [ "John6666" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Exciting Research Alert: Revolutionizing Dense Passage Retrieval with Entailment Tuning!", "raw": "Exciting Research Alert: Revolutionizing Dense Passage Retrieval with Entailment Tuning!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The good folks at HKUST have developed a novel approach that significantly improves information retrieval by leveraging natural language inference.", "raw": "The good folks at HKUST have developed a novel approach that significantly improves information retrieval by leveraging natural language inference.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The entailment tuning approach consists of several key steps to enhance dense passage retrieval performance.", "raw": "The entailment tuning approach consists of several key steps to enhance dense passage retrieval performance.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Data Preparation", "raw": "Data Preparation", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Convert questions into existence claims using rule-based transformations.", "raw": "- Convert questions into existence claims using rule-based transformations.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Combine retrieval data with NLI data from SNLI and MNLI datasets.", "raw": "- Combine retrieval data with NLI data from SNLI and MNLI datasets.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Unify the format of both data types using a consistent prompting framework.", "raw": "- Unify the format of both data types using a consistent prompting framework.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Entailment Tuning Process", "raw": "Entailment Tuning Process", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Initialize the model using pre-trained language models like BERT or RoBERTa.", "raw": "- Initialize the model using pre-trained language models like BERT or RoBERTa.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Apply aggressive masking (β=0.8) specifically to the hypothesis components while preserving premise information.", "raw": "- Apply aggressive masking (β=0.8) specifically to the hypothesis components while preserving premise information.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Train the model to predict the masked hypothesis tokens from the premise content.", "raw": "- Train the model to predict the masked hypothesis tokens from the premise content.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Run the training for 10 epochs using 8 GPUs, taking approximately 1.5-3.5 hours.", "raw": "- Run the training for 10 epochs using 8 GPUs, taking approximately 1.5-3.5 hours.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Training Arguments for Entailment Tuning (Yes! They Shared Them)", "raw": "Training Arguments for Entailment Tuning (Yes! They Shared Them)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Use a learning rate of 2e-5 with 100 warmup steps.", "raw": "- Use a learning rate of 2e-5 with 100 warmup steps.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Set batch size to 128.", "raw": "- Set batch size to 128.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Apply weight decay of 0.01.", "raw": "- Apply weight decay of 0.01.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Utilize the Adam optimizer with beta values (0.9, 0.999).", "raw": "- Utilize the Adam optimizer with beta values (0.9, 0.999).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Maintain maximum gradient norm at 1.0.", "raw": "- Maintain maximum gradient norm at 1.0.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Deployment", "raw": "Deployment", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Index passages using FAISS for efficient retrieval.", "raw": "- Index passages using FAISS for efficient retrieval.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Shard vector store across multiple GPUs.", "raw": "- Shard vector store across multiple GPUs.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Enable sub-millisecond retrieval of the top-100 passages per query.", "raw": "- Enable sub-millisecond retrieval of the top-100 passages per query.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Integration with Existing Systems", "raw": "Integration with Existing Systems", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Insert entailment tuning between pre-training and fine-tuning stages.", "raw": "- Insert entailment tuning between pre-training and fine-tuning stages.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Maintain compatibility with current dense retrieval methods.", "raw": "- Maintain compatibility with current dense retrieval methods.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Preserve existing contrastive learning approaches during fine-tuning.", "raw": "- Preserve existing contrastive learning approaches during fine-tuning.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Simple, intuitive, and effective!", "raw": "Simple, intuitive, and effective!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This advancement significantly improves the quality of retrieved passages for question-answering systems and retrieval-augmented generation tasks.", "raw": "This advancement significantly improves the quality of retrieved passages for question-answering systems and retrieval-augmented generation tasks.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/662bf5bfe93bb73804ef9344/kffXQV_Z_yfdXofVCTzxK.jpeg" } ]

[]

[ { "reaction": "👀", "users": [ "John6666", "Kaba", "louisbrulenaudet", "tiendung", "KarthikaRajagopal", "AtAndDev", "Bruno" ], "count": 7 } ]

[]

[ { "type": "text", "value": "Smol TTS models are here! OuteTTS-0.1-350M - Zero shot voice cloning, built on LLaMa architecture, CC-BY license! 🔥 ", "raw": "Smol TTS models are here! OuteTTS-0.1-350M - Zero shot voice cloning, built on LLaMa architecture, CC-BY license! 🔥 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> Pure language modeling approach to TTS", "raw": "> Pure language modeling approach to TTS", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> Zero-shot voice cloning", "raw": "> Zero-shot voice cloning", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> LLaMa architecture w/ Audio tokens (WavTokenizer)", "raw": "> LLaMa architecture w/ Audio tokens (WavTokenizer)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> BONUS: Works on-device w/ llama.cpp ⚡", "raw": "> BONUS: Works on-device w/ llama.cpp ⚡", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Three-step approach to TTS:", "raw": "Three-step approach to TTS:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> Audio tokenization using WavTokenizer (75 tok per second)", "raw": "> Audio tokenization using WavTokenizer (75 tok per second)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> CTC forced alignment for word-to-audio token mapping", "raw": "> CTC forced alignment for word-to-audio token mapping", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> Structured prompt creation w/ transcription, duration, audio tokens", "raw": "> Structured prompt creation w/ transcription, duration, audio tokens", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The model is extremely impressive for 350M parameters! Kudos to the ", "raw": "The model is extremely impressive for 350M parameters! Kudos to the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "OuteAI team on such a brilliant feat - I'd love to see this be applied on larger data and smarter backbones like SmolLM 🤗", "raw": "OuteAI team on such a brilliant feat - I'd love to see this be applied on larger data and smarter backbones like SmolLM 🤗", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Check out the models here: ", "raw": "Check out the models here: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/OuteAI/outetts-6728aa71a53a076e4ba4817c", "href": null, "resource": { "type": "collection", "id": "OuteAI/outetts-6728aa71a53a076e4ba4817c", "discussionNum": null }, "url": "https://huggingface.co/collections/OuteAI/outetts-6728aa71a53a076e4ba4817c", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/61b85ce86eb1f2c5e6233736/jH7WoESriWkSMxZ5AFHKq.mp4" } ]

[]

[ { "reaction": "👀", "users": [ "John6666", "s3nh" ], "count": 2 } ]

[]

[ { "type": "text", "value": "Build datasets for AI on the Hugging Face Hub—10x easier than ever!", "raw": "Build datasets for AI on the Hugging Face Hub—10x easier than ever!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Today, I'm excited to share our biggest feature since we joined Hugging Face. ", "raw": "Today, I'm excited to share our biggest feature since we joined Hugging Face. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Here’s how it works:", "raw": "Here’s how it works:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "1. Pick a dataset—upload your own or choose from 240K open datasets.", "raw": "1. Pick a dataset—upload your own or choose from 240K open datasets.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "2. Paste the Hub dataset ID into Argilla and set up your labeling interface.", "raw": "2. Paste the Hub dataset ID into Argilla and set up your labeling interface.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3. Share the URL with your team or the whole community!", "raw": "3. Share the URL with your team or the whole community!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "And the best part? It’s:", "raw": "And the best part? It’s:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- No code – no Python needed", "raw": "- No code – no Python needed", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Integrated – all within the Hub", "raw": "- Integrated – all within the Hub", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Scalable – from solo labeling to 100s of contributors", "raw": "- Scalable – from solo labeling to 100s of contributors", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I am incredibly proud of the team for shipping this after weeks of work and many quick iterations.", "raw": "I am incredibly proud of the team for shipping this after weeks of work and many quick iterations.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Let's make this sentence obsolete: \"Everyone wants to do the model work, not the data work.\"", "raw": "Let's make this sentence obsolete: \"Everyone wants to do the model work, not the data work.\"", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Read, share, and like the HF blog post:", "raw": "Read, share, and like the HF blog post:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/argilla-ui-hub", "href": "https://huggingface.co/blog/argilla-ui-hub", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "👍", "users": [ "John6666", "Anthony10", "NickyNicky" ], "count": 3 }, { "reaction": "👀", "users": [ "John6666" ], "count": 1 } ]

[]

[ { "type": "text", "value": "😀🤓😎 New Space - OCEAN-AI (App, co-authored by ", "raw": "😀🤓😎 New Space - OCEAN-AI (App, co-authored by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DmitryRyumin", "href": null, "resource": null, "url": null, "code": null, "user": "DmitryRyumin", "label": null, "lang": null }, { "type": "text", "value": ") 😎😉😤", "raw": ") 😎😉😤", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Title: OCEAN-AI is an open-source app for Big Five personality traits assessment and HR-processes automatization.", "raw": "🚀 Title: OCEAN-AI is an open-source app for Big Five personality traits assessment and HR-processes automatization.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤗 Demo: ", "raw": "🤗 Demo: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/ElenaRyumina/OCEANAI", "href": null, "resource": { "type": "space", "id": "ElenaRyumina/OCEANAI", "discussionNum": null }, "url": "https://huggingface.co/spaces/ElenaRyumina/OCEANAI", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👥 Authors: ", "raw": "👥 Authors: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@ElenaRyumina", "href": null, "resource": null, "url": null, "code": null, "user": "ElenaRyumina", "label": null, "lang": null }, { "type": "text", "value": ", ", "raw": ", ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DmitryRyumin", "href": null, "resource": null, "url": null, "code": null, "user": "DmitryRyumin", "label": null, "lang": null }, { "type": "text", "value": ", and Alexey Karpov", "raw": ", and Alexey Karpov", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📝 Description: OCEAN-AI consists of a set of modules for intellectual analysis of human behavior based on multimodal data for automatic personality traits (PT) assessment. The app evaluates five PT: Openness to experience, Conscientiousness, Extraversion, Agreeableness, Non-Neuroticism.", "raw": "📝 Description: OCEAN-AI consists of a set of modules for intellectual analysis of human behavior based on multimodal data for automatic personality traits (PT) assessment. The app evaluates five PT: Openness to experience, Conscientiousness, Extraversion, Agreeableness, Non-Neuroticism.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The App solves practical tasks:", "raw": "The App solves practical tasks:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Ranking of potential candidates by professional responsibilities.", "raw": "- Ranking of potential candidates by professional responsibilities.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Forming effective work teams.", "raw": "- Forming effective work teams.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Predicting consumer preferences for industrial goods.", "raw": "- Predicting consumer preferences for industrial goods.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Keywords: #OCEANAI #BigFive #PersonalityTraits #PersonalityAnalysis #MultimodalData #Transformers #FirstImpressionsV2 #DeepLearning #Innovation #BehaviorAnalysis #AffectiveRecognition #TeamFormation #ConsumerPreferences #CandidateRanking", "raw": "🔍 Keywords: #OCEANAI #BigFive #PersonalityTraits #PersonalityAnalysis #MultimodalData #Transformers #FirstImpressionsV2 #DeepLearning #Innovation #BehaviorAnalysis #AffectiveRecognition #TeamFormation #ConsumerPreferences #CandidateRanking", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/NdHXteGpwfdYQ7g36TAHX.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/8l9Gmh7mKkZG2OVOvXb3c.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/7Jg0Gj2KuVEmKxboV3qwP.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/74sdvvKu8ENEkTsDL55Cp.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/RChhTv0qB8cOI3Ge-3l18.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/vyOj3rcrsd6lNe0U0ppRn.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/9Q8pK5NmjEJkfZgNGDyje.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/EsK-RBlvcJswmaSChgmko.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/kqK5uxuM_wBbel65jmY5s.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/0Kz1SoMF8TovCj9C5M5Km.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/7WWeqGGMBULz2vw_NVifI.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/nRCxbVng_PPBqKd-Z3KVc.jpeg", "fullname": "Dmitry Ryumin", "name": "DmitryRyumin", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 377 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/yQDxSx0Il0jwma_u2UyJe.jpeg", "fullname": "Elena Ryumina", "name": "ElenaRyumina", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 15 } ]

[ { "reaction": "🔥", "users": [ "DmitryRyumin", "ElenaRyumina", "Mihaj", "nilacadenas", "arturoxdev", "MiSTe-R", "Ramikan-BR", "kramp" ], "count": 8 }, { "reaction": "❤️", "users": [ "DmitryRyumin", "ElenaRyumina", "Mihaj", "MiSTe-R", "beno-rohman" ], "count": 5 }, { "reaction": "🚀", "users": [ "DmitryRyumin", "ElenaRyumina", "Mihaj", "MiSTe-R" ], "count": 4 }, { "reaction": "🤗", "users": [ "DmitryRyumin", "ElenaRyumina", "Mihaj", "MiSTe-R" ], "count": 4 }, { "reaction": "➕", "users": [ "raincandy-u" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6471d2075ffbc18f197a1e16/t1NA_KMUkYHKokuzfijLy.jpeg", "fullname": "ScottzModelz", "name": "ScottzModelz", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "𝐏𝐚𝐩𝐞𝐫 𝐑𝐞𝐯𝐢𝐞𝐰: 𝐑𝐡𝐨-𝟏 - 𝐃𝐨 𝐧𝐨𝐭 𝐮𝐬𝐞 𝐚𝐥𝐥 𝐭𝐨𝐤𝐞𝐧𝐬 𝐞𝐪𝐮𝐚𝐥𝐥𝐲 𝐢𝐧 𝐲𝐨𝐮𝐫 𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠! ⚖️⛔️", "raw": "𝐏𝐚𝐩𝐞𝐫 𝐑𝐞𝐯𝐢𝐞𝐰: 𝐑𝐡𝐨-𝟏 - 𝐃𝐨 𝐧𝐨𝐭 𝐮𝐬𝐞 𝐚𝐥𝐥 𝐭𝐨𝐤𝐞𝐧𝐬 𝐞𝐪𝐮𝐚𝐥𝐥𝐲 𝐢𝐧 𝐲𝐨𝐮𝐫 𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠! ⚖️⛔️", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "A new paper topping Daily papers questions a hidden assumption in LLM training:", "raw": "A new paper topping Daily papers questions a hidden assumption in LLM training:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤔 𝙎𝙝𝙤𝙪𝙡𝙙 𝙬𝙚 𝙧𝙚𝙖𝙡𝙡𝙮 𝙪𝙨𝙚 𝙖𝙡𝙡 𝙩𝙤𝙠𝙚𝙣𝙨 𝙚𝙦𝙪𝙖𝙡𝙡𝙮 𝙞𝙣 𝙤𝙪𝙧 𝙇𝙇𝙈'𝙨 𝙩𝙧𝙖𝙞𝙣𝙞𝙣𝙜 ?", "raw": "🤔 𝙎𝙝𝙤𝙪𝙡𝙙 𝙬𝙚 𝙧𝙚𝙖𝙡𝙡𝙮 𝙪𝙨𝙚 𝙖𝙡𝙡 𝙩𝙤𝙠𝙚𝙣𝙨 𝙚𝙦𝙪𝙖𝙡𝙡𝙮 𝙞𝙣 𝙤𝙪𝙧 𝙇𝙇𝙈'𝙨 𝙩𝙧𝙖𝙞𝙣𝙞𝙣𝙜 ?", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Some tokens are more relevant than others, and some are mostly noise (just look up the history of 𝘚𝘰𝘭𝘪𝘥𝘎𝘰𝘭𝘥𝘔𝘢𝘨𝘪𝘬𝘢𝘳𝘱).", "raw": "Some tokens are more relevant than others, and some are mostly noise (just look up the history of 𝘚𝘰𝘭𝘪𝘥𝘎𝘰𝘭𝘥𝘔𝘢𝘨𝘪𝘬𝘢𝘳𝘱).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "So this paper introduces 𝗦𝗲𝗹𝗲𝗰𝘁𝗶𝘃𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝗶𝗻𝗴, which is actually really simple:", "raw": "So this paper introduces 𝗦𝗲𝗹𝗲𝗰𝘁𝗶𝘃𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝗶𝗻𝗴, which is actually really simple:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "➡️ A specific metric measures the relevance of each token. Then during training, only the top k% tokens for this relevance metric count in the loss calculation.", "raw": "➡️ A specific metric measures the relevance of each token. Then during training, only the top k% tokens for this relevance metric count in the loss calculation.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Authors test this method by training models on the difficult MATH dataset (only competition mathematics problems).", "raw": "Authors test this method by training models on the difficult MATH dataset (only competition mathematics problems).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "➡️ Their technique seems like a new must-do in LLM training: Training is much faster and reaches an impressive performance!", "raw": "➡️ Their technique seems like a new must-do in LLM training: Training is much faster and reaches an impressive performance!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "𝐑𝐞𝐬𝐮𝐥𝐭𝐬:", "raw": "𝐑𝐞𝐬𝐮𝐥𝐭𝐬:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "◆ ⏱️ Training is x5 to x10 faster to reach equivalent performance compared to standard language modeling.", "raw": "◆ ⏱️ Training is x5 to x10 faster to reach equivalent performance compared to standard language modeling.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "◆ 💪 Their 1B model achieves close to GPT4 Chain-of-Thought performance on MATH!", "raw": "◆ 💪 Their 1B model achieves close to GPT4 Chain-of-Thought performance on MATH!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "◆ 🚀 Their 7B model match performance of the state-of-the-art DeepSeek for the same size, while trained on only 3% of tokens", "raw": "◆ 🚀 Their 7B model match performance of the state-of-the-art DeepSeek for the same size, while trained on only 3% of tokens", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "𝐀𝐝𝐝𝐢𝐭𝐢𝐨𝐧𝐚𝐥 𝐢𝐧𝐬𝐢𝐠𝐡𝐭𝐬 💡", "raw": "𝐀𝐝𝐝𝐢𝐭𝐢𝐨𝐧𝐚𝐥 𝐢𝐧𝐬𝐢𝐠𝐡𝐭𝐬 💡", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "◆ Datasets used for pre-training, even after pre-filtering, still contain a large proportion of noisy tokens 😖", "raw": "◆ Datasets used for pre-training, even after pre-filtering, still contain a large proportion of noisy tokens 😖", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "◆ Authors show that when you reduce loss on noisy tokens, you actually reduce accuracy (Figure 7). So Selective Language Modeling seems fundamental! ✅", "raw": "◆ Authors show that when you reduce loss on noisy tokens, you actually reduce accuracy (Figure 7). So Selective Language Modeling seems fundamental! ✅", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Find great reads in ", "raw": "Find great reads in ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@akhaliq", "href": null, "resource": null, "url": null, "code": null, "user": "akhaliq", "label": null, "lang": null }, { "type": "text", "value": " 's Daily Papers 👉 ", "raw": " 's Daily Papers 👉 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/papers", "href": "https://huggingface.co/papers", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Paper added to my collection 👉 ", "raw": "Paper added to my collection 👉 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/m-ric/spinning-up-in-llms-659e698f9dd5a71bd3f579a7", "href": null, "resource": { "type": "collection", "id": "m-ric/spinning-up-in-llms-659e698f9dd5a71bd3f579a7", "discussionNum": null }, "url": "https://huggingface.co/collections/m-ric/spinning-up-in-llms-659e698f9dd5a71bd3f579a7", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/63d10d4e8eaa4831005e92b5/i3KHQs8KnAD6HEAHystvi.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1674929746905-60f1abe7544c2adfd699860c.jpeg", "fullname": "AK", "name": "akhaliq", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 5205 } ]

[ { "reaction": "🔥", "users": [ "PatrickHaller", "alielfilali01" ], "count": 2 } ]

[]

[ { "type": "text", "value": "We just released a new MoE model (meraGPT/mera-mix-4x7B) that is half as large as Mixtral-8x7B while still been competitive with it across different benchmarks. mera-mix-4x7B achieves 76.37 on the open LLM eval.", "raw": "We just released a new MoE model (meraGPT/mera-mix-4x7B) that is half as large as Mixtral-8x7B while still been competitive with it across different benchmarks. mera-mix-4x7B achieves 76.37 on the open LLM eval.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "You can check mera-mix-4x7B out on HF here - ", "raw": "You can check mera-mix-4x7B out on HF here - ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/meraGPT/mera-mix-4x7B", "href": null, "resource": { "type": "model", "id": "meraGPT/mera-mix-4x7B", "discussionNum": null }, "url": "https://huggingface.co/meraGPT/mera-mix-4x7B", "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🔥", "users": [ "codelion", "osanseviero", "victor", "sosoai" ], "count": 4 } ]

[]

[ { "type": "text", "value": "Could more DPO-style preference data be crucial for enhancing open LLMs across different languages? ", "raw": "Could more DPO-style preference data be crucial for enhancing open LLMs across different languages? ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Leveraging a 7k preference dataset Argilla (", "raw": "Leveraging a 7k preference dataset Argilla (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@alvarobartt", "href": null, "resource": null, "url": null, "code": null, "user": "alvarobartt", "label": null, "lang": null }, { "type": "text", "value": "), Hugging Face (", "raw": "), Hugging Face (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@lewtun", "href": null, "resource": null, "url": null, "code": null, "user": "lewtun", "label": null, "lang": null }, { "type": "text", "value": ") and Kaist AI (", "raw": ") and Kaist AI (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@JW17", "href": null, "resource": null, "url": null, "code": null, "user": "JW17", "label": null, "lang": null }, { "type": "text", "value": " & ", "raw": " & ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@nlee-208", "href": null, "resource": null, "url": null, "code": null, "user": "nlee-208", "label": null, "lang": null }, { "type": "text", "value": ")", "raw": ")", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "utilized Kaist AI's recently introduced ORPO technique ", "raw": "utilized Kaist AI's recently introduced ORPO technique ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2403.07691", "href": null, "resource": { "type": "paper", "id": "2403.07691", "discussionNum": null }, "url": "https://huggingface.co/papers/2403.07691", "code": null, "user": null, "label": "ORPO: Monolithic Preference Optimization without Reference Model (2403.07691)", "lang": null }, { "type": "text", "value": " with the latest MistralAI MOE model to create a very high-performing open LLM: ", "raw": " with the latest MistralAI MOE model to create a very high-performing open LLM: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1", "href": null, "resource": { "type": "model", "id": "HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1", "discussionNum": null }, "url": "https://huggingface.co/HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Since ORPO doesn't require a separate SFT stage, all that is needed is a strong base model + high-quality DPO-style datasets.", "raw": "Since ORPO doesn't require a separate SFT stage, all that is needed is a strong base model + high-quality DPO-style datasets.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Currently, there is a significant lack of non-English DPO datasets. Filling this gap could significantly improve open LLMs in various languages.", "raw": "Currently, there is a significant lack of non-English DPO datasets. Filling this gap could significantly improve open LLMs in various languages.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "You can get an overview of the current state of DPO datasets across different languages here: ", "raw": "You can get an overview of the current state of DPO datasets across different languages here: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/spaces/DIBT/preference_data_by_language", "href": "https://huggingface.co/spaces/DIBT/preference_data_by_language", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/60f0608166e5701b80ed3f02/BHso-wSWpR9b8b8CKvodC.jpeg", "fullname": "Alvaro Bartolome", "name": "alvarobartt", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1739 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6415c043486c7c9a5d151583/fUdYFh6iVh57swCkBEy-y.jpeg", "fullname": "Jiwoo Hong", "name": "JW17", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 11 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1594651707950-noauth.jpeg", "fullname": "Lewis Tunstall", "name": "lewtun", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 678 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6296fe6f7d586f66c3957f33/lCHe1mBVRzQ1sKsXD2OUd.jpeg", "fullname": "Noah Lee", "name": "nlee-208", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 7 } ]

[ { "reaction": "🚀", "users": [ "JW17", "alvarobartt", "osanseviero", "nlee-208", "NMontanaBrown", "Violette", "Heng666", "raincandy-u", "DataSoul" ], "count": 9 } ]

[]

[ { "type": "text", "value": "On Agentic AI: Autonomy Is All You Need!", "raw": "On Agentic AI: Autonomy Is All You Need!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "There is this remarkable beauty in witnessing an AI system autonomously completes complex tasks with a level of brilliance that supersedes our reasoning capabilities/expectations - It is the holy grail of creation.", "raw": "There is this remarkable beauty in witnessing an AI system autonomously completes complex tasks with a level of brilliance that supersedes our reasoning capabilities/expectations - It is the holy grail of creation.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Giving your AI agents autonomy is analogous to us having \"free will\" and everything else thereafter is a cascade of possibilities and potentials waiting to unfold.", "raw": "Giving your AI agents autonomy is analogous to us having \"free will\" and everything else thereafter is a cascade of possibilities and potentials waiting to unfold.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "As brilliantly said by ", "raw": "As brilliantly said by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@AndrewNg", "href": null, "resource": null, "url": null, "code": null, "user": "AndrewNg", "label": null, "lang": null }, { "type": "text", "value": " \"It’s a beautiful thing when you see an agent autonomously decide to do things in ways that you had not anticipated, and succeed as a result!\"", "raw": " \"It’s a beautiful thing when you see an agent autonomously decide to do things in ways that you had not anticipated, and succeed as a result!\"", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Autonomy in Agentic AI", "raw": "Autonomy in Agentic AI", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " - augments agents' decision-making capabilities.", "raw": " - augments agents' decision-making capabilities.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " - enables adaptation to diverse environments.", "raw": " - enables adaptation to diverse environments.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " - facilitates real-time learning and improvement.", "raw": " - facilitates real-time learning and improvement.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " - fosters dynamic multi-agent collaboration.", "raw": " - fosters dynamic multi-agent collaboration.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " - promotes efficient and independent task execution.", "raw": " - promotes efficient and independent task execution.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " - drives innovation in dynamic and unpredictable ", "raw": " - drives innovation in dynamic and unpredictable ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " scenarios.", "raw": " scenarios.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This what the AutoAgents paper conveyed - A fully autonomous agentic framework that basically gives agents the free will for authentic/compelling creativity. ", "raw": "This what the AutoAgents paper conveyed - A fully autonomous agentic framework that basically gives agents the free will for authentic/compelling creativity. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "With just three dynamically predefined agents [ Planner, Agent Observer and Plan Observer ] acting collaboratively they are able to generically create autonomous agents:", "raw": "With just three dynamically predefined agents [ Planner, Agent Observer and Plan Observer ] acting collaboratively they are able to generically create autonomous agents:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "code_fence", "value": null, "raw": "```\nAgent (A) = { \n Prompt (P) - defines agent's identity fully , \n Description (D) - adds specific role identity, \n Toolset (T) - equips the agent with tools, \n Suggestion (S) - offers task execution tips\n }\n```", "href": null, "resource": null, "url": null, "code": "Agent (A) = { \n Prompt (P) - defines agent's identity fully , \n Description (D) - adds specific role identity, \n Toolset (T) - equips the agent with tools, \n Suggestion (S) - offers task execution tips\n }", "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Demo: ", "raw": "Demo: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/LinkSoul/AutoAgents", "href": null, "resource": { "type": "space", "id": "LinkSoul/AutoAgents", "discussionNum": null }, "url": "https://huggingface.co/spaces/LinkSoul/AutoAgents", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Code: ", "raw": "Code: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/Link-AGI/AutoAgents", "href": "https://github.com/Link-AGI/AutoAgents", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Paper: ", "raw": "Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://arxiv.org/abs/2309.17288", "href": "https://arxiv.org/abs/2309.17288", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/-pSkN00wmzwIGpK9cp3mS.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/2FFSflMjzZK-LMDZhRFeX.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/GkMPi4BS72CZPLpI7L4-7.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/oBzDMxBC_CRKBpIi8uXeq.jpeg" } ]

[ { "avatarUrl": "/avatars/c8f08de4450d521e9f8de5fc6d594432.svg", "fullname": "Andrew Ng", "name": "AndrewNg", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3 } ]

[ { "reaction": "🔥", "users": [ "owtx", "osanseviero", "victor", "NMontanaBrown" ], "count": 4 } ]

[]

[ { "type": "text", "value": "Fun dataset added last week by ", "raw": "Fun dataset added last week by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@esind", "href": null, "resource": null, "url": null, "code": null, "user": "esind", "label": null, "lang": null }, { "type": "text", "value": " from ", "raw": " from ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/Anthropic", "href": "https://huggingface.co/Anthropic", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " to compare persuasiveness between AI and human outputs: ", "raw": " to compare persuasiveness between AI and human outputs: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/Anthropic/persuasion", "href": null, "resource": { "type": "dataset", "id": "Anthropic/persuasion", "discussionNum": null }, "url": "https://huggingface.co/datasets/Anthropic/persuasion", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e67bdd61009063689407479/77UIqMwhFodTRi35sy5EP.png" } ]

[ { "avatarUrl": "/avatars/c515e502a45156b28d845af6033ad38c.svg", "fullname": "Esin Durmus", "name": "esind", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 8 } ]

[ { "reaction": "🚀", "users": [ "Jaward", "kgourgou", "osanseviero", "not-lain", "KvrParaskevi", "InferenceIllusionist", "nogully" ], "count": 7 }, { "reaction": "👍", "users": [ "pierrci", "not-lain" ], "count": 2 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6438a9027de34e8ea7e4b257/vib8QSd1AWMr_bR9ig_xJ.jpeg", "fullname": "Jaward Sesay", "name": "Jaward", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 191, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6032802e1f993496bc14d9e3/w6hr-DEQot4VVkoyRIBiy.png", "fullname": "Omar Sanseviero", "name": "osanseviero", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2868, "isFollowing": false } ]

[ { "type": "text", "value": "We're introducing experimental support for ", "raw": "We're introducing experimental support for ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "inline_code", "value": null, "raw": "`device_map`", "href": null, "resource": null, "url": null, "code": "device_map", "user": null, "label": null, "lang": null }, { "type": "text", "value": " in Diffusers 🤗", "raw": " in Diffusers 🤗", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "If you have multiple GPUs you want to use to distribute the pipeline models, you can do so. Additionally, this becomes more useful when you have multiple low-VRAM GPUs.", "raw": "If you have multiple GPUs you want to use to distribute the pipeline models, you can do so. Additionally, this becomes more useful when you have multiple low-VRAM GPUs.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Documentation:", "raw": "Documentation:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/docs/diffusers/main/en/training/distributed_inference#device-placement", "href": "https://huggingface.co/docs/diffusers/main/en/training/distributed_inference#device-placement", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚨 Currently, only \"balanced\" device mapping strategy is supported. ", "raw": "🚨 Currently, only \"balanced\" device mapping strategy is supported. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5f7fbd813e94f16a85448745/xfC4AkW9bWPeLskzsoue9.png" } ]

[]

[ { "reaction": "🔥", "users": [ "ajibawa-2023", "clem", "osanseviero", "Bruno", "Paolo16", "alvarobartt", "ingTikna" ], "count": 7 }, { "reaction": "🤗", "users": [ "alvarobartt" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Watch the full tutorial here : ", "raw": "Watch the full tutorial here : ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://youtu.be/0t5l6CP9eBg", "href": "https://youtu.be/0t5l6CP9eBg", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The tutorial is over 2 hours literally with manually fixed captions and perfect video chapters.", "raw": "The tutorial is over 2 hours literally with manually fixed captions and perfect video chapters.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Most Awaited Full Fine Tuning (with DreamBooth effect) Tutorial Generated Images - Full Workflow Shared In The Comments - NO Paywall This Time - Explained OneTrainer - Cumulative Experience of 16 Months Stable Diffusion", "raw": "Most Awaited Full Fine Tuning (with DreamBooth effect) Tutorial Generated Images - Full Workflow Shared In The Comments - NO Paywall This Time - Explained OneTrainer - Cumulative Experience of 16 Months Stable Diffusion", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "In this tutorial, I am going to show you how to install OneTrainer from scratch on your computer and do a Stable Diffusion SDXL (Full Fine-Tuning 10.3 GB VRAM) and SD 1.5 (Full Fine-Tuning 7GB VRAM) based models training on your computer and also do the same training on a very cheap cloud machine from MassedCompute if you don't have such computer.", "raw": "In this tutorial, I am going to show you how to install OneTrainer from scratch on your computer and do a Stable Diffusion SDXL (Full Fine-Tuning 10.3 GB VRAM) and SD 1.5 (Full Fine-Tuning 7GB VRAM) based models training on your computer and also do the same training on a very cheap cloud machine from MassedCompute if you don't have such computer.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Tutorial Readme File ⤵️", "raw": "Tutorial Readme File ⤵️", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Tutorials/OneTrainer-Master-SD-1_5-SDXL-Windows-Cloud-Tutorial.md", "href": "https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Tutorials/OneTrainer-Master-SD-1_5-SDXL-Windows-Cloud-Tutorial.md", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Register Massed Compute From Below Link (could be necessary to use our Special Coupon for A6000 GPU for 31 cents per hour) ⤵️", "raw": "Register Massed Compute From Below Link (could be necessary to use our Special Coupon for A6000 GPU for 31 cents per hour) ⤵️", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://bit.ly/Furkan-Gözükara", "href": "https://bit.ly/Furkan-Gözükara", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Coupon Code for A6000 GPU is : SECourses", "raw": "Coupon Code for A6000 GPU is : SECourses", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "0:00 Introduction to Zero-to-Hero Stable Diffusion (SD) Fine-Tuning with OneTrainer (OT) tutorial", "raw": "0:00 Introduction to Zero-to-Hero Stable Diffusion (SD) Fine-Tuning with OneTrainer (OT) tutorial", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3:54 Intro to instructions GitHub readme", "raw": "3:54 Intro to instructions GitHub readme", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "4:32 How to register Massed Compute (MC) and start virtual machine (VM)", "raw": "4:32 How to register Massed Compute (MC) and start virtual machine (VM)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "5:48 Which template to choose on MC", "raw": "5:48 Which template to choose on MC", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "6:36 How to apply MC coupon", "raw": "6:36 How to apply MC coupon", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "8:41 How to install OT on your computer to train", "raw": "8:41 How to install OT on your computer to train", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "9:15 How to verify your Python, Git, FFmpeg and Git installation", "raw": "9:15 How to verify your Python, Git, FFmpeg and Git installation", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "12:00 How to install ThinLinc and start using your MC VM", "raw": "12:00 How to install ThinLinc and start using your MC VM", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "12:26 How to setup folder synchronization and file sharing between your computer and MC VM", "raw": "12:26 How to setup folder synchronization and file sharing between your computer and MC VM", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "13:56 End existing session in ThinClient", "raw": "13:56 End existing session in ThinClient", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "14:06 How to turn off MC VM", "raw": "14:06 How to turn off MC VM", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "14:24 How to connect and start using VM", "raw": "14:24 How to connect and start using VM", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "14:41 When use end existing session", "raw": "14:41 When use end existing session", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "16:38 How to download very best OT preset training configuration for SD 1.5 & SDXL models", "raw": "16:38 How to download very best OT preset training configuration for SD 1.5 & SDXL models", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "18:00 How to load configuration preset", "raw": "18:00 How to load configuration preset", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "18:38 Full explanation of OT configuration and best hyper parameters for SDXL", "raw": "18:38 Full explanation of OT configuration and best hyper parameters for SDXL", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ".", "raw": ".", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ".", "raw": ".", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ".", "raw": ".", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/k-5rEiLeDNt1t_WnlQRrT.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/JZrg_newFz497OBa9rkNN.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/8ocDtPIpZXUMwBrCCHiH2.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/VbvDH12JZyozDzS2suZS_.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/iOOCTnjxBvJzEr72evb-B.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/j53cJRPk5514Sgq73dt9l.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/xz-KXTEEtosaGlpfcO9kx.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/n2xy4t7z9mqwcR7cZvKkk.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/HJLT8leUxpbpRt-K9bo9w.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/Dn7PsCc1OmuFCEoIjm_D0.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/2tLcWMkMUpJCBc14G9Gn-.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/DwcaX7BDaeiLTF42oUT6i.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/tn-LZghS00xUhfGXglXKX.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/Ah5CQDVVNZ8geQeRn96Ap.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/1qlaRmJCIAacNKVdu9kQw.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/o9mqtw0iFSP6yaebcixTQ.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/ayHRjnxDa_PCxiDSolFW6.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/zcCuGcoBJE9P4OLoFQ1En.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/xXtzTTABOiSjlRNxQ7s8p.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6345bd89fe134dfd7a0dba40/HpnRecxUiIh_BxWaSY4GD.png" } ]

[]

[ { "reaction": "🔥", "users": [ "rafalwiatrowski", "mehdicode123", "samusenps", "kadirnar", "HeavyLvy", "clem", "Aakashk", "Matix200", "AlexSonar", "levulinh" ], "count": 10 }, { "reaction": "❤️", "users": [ "kadirnar", "AlexSonar", "levulinh" ], "count": 3 } ]

[]

[ { "type": "text", "value": "Here's a utility component for integrating your Gradio app with Hugging Face. This custom component enables you to search for models, spaces, datasets, and users. ", "raw": "Here's a utility component for integrating your Gradio app with Hugging Face. This custom component enables you to search for models, spaces, datasets, and users. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "inline_code", "value": null, "raw": "`pip install gradio_huggingfacehub_search`", "href": null, "resource": null, "url": null, "code": "pip install gradio_huggingfacehub_search", "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "You can see it in action here. ", "raw": "You can see it in action here. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/arcee-ai/mergekit-config-generator", "href": null, "resource": { "type": "space", "id": "arcee-ai/mergekit-config-generator", "discussionNum": null }, "url": "https://huggingface.co/spaces/arcee-ai/mergekit-config-generator", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "And learn how to use it here ", "raw": "And learn how to use it here ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/radames/gradio_huggingfacehub_search", "href": null, "resource": { "type": "space", "id": "radames/gradio_huggingfacehub_search", "discussionNum": null }, "url": "https://huggingface.co/spaces/radames/gradio_huggingfacehub_search", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6064e095abd8d3692e3e2ed6/D6NSplfrz7YHb_XfuBlOi.mp4" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "osanseviero", "srisree", "clem", "Gevennou", "multimodalart" ], "count": 6 }, { "reaction": "➕", "users": [ "samusenps", "clem", "multimodalart" ], "count": 3 }, { "reaction": "🔥", "users": [ "osanseviero", "clem", "multimodalart" ], "count": 3 } ]

[]

[ { "type": "text", "value": "\"How expensive is it actually to teach a #LanguageModel German through #finetuning 💰💰💰? We get asked this quite often.", "raw": "\"How expensive is it actually to teach a #LanguageModel German through #finetuning 💰💰💰? We get asked this quite often.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "There is no one-size-fits-all answer to this question, as among other factors:", "raw": "There is no one-size-fits-all answer to this question, as among other factors:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "⏹ each fine-tuning is different,", "raw": "⏹ each fine-tuning is different,", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "⏹ the hardware used can be a major cost driver,", "raw": "⏹ the hardware used can be a major cost driver,", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "⏹ the amount and type of training data can extend the process,", "raw": "⏹ the amount and type of training data can extend the process,", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "⏹ and the skills to be trained can increase the difficulty of fine-tuning.", "raw": "⏹ and the skills to be trained can increase the difficulty of fine-tuning.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "However, we have broken down the costs incurred for our latest fine-tune ( ", "raw": "However, we have broken down the costs incurred for our latest fine-tune ( ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/VAGOsolutions/SauerkrautLM-Qwen-32b", "href": null, "resource": { "type": "model", "id": "VAGOsolutions/SauerkrautLM-Qwen-32b", "discussionNum": null }, "url": "https://huggingface.co/VAGOsolutions/SauerkrautLM-Qwen-32b", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ")", "raw": ")", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Base model: ", "raw": "Base model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/Qwen/Qwen1.5-32B", "href": null, "resource": { "type": "model", "id": "Qwen/Qwen1.5-32B", "discussionNum": null }, "url": "https://huggingface.co/Qwen/Qwen1.5-32B", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Fine-Tuning Goal: Train German language", "raw": "Fine-Tuning Goal: Train German language", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Training dataset size: 160,000 SFT data / 110,000 DPO data", "raw": "Training dataset size: 160,000 SFT data / 110,000 DPO data", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Training duration: 72.5 hours (2 epochs SFT / 1 epoch DPO)", "raw": "Training duration: 72.5 hours (2 epochs SFT / 1 epoch DPO)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "GPU: 2x A100 SXM", "raw": "GPU: 2x A100 SXM", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "New model: ", "raw": "New model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/VAGOsolutions/SauerkrautLM-Qwen-32b", "href": null, "resource": { "type": "model", "id": "VAGOsolutions/SauerkrautLM-Qwen-32b", "discussionNum": null }, "url": "https://huggingface.co/VAGOsolutions/SauerkrautLM-Qwen-32b", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Total cost: 312 euros 💶", "raw": "Total cost: 312 euros 💶", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "These are quite reasonable training costs considering the model now speaks passable German (previously very broken). Depending on the use case and process requirements, this can even be a real alternative to the costly continuous pre-training of foreign language models.", "raw": "These are quite reasonable training costs considering the model now speaks passable German (previously very broken). Depending on the use case and process requirements, this can even be a real alternative to the costly continuous pre-training of foreign language models.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🔥", "users": [ "jed-tiotuico", "osanseviero", "clem", "victor", "fpreiss", "philipp-zettl", "joennlae", "Kearm" ], "count": 8 } ]

[]

[ { "type": "text", "value": "Yesterday, xAI announced Grok-1.5 Vision - ", "raw": "Yesterday, xAI announced Grok-1.5 Vision - ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://x.ai/blog/grok-1.5v", "href": "https://x.ai/blog/grok-1.5v", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ". But more importantly, they also released a new VLM benchmark dataset - RealWorldQA. The only problem was that they released it as a ZIP archive. I fixed that! Now you can use it in your evaluations as a regular HF dataset: ", "raw": ". But more importantly, they also released a new VLM benchmark dataset - RealWorldQA. The only problem was that they released it as a ZIP archive. I fixed that! Now you can use it in your evaluations as a regular HF dataset: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/visheratin/realworldqa", "href": null, "resource": { "type": "dataset", "id": "visheratin/realworldqa", "discussionNum": null }, "url": "https://huggingface.co/datasets/visheratin/realworldqa", "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "👍", "users": [ "musfiqdehan", "fynnkroeger", "osanseviero", "Bruno", "ajibawa-2023", "VictorSanh", "clem", "not-lain", "ShihHong", "alielfilali01", "dinhanhx", "jyoung105" ], "count": 12 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1583857146757-5e67bdd61009063689407479.jpeg", "fullname": "Clem 🤗", "name": "clem", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 1763, "isFollowing": false } ]

[ { "type": "text", "value": "Introducing MusicGen Web: AI-powered music generation directly in your browser, built with 🤗 Transformers.js! 🎵", "raw": "Introducing MusicGen Web: AI-powered music generation directly in your browser, built with 🤗 Transformers.js! 🎵", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Everything runs 100% locally, meaning there are no calls to an API! 🤯 Since it's served as a static HF space, it costs $0 to host and run! 🔥", "raw": "Everything runs 100% locally, meaning there are no calls to an API! 🤯 Since it's served as a static HF space, it costs $0 to host and run! 🔥", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We also added the ability to share your generated music to the discussion tab, so give it a try! 👇", "raw": "We also added the ability to share your generated music to the discussion tab, so give it a try! 👇", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/Xenova/musicgen-web", "href": null, "resource": { "type": "space", "id": "Xenova/musicgen-web", "discussionNum": null }, "url": "https://huggingface.co/spaces/Xenova/musicgen-web", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/61b253b7ac5ecaae3d1efe0c/xvWuxGhP6T_hTnRKvrRxf.mp4" } ]

[]

[ { "reaction": "🔥", "users": [ "DmitryRyumin", "adamelliotfields", "samusenps", "andrewmilesmurphy1992", "Lyte", "rocca", "Sword666", "ajibawa-2023", "ashwinlimaye", "Retr0REG", "cdosoftei", "Cksampa", "osanseviero", "clem", "AlekseiPravdin", "victor", "ylacombe", "alt2023", "DataSoul", "b4s36t4", "Noomam", "nekomeowww" ], "count": 22 }, { "reaction": "😎", "users": [ "luoweiqiang9527", "cdosoftei", "raincandy-u", "Cksampa", "osanseviero", "clem", "victor", "Kce", "niwram69" ], "count": 9 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64b3eb9b93679cd0094dc95d/eBX-78aOs6cVencjM-3ME.png", "fullname": "DataSoul", "name": "DataSoul", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/61b253b7ac5ecaae3d1efe0c/hwiQ0uvz3t-L5a-NtBIO6.png", "fullname": "Joshua", "name": "Xenova", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 3792, "isFollowing": false } ]

[ { "type": "text", "value": "Spent the morning designing a super important product: SMALL LANGUAGE MODEL BABY ONESIE ", "raw": "Spent the morning designing a super important product: SMALL LANGUAGE MODEL BABY ONESIE ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://www.shop.hackernoon.com/all-merch/p/small-language-model-baby-onesie", "href": "https://www.shop.hackernoon.com/all-merch/p/small-language-model-baby-onesie", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/64862a25cf5ad5e1f0482ef2/neXR9MekpTqT8Be9LgV0o.png" } ]

[]

[ { "reaction": "🔥", "users": [ "samusenps", "visheratin", "travisking", "Violette" ], "count": 4 }, { "reaction": "👀", "users": [ "samusenps" ], "count": 1 }, { "reaction": "➕", "users": [ "samusenps" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6538119803519fddb4a17e10/ffJMkdx-rM7VvLTCM6ri_.jpeg", "fullname": "samusenps", "name": "samusenps", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 91, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/657eb5b256c9c67605a6e8b5/RPblnGJX57oiIcASEz_S8.png", "fullname": "raincandy_U", "name": "raincandy-u", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 30, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64862a25cf5ad5e1f0482ef2/61qPUtw9jIl7zpPYmi0VW.jpeg", "fullname": "David Smooke", "name": "Smooke", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 43, "isFollowing": false } ]

[ { "type": "text", "value": "Hi everyone! I lead a team of students to create a game board community of LLMs that interact to help one of the LLMs create a presentation on an idea. ", "raw": "Hi everyone! I lead a team of students to create a game board community of LLMs that interact to help one of the LLMs create a presentation on an idea. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/GooseCube/WWPD.ai", "href": "https://github.com/GooseCube/WWPD.ai", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This was based on Standfords unusable initial version: ", "raw": "This was based on Standfords unusable initial version: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "\"A group of researchers at Stanford University and Google have created a miniature RPG-style virtual world similar to The Sims, where 25 characters, controlled by ChatGPT and custom code, live out their lives independently with a high degree of realistic behavior. They wrote about their experiment in a preprint academic paper released on Friday.\"", "raw": "\"A group of researchers at Stanford University and Google have created a miniature RPG-style virtual world similar to The Sims, where 25 characters, controlled by ChatGPT and custom code, live out their lives independently with a high degree of realistic behavior. They wrote about their experiment in a preprint academic paper released on Friday.\"", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We've come so far on this open source project, however the class that developed it is now finished. The UI needs a work over, but the giant success is that this is a LIVE version of the Stanfords version. We're super close to finishing something that is super cool. ", "raw": "We've come so far on this open source project, however the class that developed it is now finished. The UI needs a work over, but the giant success is that this is a LIVE version of the Stanfords version. We're super close to finishing something that is super cool. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Check it out at: ", "raw": "Check it out at: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://wwpd-ai.vercel.app/", "href": "https://wwpd-ai.vercel.app/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I'm hoping that we can get some help from the community to help bring it to a place where the LLM gives a ted talk on whatever a topic that the user chooses. ", "raw": "I'm hoping that we can get some help from the community to help bring it to a place where the LLM gives a ted talk on whatever a topic that the user chooses. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Username and pass: ", "raw": "Username and pass: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "[email protected]", "raw": "[email protected]", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "12345678", "raw": "12345678", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/65f46f95d64a62e375c5d31d/zSMfoFJNglSA3DoyN5P7n.png" } ]

[]

[ { "reaction": "😎", "users": [ "raincandy-u", "DRRRRDFFD", "osanseviero", "oceansweep", "clem", "abidlabs" ], "count": 6 }, { "reaction": "🚀", "users": [ "Kukedlc", "oceansweep", "abidlabs", "kramp" ], "count": 4 }, { "reaction": "🧠", "users": [ "Kukedlc", "oceansweep", "abidlabs" ], "count": 3 }, { "reaction": "🔥", "users": [ "levulinh" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/5f17f0a0925b9863e28ad517/X7QKoiXbUtEZSG9jyvfk3.jpeg", "fullname": "Victor Mustar", "name": "victor", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 2607, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/65f46f95d64a62e375c5d31d/or1-4C8fixlfIWpCXNBW4.png", "fullname": "Steven Ochs", "name": "theStevenochs", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 9, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/657eb5b256c9c67605a6e8b5/RPblnGJX57oiIcASEz_S8.png", "fullname": "raincandy_U", "name": "raincandy-u", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 30, "isFollowing": false }, { "avatarUrl": "/avatars/735b678bd5863a0c1b1bdd3bbf8858fa.svg", "fullname": "r", "name": "oceansweep", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 7, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/65aafba82ed95c799f4f23cf/ti4NGr_1Ltg9EcwlE-mf-.png", "fullname": "Oleh Puhaiev", "name": "websymbol", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "Hello everyone, looking to share and learn more most of my experience is security so I will be doing more learning in order to secure AI", "raw": "Hello everyone, looking to share and learn more most of my experience is security so I will be doing more learning in order to secure AI", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🚀", "users": [ "adaboost19", "raincandy-u", "nachoyawn", "sbrandeis" ], "count": 4 } ]

[]

[ { "type": "text", "value": "Hey everyone! 👋 ", "raw": "Hey everyone! 👋 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This is my first post here and I’m super excited to start with not just one, but two awesome updates! 🚀", "raw": "This is my first post here and I’m super excited to start with not just one, but two awesome updates! 🚀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Some of you might already know that I recently started my internship at Hugging Face. I’m grateful to be a part of the LLMs evaluation team and the Open LLM Leaderboard! 🤗", "raw": "Some of you might already know that I recently started my internship at Hugging Face. I’m grateful to be a part of the LLMs evaluation team and the Open LLM Leaderboard! 🤗", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "First up, we’ve got some big news: we’ve just completed the evaluations for the ", "raw": "First up, we’ve got some big news: we’ve just completed the evaluations for the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/mistral-community/Mixtral-8x22B-v0.1", "href": null, "resource": { "type": "model", "id": "mistral-community/Mixtral-8x22B-v0.1", "discussionNum": null }, "url": "https://huggingface.co/mistral-community/Mixtral-8x22B-v0.1", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ", and guess what? It’s now the top-performing pretrained model on the Open LLM Leaderboard! A huge shoutout to Mistral! 🏆👏 You can see more details and check out the evaluation results right here – ", "raw": ", and guess what? It’s now the top-performing pretrained model on the Open LLM Leaderboard! A huge shoutout to Mistral! 🏆👏 You can see more details and check out the evaluation results right here – ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/datasets/open-llm-leaderboard/details_mistral-community__Mixtral-8x22B-v0.1", "href": "https://huggingface.co/datasets/open-llm-leaderboard/details_mistral-community__Mixtral-8x22B-v0.1", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Next, I’m excited to share a cool new feature – you can now search for models on the Open LLM Leaderboard by their licenses! 🕵️‍♂️ This feature will help you find the perfect model for your projects way faster. Just type \"license: MIT\" as a test run!", "raw": "Next, I’m excited to share a cool new feature – you can now search for models on the Open LLM Leaderboard by their licenses! 🕵️‍♂️ This feature will help you find the perfect model for your projects way faster. Just type \"license: MIT\" as a test run!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I'd be super happy if you'd follow me here for more updates on the Leaderboard and other exciting developments. Can’t wait to share more with you soon! ✨ ", "raw": "I'd be super happy if you'd follow me here for more updates on the Leaderboard and other exciting developments. Can’t wait to share more with you soon! ✨ ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/63f5010dfcf95ecac2ad8652/I5u8gdEv32SZQKXjT5ZiP.png" } ]

[]

[ { "reaction": "❤️", "users": [ "clefourrier", "julien-c", "osanseviero", "santiagokomadina", "DmitryRyumin", "victor", "jkepler81", "samusenps", "sugatoray", "clem", "oceansweep" ], "count": 11 }, { "reaction": "🤗", "users": [ "DmitryRyumin", "clem", "oceansweep" ], "count": 3 }, { "reaction": "🚀", "users": [ "Felladrin", "pwab" ], "count": 2 } ]

[]

[ { "type": "text", "value": "🕺🎬🔥 New Research Alert - CVPR 2024 (Avatars Collection)! 🔥🤖⚡", "raw": "🕺🎬🔥 New Research Alert - CVPR 2024 (Avatars Collection)! 🔥🤖⚡", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Title: GoMAvatar: Efficient Animatable Human Modeling from Monocular Video Using Gaussians-on-Mesh", "raw": "📄 Title: GoMAvatar: Efficient Animatable Human Modeling from Monocular Video Using Gaussians-on-Mesh", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📝 Description: GoMAvatar is an efficient method for real-time, high-quality, animatable human modeling from a single monocular video. It combines the rendering quality of Gaussian splatting with the geometry modeling capabilities of deformable meshes, enabling realistic digital avatars that can be rearticulated in new poses and rendered from novel angles, while seamlessly integrating with graphics pipelines.", "raw": "📝 Description: GoMAvatar is an efficient method for real-time, high-quality, animatable human modeling from a single monocular video. It combines the rendering quality of Gaussian splatting with the geometry modeling capabilities of deformable meshes, enabling realistic digital avatars that can be rearticulated in new poses and rendered from novel angles, while seamlessly integrating with graphics pipelines.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👥 Authors: Jing Wen, Xiaoming Zhao, Zhongzheng Ren, Alexander G. Schwing, Shenlong Wang", "raw": "👥 Authors: Jing Wen, Xiaoming Zhao, Zhongzheng Ren, Alexander G. Schwing, Shenlong Wang", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📅 Conference: CVPR, Jun 17-21, 2024 | Seattle WA, USA 🇺🇸", "raw": "📅 Conference: CVPR, Jun 17-21, 2024 | Seattle WA, USA 🇺🇸", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔗 Paper: ", "raw": "🔗 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.07991", "href": null, "resource": { "type": "paper", "id": "2404.07991", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.07991", "code": null, "user": null, "label": "GoMAvatar: Efficient Animatable Human Modeling from Monocular Video\n Using Gaussians-on-Mesh (2404.07991)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌐 Github Page: ", "raw": "🌐 Github Page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://wenj.github.io/GoMAvatar/", "href": "https://wenj.github.io/GoMAvatar/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📁 Repository: ", "raw": "📁 Repository: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/wenj/GoMAvatar", "href": "https://github.com/wenj/GoMAvatar", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "raw": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "href": null, "resource": { "type": "space", "id": "DmitryRyumin/NewEraAI-Papers", "discussionNum": null }, "url": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " curated by ", "raw": " curated by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DmitryRyumin", "href": null, "resource": null, "url": null, "code": null, "user": "DmitryRyumin", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Added to the Avatars Collection: ", "raw": "🚀 Added to the Avatars Collection: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "href": null, "resource": { "type": "collection", "id": "DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "discussionNum": null }, "url": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Keywords: #GoMAvatar #3DAvatar #3DAnimation #AnimatableAvatars #MonocularVideo #RealTimeRendering #HumanModeling #CVPR2024 #DeepLearning #Innovation", "raw": "🔍 Keywords: #GoMAvatar #3DAvatar #3DAnimation #AnimatableAvatars #MonocularVideo #RealTimeRendering #HumanModeling #CVPR2024 #DeepLearning #Innovation", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/QrQRKYgqUYrtK1tlffGzb.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/tdiWs3pYbcMXb0FOy1qMi.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/yvv0FJZ7G92ikEfY6XkSp.gif" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/xKqMHnANG1KMDHcQaEplP.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/pM0ddjuQKFfadjBV2GG4P.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/nRCxbVng_PPBqKd-Z3KVc.jpeg", "fullname": "Dmitry Ryumin", "name": "DmitryRyumin", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 377 } ]

[ { "reaction": "👍", "users": [ "DmitryRyumin", "osanseviero", "clem", "abbott" ], "count": 4 }, { "reaction": "❤️", "users": [ "samusenps", "clem" ], "count": 2 } ]

[]

[ { "type": "text", "value": "Introducing Zephyr 141B-A35B 🪁: ", "raw": "Introducing Zephyr 141B-A35B 🪁: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1", "href": null, "resource": { "type": "model", "id": "HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1", "discussionNum": null }, "url": "https://huggingface.co/HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Yesterday, Mistral released their latest base model (via magnet link of course 😅) and the community quickly converted it to transformers format and pushed it to the Hub: ", "raw": "Yesterday, Mistral released their latest base model (via magnet link of course 😅) and the community quickly converted it to transformers format and pushed it to the Hub: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/mistral-community/Mixtral-8x22B-v0.1", "href": null, "resource": { "type": "model", "id": "mistral-community/Mixtral-8x22B-v0.1", "discussionNum": null }, "url": "https://huggingface.co/mistral-community/Mixtral-8x22B-v0.1", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Early evals of this model looked extremely strong, so we teamed up with Argilla and KAIST AI to cook up a Zephyr recipe with a few new alignment techniques that came out recently:", "raw": "Early evals of this model looked extremely strong, so we teamed up with Argilla and KAIST AI to cook up a Zephyr recipe with a few new alignment techniques that came out recently:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🧑‍🍳 Align the base model with Odds Ratio Preference Optimisation (ORPO). This novel algorithm developed by ", "raw": "🧑‍🍳 Align the base model with Odds Ratio Preference Optimisation (ORPO). This novel algorithm developed by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@JW17", "href": null, "resource": null, "url": null, "code": null, "user": "JW17", "label": null, "lang": null }, { "type": "text", "value": " and ", "raw": " and ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@nlee-208", "href": null, "resource": null, "url": null, "code": null, "user": "nlee-208", "label": null, "lang": null }, { "type": "text", "value": " and ", "raw": " and ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@j6mes", "href": null, "resource": null, "url": null, "code": null, "user": "j6mes", "label": null, "lang": null }, { "type": "text", "value": " and does not require an SFT step to achieve high performance and is thus much more computationally efficient than methods like DPO and PPO. ", "raw": " and does not require an SFT step to achieve high performance and is thus much more computationally efficient than methods like DPO and PPO. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🦫 Use a brand new dataset of 7k high-quality, multi-turn preferences that has been developed by our friends at Argilla. To create this dataset, they took the excellent Capybara SFT dataset from ", "raw": "🦫 Use a brand new dataset of 7k high-quality, multi-turn preferences that has been developed by our friends at Argilla. To create this dataset, they took the excellent Capybara SFT dataset from ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@LDJnr", "href": null, "resource": null, "url": null, "code": null, "user": "LDJnr", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/LDJnr/Capybara", "href": null, "resource": { "type": "dataset", "id": "LDJnr/Capybara", "discussionNum": null }, "url": "https://huggingface.co/datasets/LDJnr/Capybara", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " and converted it into a preference dataset by augmenting the final turn with responses from new LLMs that were then ranked by GPT-4.", "raw": " and converted it into a preference dataset by augmenting the final turn with responses from new LLMs that were then ranked by GPT-4.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "What we find especially neat about this approach is that training on 7k samples only takes ~1.3h on 4 H100 nodes, yet produces a model that is very strong on chat benchmarks like IFEval and BBH.", "raw": "What we find especially neat about this approach is that training on 7k samples only takes ~1.3h on 4 H100 nodes, yet produces a model that is very strong on chat benchmarks like IFEval and BBH.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Kudos to ", "raw": "Kudos to ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@alvarobartt", "href": null, "resource": null, "url": null, "code": null, "user": "alvarobartt", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@JW17", "href": null, "resource": null, "url": null, "code": null, "user": "JW17", "label": null, "lang": null }, { "type": "text", "value": " and ", "raw": " and ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@nlee-208", "href": null, "resource": null, "url": null, "code": null, "user": "nlee-208", "label": null, "lang": null }, { "type": "text", "value": " for this very nice and fast-paced collab!", "raw": " for this very nice and fast-paced collab!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "For more details on the paper and dataset, checkout our collection: ", "raw": "For more details on the paper and dataset, checkout our collection: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/HuggingFaceH4/zephyr-orpo-6617eba2c5c0e2cc3c151524", "href": null, "resource": { "type": "collection", "id": "HuggingFaceH4/zephyr-orpo-6617eba2c5c0e2cc3c151524", "discussionNum": null }, "url": "https://huggingface.co/collections/HuggingFaceH4/zephyr-orpo-6617eba2c5c0e2cc3c151524", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5f0c746619cb630495b814fd/FlnsqNXJIX4qhxvFkko09.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/60f0608166e5701b80ed3f02/BHso-wSWpR9b8b8CKvodC.jpeg", "fullname": "Alvaro Bartolome", "name": "alvarobartt", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1739 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/63b54db5889aa6707f08f2d1/zUnyx-8Zabxr54poZtUo4.jpeg", "fullname": "James Thorne", "name": "j6mes", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 6 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6415c043486c7c9a5d151583/fUdYFh6iVh57swCkBEy-y.jpeg", "fullname": "Jiwoo Hong", "name": "JW17", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 11 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/2GjRE8z9lYNeelnBTz5ym.png", "fullname": "Luigi D", "name": "LDJnr", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 125 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6296fe6f7d586f66c3957f33/lCHe1mBVRzQ1sKsXD2OUd.jpeg", "fullname": "Noah Lee", "name": "nlee-208", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 7 } ]

[ { "reaction": "❤️", "users": [ "alvarobartt", "mehungrymunchie", "nlee-208", "SilentWraith", "JW17", "samusenps", "hakunamatata1997", "dakshdeephere", "NouRed", "victor", "osanseviero", "avinash02", "fynnkroeger", "dillfrescott", "clem", "philschmid", "srisree", "rafaelsandroni", "not-lain", "netynet" ], "count": 20 }, { "reaction": "🤗", "users": [ "beta3", "clem", "SilentWraith", "JW17", "mathiasn1", "victor", "osanseviero", "NickyNicky", "alvarobartt", "dillfrescott", "philschmid", "not-lain" ], "count": 12 } ]

[]

[ { "type": "text", "value": "The first part of ChemLLM Training Dataset Opensourced on HuggingFace🤗 Now!", "raw": "The first part of ChemLLM Training Dataset Opensourced on HuggingFace🤗 Now!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "You can use this curated DPO dataset to enhance your own model's performance on Chemistry or Materials tasks.", "raw": "You can use this curated DPO dataset to enhance your own model's performance on Chemistry or Materials tasks.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "English version", "raw": "English version", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/AI4Chem/ChemPref-DPO-for-Chemistry-data-en", "href": null, "resource": { "type": "dataset", "id": "AI4Chem/ChemPref-DPO-for-Chemistry-data-en", "discussionNum": null }, "url": "https://huggingface.co/datasets/AI4Chem/ChemPref-DPO-for-Chemistry-data-en", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Chinese Version", "raw": "Chinese Version", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/AI4Chem/ChemPref-DPO-for-Chemistry-data-cn", "href": null, "resource": { "type": "dataset", "id": "AI4Chem/ChemPref-DPO-for-Chemistry-data-cn", "discussionNum": null }, "url": "https://huggingface.co/datasets/AI4Chem/ChemPref-DPO-for-Chemistry-data-cn", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Keep following our [work](", "raw": "Keep following our [work](", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/AI4Chem", "href": "https://huggingface.co/AI4Chem", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ") and [paper](", "raw": ") and [paper](", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2402.06852", "href": null, "resource": { "type": "paper", "id": "2402.06852", "discussionNum": null }, "url": "https://huggingface.co/papers/2402.06852", "code": null, "user": null, "label": "ChemLLM: A Chemical Large Language Model (2402.06852)", "lang": null }, { "type": "text", "value": ").", "raw": ").", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "More Data will be available in weeks, All hail to open source community!", "raw": "More Data will be available in weeks, All hail to open source community!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/64bce15bafd1e46c5504ad38/uqM-u-lZ9bTQOBwFDLTUA.png" } ]

[]

[ { "reaction": "🔥", "users": [ "samusenps", "Csplk", "osanseviero", "mmhamdy", "Joseph717171", "ucyang" ], "count": 6 }, { "reaction": "❤️", "users": [ "clem", "samusenps", "osanseviero", "Joseph717171" ], "count": 4 } ]

[]

[ { "type": "text", "value": "Leave No Context Behind", "raw": "Leave No Context Behind", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Efficient Infinite Context Transformers with Infini-attention", "raw": "Efficient Infinite Context Transformers with Infini-attention", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.07143", "href": null, "resource": { "type": "paper", "id": "2404.07143", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.07143", "code": null, "user": null, "label": "Leave No Context Behind: Efficient Infinite Context Transformers with\n Infini-attention (2404.07143)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This work introduces an efficient method to scale Transformer-based Large Language Models (LLMs) to infinitely long inputs with bounded memory and computation. A key component in our proposed approach is a new attention technique dubbed Infini-attention. The Infini-attention incorporates a compressive memory into the vanilla attention mechanism and builds in both masked local attention and long-term linear attention mechanisms in a single Transformer block. We demonstrate the effectiveness of our approach on long-context language modeling benchmarks, 1M sequence length passkey context block retrieval and 500K length book summarization tasks with 1B and 8B LLMs. Our approach introduces minimal bounded memory parameters and enables fast streaming inference for LLMs.", "raw": "This work introduces an efficient method to scale Transformer-based Large Language Models (LLMs) to infinitely long inputs with bounded memory and computation. A key component in our proposed approach is a new attention technique dubbed Infini-attention. The Infini-attention incorporates a compressive memory into the vanilla attention mechanism and builds in both masked local attention and long-term linear attention mechanisms in a single Transformer block. We demonstrate the effectiveness of our approach on long-context language modeling benchmarks, 1M sequence length passkey context block retrieval and 500K length book summarization tasks with 1B and 8B LLMs. Our approach introduces minimal bounded memory parameters and enables fast streaming inference for LLMs.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60f1abe7544c2adfd699860c/_zUgrQgem4rGtqPTYVEYk.png" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "Ahmedelm", "osanseviero", "clem", "hiyouga", "victor", "xpgx1", "chen-yingfa" ], "count": 8 }, { "reaction": "🚀", "users": [ "samusenps", "clem", "KvrParaskevi" ], "count": 3 }, { "reaction": "➕", "users": [ "samusenps", "clem" ], "count": 2 }, { "reaction": "🔥", "users": [ "hakunamatata1997" ], "count": 1 } ]

[]

[ { "type": "text", "value": "🔍 Today's pick in Interpretability & Analysis of LMs: x2 edition!", "raw": "🔍 Today's pick in Interpretability & Analysis of LMs: x2 edition!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Today's highlighted works aim reproduce findings from Transformer-centric interpretability literature on new RNN-based architectures such as Mamba and RWKV:", "raw": "Today's highlighted works aim reproduce findings from Transformer-centric interpretability literature on new RNN-based architectures such as Mamba and RWKV:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.05971", "href": null, "resource": { "type": "paper", "id": "2404.05971", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.05971", "code": null, "user": null, "label": "Does Transformer Interpretability Transfer to RNNs? (2404.05971)", "lang": null }, { "type": "text", "value": " by ", "raw": " by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@MrGonao", "href": null, "resource": null, "url": null, "code": null, "user": "MrGonao", "label": null, "lang": null }, { "type": "text", "value": " T. Marshall ", "raw": " T. Marshall ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@norabelrose", "href": null, "resource": null, "url": null, "code": null, "user": "norabelrose", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.03646", "href": null, "resource": { "type": "paper", "id": "2404.03646", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.03646", "code": null, "user": null, "label": "Locating and Editing Factual Associations in Mamba (2404.03646)", "lang": null }, { "type": "text", "value": " by ", "raw": " by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@sensharma", "href": null, "resource": null, "url": null, "code": null, "user": "sensharma", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@datkinson", "href": null, "resource": null, "url": null, "code": null, "user": "datkinson", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@davidbau", "href": null, "resource": null, "url": null, "code": null, "user": "davidbau", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The first paper applies contrastive activation addition, the tuned lens and probing for eliciting latent knowledge in quirky models to Mamba and RWKV LMs, finding these Transformer-specific methods can be applied with slight adaptation to these architectures, obtaining similar results.", "raw": "The first paper applies contrastive activation addition, the tuned lens and probing for eliciting latent knowledge in quirky models to Mamba and RWKV LMs, finding these Transformer-specific methods can be applied with slight adaptation to these architectures, obtaining similar results.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The second work applies the ROME method to Mamba, finding weights playing the role of MLPs in encoding factual relations across several Mamba layers, and can be patched to perform model editing. A new SSM-specific technique is also introduced to emulate attention knockout (value zeroing) revealing information flows similar to the ones in Transformers when processing factual statements.", "raw": "The second work applies the ROME method to Mamba, finding weights playing the role of MLPs in encoding factual relations across several Mamba layers, and can be patched to perform model editing. A new SSM-specific technique is also introduced to emulate attention knockout (value zeroing) revealing information flows similar to the ones in Transformers when processing factual statements.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💻 Code: ", "raw": "💻 Code: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/arnab-api/romba", "href": "https://github.com/arnab-api/romba", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 All daily picks: ", "raw": "🔍 All daily picks: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-ofc-lms-65ae3339949c5675d25de2f9", "href": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-ofc-lms-65ae3339949c5675d25de2f9", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/_5PUMcgJ6Di5LJ1xzao4K.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/2jiq8zbgNvd5dLLNQmqOJ.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/h3oDgV9E7ad5CBja_8r2f.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/tbKl7Ayuuwzv4NbAuKNRV.png" } ]

[ { "avatarUrl": "/avatars/c09f6f1df52a710db6a65e37a4545458.svg", "fullname": "David Atkinson", "name": "datkinson", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null }, { "avatarUrl": "/avatars/ac208cd180b4f3ed1ec367e581facfcf.svg", "fullname": "David Bau", "name": "davidbau", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 4 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/643291902bfb2b0ec7556fa9/d68mx--IVqE_b17W6dRsk.png", "fullname": "Gonçalo Paulo", "name": "MrGonao", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1 }, { "avatarUrl": "/avatars/7e361e249d836a9ac01d111c6347e23a.svg", "fullname": "Nora Belrose", "name": "norabelrose", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 6 }, { "avatarUrl": "/avatars/aa4386234a8b1db5dbd2cefb3d90ed08.svg", "fullname": "Arnab Sen Sharma", "name": "sensharma", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3 } ]

[ { "reaction": "❤️", "users": [ "samusenps", "osanseviero", "clem", "datkinson" ], "count": 4 }, { "reaction": "🔥", "users": [ "samusenps", "osanseviero", "clem", "datkinson" ], "count": 4 } ]

[]

[ { "type": "text", "value": "I stopped procrastinating and finally took the time to write the second article of my series of blog posts on SSM: ", "raw": "I stopped procrastinating and finally took the time to write the second article of my series of blog posts on SSM: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/lbourdois/ssm-2022", "href": "https://huggingface.co/blog/lbourdois/ssm-2022", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ".", "raw": ".", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "In this blog post, I review the history of SSM models released in 2022, with over 14 models discussed in a synthetic format. ", "raw": "In this blog post, I review the history of SSM models released in 2022, with over 14 models discussed in a synthetic format. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "They are separated into two parts: \"theoretical\" (DSS, S4D, GSS, Mega, S5, etc.) and \"applications\" (Sashimi, ViS4mer, CCNN, etc.).", "raw": "They are separated into two parts: \"theoretical\" (DSS, S4D, GSS, Mega, S5, etc.) and \"applications\" (Sashimi, ViS4mer, CCNN, etc.).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "To understand everything, it's best to have read the introduction to S4 to SSM blog post first: ", "raw": "To understand everything, it's best to have read the introduction to S4 to SSM blog post first: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/lbourdois/get-on-the-ssm-train", "href": "https://huggingface.co/blog/lbourdois/get-on-the-ssm-train", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ".", "raw": ".", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "All the articles in the series are listed in this space: ", "raw": "All the articles in the series are listed in this space: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/lbourdois/SSM_blog_posts", "href": null, "resource": { "type": "space", "id": "lbourdois/SSM_blog_posts", "discussionNum": null }, "url": "https://huggingface.co/spaces/lbourdois/SSM_blog_posts", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Wishing you a good reading :)", "raw": "Wishing you a good reading :)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "❤️", "users": [ "clem", "samusenps", "PheelaV", "sourabhxiii" ], "count": 4 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/613b0a62a14099d5afed7830/pLuqSIYaNYhUqdjxlNrFn.png", "fullname": "Loïck BOURDOIS", "name": "lbourdois", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 90, "isFollowing": false }, { "avatarUrl": "/avatars/e5b9cc41377f6a06c56e4a1bd59978bc.svg", "fullname": "FilipV", "name": "PheelaV", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "Yesterday, we released Parler-TTS and Data-Speech, fully open-source reproduction of work from the paper: ", "raw": "Yesterday, we released Parler-TTS and Data-Speech, fully open-source reproduction of work from the paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2402.01912", "href": null, "resource": { "type": "paper", "id": "2402.01912", "discussionNum": null }, "url": "https://huggingface.co/papers/2402.01912", "code": null, "user": null, "label": "Natural language guidance of high-fidelity text-to-speech with synthetic\n annotations (2402.01912)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Parler-TTS is a lightweight text-to-speech (TTS) model that can generate high-quality, natural sounding speech in the style of a given speaker (gender, pitch, speaking style, etc).", "raw": "Parler-TTS is a lightweight text-to-speech (TTS) model that can generate high-quality, natural sounding speech in the style of a given speaker (gender, pitch, speaking style, etc).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/collections/parler-tts/parler-tts-fully-open-source-high-quality-tts-models-66164ad285ba03e8ffde214c", "href": "https://huggingface.co/collections/parler-tts/parler-tts-fully-open-source-high-quality-tts-models-66164ad285ba03e8ffde214c", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Parler-TTS Mini v0.1, is the first iteration Parler-TTS model trained using 10k hours of narrated audiobooks. It generates high-quality speech with features that can be controlled using a simple text prompt (e.g. gender, background noise, speaking rate, pitch and reverberation).", "raw": "Parler-TTS Mini v0.1, is the first iteration Parler-TTS model trained using 10k hours of narrated audiobooks. It generates high-quality speech with features that can be controlled using a simple text prompt (e.g. gender, background noise, speaking rate, pitch and reverberation).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "To improve the prosody and naturalness of the speech further, we're scaling up the amount of training data to 50k hours of speech. The v1 release of the model will be trained on this data, as well as inference optimisations, such as flash attention and torch compile.", "raw": "To improve the prosody and naturalness of the speech further, we're scaling up the amount of training data to 50k hours of speech. The v1 release of the model will be trained on this data, as well as inference optimisations, such as flash attention and torch compile.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/parler-tts/parler_tts_mini_v0.1", "href": null, "resource": { "type": "model", "id": "parler-tts/parler_tts_mini_v0.1", "discussionNum": null }, "url": "https://huggingface.co/parler-tts/parler_tts_mini_v0.1", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Data-Speech can be used for annotating speech characteristics in a large-scale setting.", "raw": "Data-Speech can be used for annotating speech characteristics in a large-scale setting.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/parler-tts/open-source-speech-datasets-annotated-using-data-speech-661648ffa0d3d76bfa23d534", "href": null, "resource": { "type": "collection", "id": "parler-tts/open-source-speech-datasets-annotated-using-data-speech-661648ffa0d3d76bfa23d534", "discussionNum": null }, "url": "https://huggingface.co/collections/parler-tts/open-source-speech-datasets-annotated-using-data-speech-661648ffa0d3d76bfa23d534", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This work is both scalable and easily modifiable and will hopefully help the TTS research community explore new ways of conditionning speech synthesis.", "raw": "This work is both scalable and easily modifiable and will hopefully help the TTS research community explore new ways of conditionning speech synthesis.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "All of the datasets, pre-processing, training code and weights are released publicly under permissive license, enabling the community to build on our work and develop their own powerful TTS models.", "raw": "All of the datasets, pre-processing, training code and weights are released publicly under permissive license, enabling the community to build on our work and develop their own powerful TTS models.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/62611fcabbcbd1c34f1615f6/waoP_QoRLhv-NkzYdueIY.mp4" } ]

[]

[ { "reaction": "🔥", "users": [ "sanchit-gandhi", "reach-vb", "thomwolf", "DmitryRyumin", "osanseviero", "clem", "Lyte", "maxugly", "sroman3", "blanchon", "finefin", "erkhem-gantulga", "vedadom", "microdiploma" ], "count": 14 }, { "reaction": "❤️", "users": [ "reach-vb", "thomwolf", "samusenps", "osanseviero", "clem", "maxugly", "vedadom" ], "count": 7 }, { "reaction": "🤯", "users": [ "reach-vb", "thomwolf", "osanseviero", "clem", "maxugly", "vedadom" ], "count": 6 }, { "reaction": "👍", "users": [ "wsuff", "maxugly", "abdouaziiz", "vedadom" ], "count": 4 }, { "reaction": "👀", "users": [ "reach-vb", "clem", "maxugly" ], "count": 3 } ]

[]

[ { "type": "text", "value": "Writer team had the opportunity to run an eval for Mixtral-8x22b, results were interesting.", "raw": "Writer team had the opportunity to run an eval for Mixtral-8x22b, results were interesting.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| ---------------------------- |", "raw": "| ---------------------------- |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| #mmlu 77.26 |", "raw": "| #mmlu 77.26 |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| ---------------------------- |", "raw": "| ---------------------------- |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| #hellaswag 88.81 |", "raw": "| #hellaswag 88.81 |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| ---------------------------- |", "raw": "| ---------------------------- |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| #truthfulqa 52.05 |", "raw": "| #truthfulqa 52.05 |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| ---------------------------- |", "raw": "| ---------------------------- |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| #arc_challenge 70.31 |", "raw": "| #arc_challenge 70.31 |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| ---------------------------- |", "raw": "| ---------------------------- |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| #winogrande 84.93 | ", "raw": "| #winogrande 84.93 | ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| ---------------------------- |", "raw": "| ---------------------------- |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| #gsm8k 76.65 |", "raw": "| #gsm8k 76.65 |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "| ---------------------------- |", "raw": "| ---------------------------- |", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60cd486d723acf5eb46fe8d3/3uzTjarADCM_X435pFvQ6.png" } ]

[]

[ { "reaction": "🔥", "users": [ "den0620", "kur4i", "osanseviero", "clem", "Jason233", "wannaphong", "firstrow", "ChuxiJ", "nbroad" ], "count": 9 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/657eb5b256c9c67605a6e8b5/RPblnGJX57oiIcASEz_S8.png", "fullname": "raincandy_U", "name": "raincandy-u", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 30, "isFollowing": false } ]

[ { "type": "text", "value": "Mistral AI recently released a new Mixtral model. It's another Mixture of Experts model with 8 experts, each with 22B parameters. It requires over 200GB of VRAM to run in float16, and over 70GB of VRAM to run in int4. However, individuals have been successful at finetuning it on Apple Silicon laptops using the MLX framework. It features a 64K context window, twice that of their previous models (32K).", "raw": "Mistral AI recently released a new Mixtral model. It's another Mixture of Experts model with 8 experts, each with 22B parameters. It requires over 200GB of VRAM to run in float16, and over 70GB of VRAM to run in int4. However, individuals have been successful at finetuning it on Apple Silicon laptops using the MLX framework. It features a 64K context window, twice that of their previous models (32K).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The model was released over torrent, a method Mistral has recently often used for their releases. While the license has not been confirmed yet, a moderator on their Discord server yesterday suggested it was Apache 2.0 licensed.", "raw": "The model was released over torrent, a method Mistral has recently often used for their releases. While the license has not been confirmed yet, a moderator on their Discord server yesterday suggested it was Apache 2.0 licensed.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Sources:", "raw": "Sources:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "• ", "raw": "• ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://twitter.com/_philschmid/status/1778051363554934874", "href": "https://twitter.com/_philschmid/status/1778051363554934874", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "• ", "raw": "• ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://twitter.com/reach_vb/status/1777946948617605384", "href": "https://twitter.com/reach_vb/status/1777946948617605384", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "👀", "users": [ "osanseviero", "lewiswatson", "tansutt", "victor", "ajibawa-2023", "samusenps", "antiven0m", "Diegg", "tcml", "simonweniger", "LucienL", "marz1", "clem" ], "count": 13 }, { "reaction": "🚀", "users": [ "alobnayis", "simonweniger", "clem", "Sshubam" ], "count": 4 }, { "reaction": "😎", "users": [ "samusenps", "clem" ], "count": 2 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6538119803519fddb4a17e10/ffJMkdx-rM7VvLTCM6ri_.jpeg", "fullname": "samusenps", "name": "samusenps", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 91, "isFollowing": false } ]

[ { "type": "text", "value": "Diaries of Open Source. Part 14 🤗", "raw": "Diaries of Open Source. Part 14 🤗", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔥CohereForAI releases Command R+, an open 104B model with:", "raw": "🔥CohereForAI releases Command R+, an open 104B model with:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Tool usage capabilities", "raw": "- Tool usage capabilities", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Specialized in RAGs", "raw": "- Specialized in RAGs", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Multilingual", "raw": "- Multilingual", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "It's one of the first models to surpass GPT-4 in the lmsys arena, check it out!", "raw": "It's one of the first models to surpass GPT-4 in the lmsys arena, check it out!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Model: ", "raw": "Model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://hf.co/CohereForAI/c4ai-command-r-plus", "href": null, "resource": { "type": "model", "id": "CohereForAI/c4ai-command-r-plus", "discussionNum": null }, "url": "https://hf.co/CohereForAI/c4ai-command-r-plus", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Official demo: ", "raw": "Official demo: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://hf.co/spaces/CohereForAI/c4ai-command-r-plus", "href": "https://hf.co/spaces/CohereForAI/c4ai-command-r-plus", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Quantized: ", "raw": "Quantized: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://hf.co/CohereForAI/c4ai-command-r-plus-4bit", "href": null, "resource": { "type": "model", "id": "CohereForAI/c4ai-command-r-plus-4bit", "discussionNum": null }, "url": "https://hf.co/CohereForAI/c4ai-command-r-plus-4bit", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🎉Google releases a new version of their Gemma instruct models, with improved quality, nicer to converse, and a fancier RL algorithm. The model is similar to Llama 2 70B in the Chat Arena!", "raw": "🎉Google releases a new version of their Gemma instruct models, with improved quality, nicer to converse, and a fancier RL algorithm. The model is similar to Llama 2 70B in the Chat Arena!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Models: ", "raw": "Models: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://hf.co/collections/google/gemma-release-65d5efbccdbb8c4202ec078b", "href": null, "resource": { "type": "collection", "id": "google/gemma-release-65d5efbccdbb8c4202ec078b", "discussionNum": null }, "url": "https://hf.co/collections/google/gemma-release-65d5efbccdbb8c4202ec078b", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Try it out in HuggingChat ", "raw": "Try it out in HuggingChat ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://hf.co/chat/models/google/gemma-1.1-7b-it", "href": "https://hf.co/chat/models/google/gemma-1.1-7b-it", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🪄VoiceCraft, a speech editing and TTS SOTA open model", "raw": "🪄VoiceCraft, a speech editing and TTS SOTA open model", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Paper: ", "raw": "Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://hf.co/papers/2403.16973", "href": null, "resource": { "type": "paper", "id": "2403.16973", "discussionNum": null }, "url": "https://hf.co/papers/2403.16973", "code": null, "user": null, "label": "VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wild (2403.16973)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Model: ", "raw": "Model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://hf.co/pyp1/VoiceCraft", "href": null, "resource": { "type": "model", "id": "pyp1/VoiceCraft", "discussionNum": null }, "url": "https://hf.co/pyp1/VoiceCraft", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💻Google released CodeGemma, a family of code generation, completion, and chat models", "raw": "💻Google released CodeGemma, a family of code generation, completion, and chat models", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Blog post: ", "raw": "Blog post: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://hf.co/blog/codegemma", "href": "https://hf.co/blog/codegemma", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Models: ", "raw": "Models: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://hf.co/collections/google/codegemma-release-66152ac7b683e2667abdee11", "href": null, "resource": { "type": "collection", "id": "google/codegemma-release-66152ac7b683e2667abdee11", "discussionNum": null }, "url": "https://hf.co/collections/google/codegemma-release-66152ac7b683e2667abdee11", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Report: ", "raw": "Report: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://storage.googleapis.com/deepmind-media/gemma/codegemma_report.pdf", "href": "https://storage.googleapis.com/deepmind-media/gemma/codegemma_report.pdf", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Misc models:", "raw": "Misc models:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🦖T-Rex2, a very powerful object detection model for many applications ", "raw": "🦖T-Rex2, a very powerful object detection model for many applications ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/IDEA-Research/T-Rex", "href": "https://github.com/IDEA-Research/T-Rex", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👀 CT-RATE : A 3D dataset paired with text reports ", "raw": "👀 CT-RATE : A 3D dataset paired with text reports ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://hf.co/datasets/ibrahimhamamci/CT-RATE", "href": null, "resource": { "type": "dataset", "id": "ibrahimhamamci/CT-RATE", "discussionNum": null }, "url": "https://hf.co/datasets/ibrahimhamamci/CT-RATE", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🐙Octopus v2: a Gemma-based model trained for Android API - extremely fast, better than Llama+RAG, great results ", "raw": "🐙Octopus v2: a Gemma-based model trained for Android API - extremely fast, better than Llama+RAG, great results ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://hf.co/NexaAIDev/Octopus-v2", "href": null, "resource": { "type": "model", "id": "NexaAIDev/Octopus-v2", "discussionNum": null }, "url": "https://hf.co/NexaAIDev/Octopus-v2", "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🔥", "users": [ "qnguyen3", "YaTharThShaRma999", "samusenps", "danielus", "linoyts", "simonweniger", "clem", "megrisdal", "ashwinlimaye", "USM-Valor", "merve" ], "count": 11 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6032802e1f993496bc14d9e3/w6hr-DEQot4VVkoyRIBiy.png", "fullname": "Omar Sanseviero", "name": "osanseviero", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2868, "isFollowing": false }, { "avatarUrl": "/avatars/6f6823095d3073e80aa2a76657938717.svg", "fullname": "hennie", "name": "bonydork", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "🔍 Today's pick in Interpretability & Analysis of LMs: Context versus Prior Knowledge in Language Models by ", "raw": "🔍 Today's pick in Interpretability & Analysis of LMs: Context versus Prior Knowledge in Language Models by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@kdu4108", "href": null, "resource": null, "url": null, "code": null, "user": "kdu4108", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@vesteinn", "href": null, "resource": null, "url": null, "code": null, "user": "vesteinn", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@niklasstoehr", "href": null, "resource": null, "url": null, "code": null, "user": "niklasstoehr", "label": null, "lang": null }, { "type": "text", "value": " J. C. White A. Schein ", "raw": " J. C. White A. Schein ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@rcotterell", "href": null, "resource": null, "url": null, "code": null, "user": "rcotterell", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This work examines the influence of context versus memorized knowledge in LMs through the lens of the shift caused by contexts at various degrees of informativeness to the models' predictive distribution. Understanding this difference is especially important in the context of knowledge conflicts between memorized and contextual information.", "raw": "This work examines the influence of context versus memorized knowledge in LMs through the lens of the shift caused by contexts at various degrees of informativeness to the models' predictive distribution. Understanding this difference is especially important in the context of knowledge conflicts between memorized and contextual information.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Authors propose disentangling context influence in terms of \"persuasion\", i.e. how impactful is the inclusion of the context for answers of a given query/entity pair, and \"susceptibility\", i.e. how much answers of a given query/entity pair are likely to be swayed by the presence of context, and operationalize these concepts using information-theoretic measures akin to mutual information.", "raw": "Authors propose disentangling context influence in terms of \"persuasion\", i.e. how impactful is the inclusion of the context for answers of a given query/entity pair, and \"susceptibility\", i.e. how much answers of a given query/entity pair are likely to be swayed by the presence of context, and operationalize these concepts using information-theoretic measures akin to mutual information.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The two metrics are validated using a synthetic dataset sourced from a knowledge graph. Analysis shows that:
", "raw": "The two metrics are validated using a synthetic dataset sourced from a knowledge graph. Analysis shows that:
", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- The degree of persuasiveness of relevant contexts increases with the increase of model size (interesting implications here for the jailbreaking of LLMs!)", "raw": "- The degree of persuasiveness of relevant contexts increases with the increase of model size (interesting implications here for the jailbreaking of LLMs!)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- assertive contexts tend to be more persuasive for closed queries (yes/no) and mid-sized models", "raw": "- assertive contexts tend to be more persuasive for closed queries (yes/no) and mid-sized models", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Negation affect context persuasiveness", "raw": "- Negation affect context persuasiveness", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Familiar entities (explored as real vs. fake, more frequent in training data and more connected in the KG) are less susceptible to context influence", "raw": "- Familiar entities (explored as real vs. fake, more frequent in training data and more connected in the KG) are less susceptible to context influence", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Finally, authors suggest applications of the persuasion/susceptibility framing for social science analyses and gender bias evaluation.", "raw": "Finally, authors suggest applications of the persuasion/susceptibility framing for social science analyses and gender bias evaluation.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💻 Code: ", "raw": "💻 Code: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/kdu4108/measureLM", "href": "https://github.com/kdu4108/measureLM", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Paper: ", "raw": "📄 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.04633", "href": null, "resource": { "type": "paper", "id": "2404.04633", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.04633", "code": null, "user": null, "label": "Context versus Prior Knowledge in Language Models (2404.04633)", "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 All daily picks: ", "raw": "🔍 All daily picks: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-ofc-lms-65ae3339949c5675d25de2f9", "href": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-ofc-lms-65ae3339949c5675d25de2f9", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/M_e5Xt_OK1gZ0llKCUku7.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/thh6WrBdyO2XFt7U0FPFE.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/Rm94A5dE2dFtSLHw6MYUV.png" } ]

[ { "avatarUrl": "/avatars/cc01563723f70516e4aefd3da625bfb7.svg", "fullname": "Kevin Du", "name": "kdu4108", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/62a06ed13df0325603b5f038/VYFIR3T6_CBfGfP5T8nIK.jpeg", "fullname": "Niklas Stoehr", "name": "niklasstoehr", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1 }, { "avatarUrl": "/avatars/b719127447ed71eb128f418765cb0cce.svg", "fullname": "Ryan Cotterell", "name": "rcotterell", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/5f915e09161dc51925934edf/2PqjbJ7FVG3FOCNmn7ORY.png", "fullname": "Vésteinn Snæbjarnarson", "name": "vesteinn", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 8 } ]

[ { "reaction": "👍", "users": [ "niklasstoehr", "Lei-bw" ], "count": 2 }, { "reaction": "❤️", "users": [ "afrideva" ], "count": 1 } ]

[]

[ { "type": "text", "value": "I'm extending the AI Comic Factory (to follow-up on the \"bring your own model\" philosophy 🤗) to support the broader LLM ecosystem of our other vendor friends!", "raw": "I'm extending the AI Comic Factory (to follow-up on the \"bring your own model\" philosophy 🤗) to support the broader LLM ecosystem of our other vendor friends!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Here is the announcement: ", "raw": "Here is the announcement: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/jbilcke-hf/ai-comic-factory/discussions/723", "href": null, "resource": { "type": "space", "id": "jbilcke-hf/ai-comic-factory", "discussionNum": 723 }, "url": "https://huggingface.co/spaces/jbilcke-hf/ai-comic-factory/discussions/723", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This is an experimental feature, some ", "raw": "This is an experimental feature, some ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "models/vendors", "raw": "models/vendors", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " might require parameter tuning but I haven't tested all of them yet (only a bit of GPT-4 👀)", "raw": " might require parameter tuning but I haven't tested all of them yet (only a bit of GPT-4 👀)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Let me know if you experience any issues!", "raw": "Let me know if you experience any issues!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/jbilcke-hf/ai-comic-factory", "href": null, "resource": { "type": "space", "id": "jbilcke-hf/ai-comic-factory", "discussionNum": null }, "url": "https://huggingface.co/spaces/jbilcke-hf/ai-comic-factory", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/64904918584563e08e84d39b/lVJupbK5Q_fPj6gOfyxiw.png" } ]

[]

[ { "reaction": "🔥", "users": [ "m-ric", "samusenps", "kramp", "thomwolf" ], "count": 4 }, { "reaction": "❤️", "users": [ "ari1000" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/2RK8J_YSNAK2ob8XZH7w2.jpeg", "fullname": "Julian Bilcke", "name": "jbilcke-hf", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1312, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/EWIWVQMz9BApVOa4tSJt9.jpeg", "fullname": "Kirito kirigaya", "name": "KIRITO4223", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "Another gem from our lab — DGInStyle! We use Stable Diffusion to generate semantic segmentation data for autonomous driving and train domain-generalizable networks.", "raw": "Another gem from our lab — DGInStyle! We use Stable Diffusion to generate semantic segmentation data for autonomous driving and train domain-generalizable networks.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📟 Website: ", "raw": "📟 Website: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://dginstyle.github.io", "href": "https://dginstyle.github.io", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🧾 Paper: ", "raw": "🧾 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://arxiv.org/abs/2312.03048", "href": "https://arxiv.org/abs/2312.03048", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤗 Hugging Face Paper: ", "raw": "🤗 Hugging Face Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2312.03048", "href": null, "resource": { "type": "paper", "id": "2312.03048", "discussionNum": null }, "url": "https://huggingface.co/papers/2312.03048", "code": null, "user": null, "label": "DGInStyle: Domain-Generalizable Semantic Segmentation with Image\n Diffusion Models and Stylized Semantic Control (2312.03048)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤗 Hugging Face Model: ", "raw": "🤗 Hugging Face Model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/yurujaja/DGInStyle", "href": null, "resource": { "type": "model", "id": "yurujaja/DGInStyle", "discussionNum": null }, "url": "https://huggingface.co/yurujaja/DGInStyle", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🐙 Code: ", "raw": "🐙 Code: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/yurujaja/DGInStyle", "href": "https://github.com/yurujaja/DGInStyle", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "In a nutshell, our pipeline overcomes the resolution loss of Stable Diffusion latent space and the style bias of ControlNet, as shown in the attached figures. This allows us to generate sufficiently high-quality pairs of images and semantic masks to train domain-generalizable semantic segmentation networks. ", "raw": "In a nutshell, our pipeline overcomes the resolution loss of Stable Diffusion latent space and the style bias of ControlNet, as shown in the attached figures. This allows us to generate sufficiently high-quality pairs of images and semantic masks to train domain-generalizable semantic segmentation networks. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Team: Yuru Jia (", "raw": "Team: Yuru Jia (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@yurujaja", "href": null, "resource": null, "url": null, "code": null, "user": "yurujaja", "label": null, "lang": null }, { "type": "text", "value": "), Lukas Hoyer, Shengyu Huang, Tianfu Wang (", "raw": "), Lukas Hoyer, Shengyu Huang, Tianfu Wang (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@Tianfwang", "href": null, "resource": null, "url": null, "code": null, "user": "Tianfwang", "label": null, "lang": null }, { "type": "text", "value": "), Luc Van Gool, Konrad Schindler, and Anton Obukhov (", "raw": "), Luc Van Gool, Konrad Schindler, and Anton Obukhov (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@toshas", "href": null, "resource": null, "url": null, "code": null, "user": "toshas", "label": null, "lang": null }, { "type": "text", "value": ").", "raw": ").", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/62f93abbc4817cfc0756b6f8/p6nz7SW15cMh8h2lHvcsV.qt" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/62f93abbc4817cfc0756b6f8/l0AckKGOOqzA8LDC5avij.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/62f93abbc4817cfc0756b6f8/GqvWNIqe09Cu_Zwjj1Yyp.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/62f93abbc4817cfc0756b6f8/IHEMpONnUZjPwpVR-asts.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/62f93abbc4817cfc0756b6f8/hTVqgpYHiXTP2sqOM3TKd.jpeg" } ]

[ { "avatarUrl": "/avatars/a06aecf91bcc005e618460ab31228fb7.svg", "fullname": "Tianfu Wang", "name": "Tianfwang", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/62f93abbc4817cfc0756b6f8/rGYLaq-rmoJJYotkC1VXk.jpeg", "fullname": "Anton Obukhov", "name": "toshas", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 68 }, { "avatarUrl": "/avatars/07b8cdfede0923a6e6e5170341f32eab.svg", "fullname": "Yuru Jia", "name": "yurujaja", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1 } ]

[ { "reaction": "❤️", "users": [ "samusenps", "clem", "Joseph717171", "oceansweep", "Timilla" ], "count": 5 }, { "reaction": "🚀", "users": [ "osanseviero", "clem", "Joseph717171", "oceansweep" ], "count": 4 }, { "reaction": "👍", "users": [ "Timilla", "Joseph717171", "oceansweep" ], "count": 3 } ]

[]

[ { "type": "text", "value": "LLM2Vec", "raw": "LLM2Vec", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Large Language Models Are Secretly Powerful Text Encoders", "raw": "Large Language Models Are Secretly Powerful Text Encoders", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.05961", "href": null, "resource": { "type": "paper", "id": "2404.05961", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.05961", "code": null, "user": null, "label": "LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders (2404.05961)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Large decoder-only language models (LLMs) are the state-of-the-art models on most of today's NLP tasks and benchmarks. Yet, the community is only slowly adopting these models for text embedding tasks, which require rich contextualized representations. In this work, we introduce LLM2Vec, a simple unsupervised approach that can transform any decoder-only LLM into a strong text encoder. LLM2Vec consists of three simple steps: 1) enabling bidirectional attention, 2) masked next token prediction, and 3) unsupervised contrastive learning. We demonstrate the effectiveness of LLM2Vec by applying it to 3 popular LLMs ranging from 1.3B to 7B parameters and evaluate the transformed models on English word- and sequence-level tasks. We outperform encoder-only models by a large margin on word-level tasks and reach a new unsupervised state-of-the-art performance on the Massive Text Embeddings Benchmark (MTEB). Moreover, when combining LLM2Vec with supervised contrastive learning, we achieve state-of-the-art performance on MTEB among models that train only on publicly available data. Our strong empirical results and extensive analysis demonstrate that LLMs can be effectively transformed into universal text encoders in a parameter-efficient manner without the need for expensive adaptation or synthetic GPT-4 generated data.", "raw": "Large decoder-only language models (LLMs) are the state-of-the-art models on most of today's NLP tasks and benchmarks. Yet, the community is only slowly adopting these models for text embedding tasks, which require rich contextualized representations. In this work, we introduce LLM2Vec, a simple unsupervised approach that can transform any decoder-only LLM into a strong text encoder. LLM2Vec consists of three simple steps: 1) enabling bidirectional attention, 2) masked next token prediction, and 3) unsupervised contrastive learning. We demonstrate the effectiveness of LLM2Vec by applying it to 3 popular LLMs ranging from 1.3B to 7B parameters and evaluate the transformed models on English word- and sequence-level tasks. We outperform encoder-only models by a large margin on word-level tasks and reach a new unsupervised state-of-the-art performance on the Massive Text Embeddings Benchmark (MTEB). Moreover, when combining LLM2Vec with supervised contrastive learning, we achieve state-of-the-art performance on MTEB among models that train only on publicly available data. Our strong empirical results and extensive analysis demonstrate that LLMs can be effectively transformed into universal text encoders in a parameter-efficient manner without the need for expensive adaptation or synthetic GPT-4 generated data.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60f1abe7544c2adfd699860c/y8ZJR3T0xtEYg6jKu6XuJ.png" } ]

[]

[ { "reaction": "🔥", "users": [ "SivilTaram", "osanseviero", "CarlLee", "oliverguhr", "KvrParaskevi", "m18coppola", "celinah" ], "count": 7 }, { "reaction": "👍", "users": [ "AshokRaja", "Sylvestre" ], "count": 2 } ]

[]

[ { "type": "text", "value": "𝗡𝗲𝘄 𝗦𝗽𝗮𝗰𝗲: 𝘼𝙄 𝙏𝙧𝙖𝙫𝙚𝙡 𝙥𝙡𝙖𝙣𝙣𝙚𝙧 🗺️🏕️ Plan your next vacation in a few minutes!", "raw": "𝗡𝗲𝘄 𝗦𝗽𝗮𝗰𝗲: 𝘼𝙄 𝙏𝙧𝙖𝙫𝙚𝙡 𝙥𝙡𝙖𝙣𝙣𝙚𝙧 🗺️🏕️ Plan your next vacation in a few minutes!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I wanted to try out if a powerful LLM like Mixtral-8x7b had geographical reasoning capabilities.", "raw": "I wanted to try out if a powerful LLM like Mixtral-8x7b had geographical reasoning capabilities.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "So I built a small space that prompts the LLM to provide a JSON list of places based on a user input.", "raw": "So I built a small space that prompts the LLM to provide a JSON list of places based on a user input.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "And the result was impressive! 🤯 ", "raw": "And the result was impressive! 🤯 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "⇒ 𝗜𝘁 𝘀𝗲𝗲𝗺𝘀 𝗹𝗶𝗸𝗲 𝗠𝗶𝘅𝘁𝗿𝗮𝗹 𝗵𝗮𝘀 𝗮 𝗴𝗿𝗮𝘀𝗽 𝗼𝗳 𝗴𝗲𝗼𝗴𝗿𝗮𝗽𝗵𝗶𝗰𝗮𝗹 𝗰𝗼𝗻𝗰𝗲𝗽𝘁𝘀 𝗹𝗶𝗸𝗲 𝗡𝗼𝗿𝘁𝗵 - 𝗦𝗼𝘂𝘁𝗵, 𝗼𝗿 𝘀𝗽𝗮𝘁𝗶𝗮𝗹 𝗮𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁.🧭 Not just describing these concepts, but really applying them in practice, for instance to successfully answer \"give me 4 European cities that are aligned on the map\". This is a 𝗻𝗶𝗰𝗲 𝗲𝘅𝗮𝗺𝗽𝗹𝗲 𝗼𝗳 𝗮𝗻 𝗲𝗺𝗲𝗿𝗴𝗲𝗻𝘁 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝘆, since nothing in the LLM's training data should prepare it for this specific task.", "raw": "⇒ 𝗜𝘁 𝘀𝗲𝗲𝗺𝘀 𝗹𝗶𝗸𝗲 𝗠𝗶𝘅𝘁𝗿𝗮𝗹 𝗵𝗮𝘀 𝗮 𝗴𝗿𝗮𝘀𝗽 𝗼𝗳 𝗴𝗲𝗼𝗴𝗿𝗮𝗽𝗵𝗶𝗰𝗮𝗹 𝗰𝗼𝗻𝗰𝗲𝗽𝘁𝘀 𝗹𝗶𝗸𝗲 𝗡𝗼𝗿𝘁𝗵 - 𝗦𝗼𝘂𝘁𝗵, 𝗼𝗿 𝘀𝗽𝗮𝘁𝗶𝗮𝗹 𝗮𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁.🧭 Not just describing these concepts, but really applying them in practice, for instance to successfully answer \"give me 4 European cities that are aligned on the map\". This is a 𝗻𝗶𝗰𝗲 𝗲𝘅𝗮𝗺𝗽𝗹𝗲 𝗼𝗳 𝗮𝗻 𝗲𝗺𝗲𝗿𝗴𝗲𝗻𝘁 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝘆, since nothing in the LLM's training data should prepare it for this specific task.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Anyway, I added API calls and a nice visualization on top of the LLM, streaming output, caching for the answers and locations... and ta-da! ✨ I got the 𝗔𝗜 𝗧𝗿𝗮𝘃𝗲𝗹 𝗣𝗹𝗮𝗻𝗻𝗲𝗿.", "raw": "Anyway, I added API calls and a nice visualization on top of the LLM, streaming output, caching for the answers and locations... and ta-da! ✨ I got the 𝗔𝗜 𝗧𝗿𝗮𝘃𝗲𝗹 𝗣𝗹𝗮𝗻𝗻𝗲𝗿.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "𝙔𝙤𝙪 𝙘𝙖𝙣 𝙙𝙚𝙨𝙘𝙧𝙞𝙗𝙚 𝙞𝙩 𝙮𝙤𝙪𝙧 𝙩𝙧𝙞𝙥, 𝙖𝙣𝙙 𝙞𝙩 𝙬𝙞𝙡𝙡 𝙘𝙤𝙢𝙚 𝙪𝙥 𝙬𝙞𝙩𝙝 𝙣𝙞𝙘𝙚 𝙖𝙣𝙙 𝙘𝙤𝙣𝙫𝙚𝙣𝙞𝙚𝙣𝙩 𝙡𝙤𝙘𝙖𝙩𝙞𝙤𝙣𝙨!", "raw": "𝙔𝙤𝙪 𝙘𝙖𝙣 𝙙𝙚𝙨𝙘𝙧𝙞𝙗𝙚 𝙞𝙩 𝙮𝙤𝙪𝙧 𝙩𝙧𝙞𝙥, 𝙖𝙣𝙙 𝙞𝙩 𝙬𝙞𝙡𝙡 𝙘𝙤𝙢𝙚 𝙪𝙥 𝙬𝙞𝙩𝙝 𝙣𝙞𝙘𝙚 𝙖𝙣𝙙 𝙘𝙤𝙣𝙫𝙚𝙣𝙞𝙚𝙣𝙩 𝙡𝙤𝙘𝙖𝙩𝙞𝙤𝙣𝙨!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "𝙏𝙧𝙮 𝙞𝙩 𝙝𝙚𝙧𝙚 👉 ", "raw": "𝙏𝙧𝙮 𝙞𝙩 𝙝𝙚𝙧𝙚 👉 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/m-ric/ai-travel-planner", "href": null, "resource": { "type": "space", "id": "m-ric/ai-travel-planner", "discussionNum": null }, "url": "https://huggingface.co/spaces/m-ric/ai-travel-planner", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Thank you ", "raw": "Thank you ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@freddyaboulton", "href": null, "resource": null, "url": null, "code": null, "user": "freddyaboulton", "label": null, "lang": null }, { "type": "text", "value": " for the 𝚐𝚛𝚊𝚍𝚒𝚘_𝚏𝚘𝚕𝚒𝚞𝚖 component, and ", "raw": " for the 𝚐𝚛𝚊𝚍𝚒𝚘_𝚏𝚘𝚕𝚒𝚞𝚖 component, and ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@clem", "href": null, "resource": null, "url": null, "code": null, "user": "clem", "label": null, "lang": null }, { "type": "text", "value": " , ", "raw": " , ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@pngwn", "href": null, "resource": null, "url": null, "code": null, "user": "pngwn", "label": null, "lang": null }, { "type": "text", "value": " , ", "raw": " , ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@abidlabs", "href": null, "resource": null, "url": null, "code": null, "user": "abidlabs", "label": null, "lang": null }, { "type": "text", "value": " for your ideas and support!", "raw": " for your ideas and support!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/63d10d4e8eaa4831005e92b5/EdcSn_rsFZeIcpPWXQnD2.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1621947938344-noauth.png", "fullname": "Abubakar Abid", "name": "abidlabs", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 487 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1583857146757-5e67bdd61009063689407479.jpeg", "fullname": "Clem 🤗", "name": "clem", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 1763 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1654278567459-626a9bfa03e2e2796f24ca11.jpeg", "fullname": "Freddy Boulton", "name": "freddyaboulton", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 164 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1641398053462-noauth.jpeg", "fullname": "Pete", "name": "pngwn", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 70 } ]

[ { "reaction": "🚀", "users": [ "pngwn", "lunarflu", "osanseviero", "victor", "clem", "monsoon-nlp", "freddyaboulton" ], "count": 7 }, { "reaction": "🔥", "users": [ "pngwn", "lunarflu", "clem", "Locutusque", "adamelliotfields", "raincandy-u", "freddyaboulton" ], "count": 7 }, { "reaction": "🤗", "users": [ "lunarflu", "clem", "freddyaboulton" ], "count": 3 }, { "reaction": "🤯", "users": [ "lunarflu", "clem", "freddyaboulton" ], "count": 3 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1641398053462-noauth.jpeg", "fullname": "Pete", "name": "pngwn", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 70, "isFollowing": false } ]

[ { "type": "text", "value": "🎉 Introducing nanoLLaVA, a powerful multimodal AI model that packs the capabilities of a 1B parameter vision language model into just 5GB of VRAM. 🚀 This makes it an ideal choice for edge devices, bringing cutting-edge visual understanding and generation to your devices like never before. 📱💻", "raw": "🎉 Introducing nanoLLaVA, a powerful multimodal AI model that packs the capabilities of a 1B parameter vision language model into just 5GB of VRAM. 🚀 This makes it an ideal choice for edge devices, bringing cutting-edge visual understanding and generation to your devices like never before. 📱💻", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Model: ", "raw": "Model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/qnguyen3/nanoLLaVA", "href": null, "resource": { "type": "model", "id": "qnguyen3/nanoLLaVA", "discussionNum": null }, "url": "https://huggingface.co/qnguyen3/nanoLLaVA", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " 🔍", "raw": " 🔍", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Spaces: ", "raw": "Spaces: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/qnguyen3/nanoLLaVA", "href": null, "resource": { "type": "space", "id": "qnguyen3/nanoLLaVA", "discussionNum": null }, "url": "https://huggingface.co/spaces/qnguyen3/nanoLLaVA", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " (thanks to ", "raw": " (thanks to ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@merve", "href": null, "resource": null, "url": null, "code": null, "user": "merve", "label": null, "lang": null }, { "type": "text", "value": ")", "raw": ")", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Under the hood, nanoLLaVA is based on the powerful ", "raw": "Under the hood, nanoLLaVA is based on the powerful ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/vilm/Quyen-SE-v0.1", "href": null, "resource": { "type": "model", "id": "vilm/Quyen-SE-v0.1", "discussionNum": null }, "url": "https://huggingface.co/vilm/Quyen-SE-v0.1", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " (my Qwen1.5-0.5B finetune) and Google's impressive ", "raw": " (my Qwen1.5-0.5B finetune) and Google's impressive ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/google/siglip-so400m-patch14-384", "href": null, "resource": { "type": "model", "id": "google/siglip-so400m-patch14-384", "discussionNum": null }, "url": "https://huggingface.co/google/siglip-so400m-patch14-384", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ". 🧠 The model is trained using a data-centric approach to ensure optimal performance. 📊", "raw": ". 🧠 The model is trained using a data-centric approach to ensure optimal performance. 📊", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "In the spirit of transparency and collaboration, all code and model weights are open-sourced under the Apache 2.0 license. 🤝", "raw": "In the spirit of transparency and collaboration, all code and model weights are open-sourced under the Apache 2.0 license. 🤝", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/630430583926de1f7ec62c6b/f1Q_foPsVfKhyge9JeMSg.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1648113222875-6141a88b3a0ec78603c9e784.png", "fullname": "Merve Noyan", "name": "merve", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 5589 } ]

[ { "reaction": "🔥", "users": [ "samusenps", "SivilTaram", "DmitryRyumin", "m-ric", "osanseviero", "ArthurFischel", "den0620", "2dts", "merve", "qnguyen3", "ggcristian" ], "count": 11 }, { "reaction": "❤️", "users": [ "samusenps", "SivilTaram", "ndamulelonemakh", "osanseviero", "penma", "2dts", "cnmoro", "merve" ], "count": 8 }, { "reaction": "🚀", "users": [ "SivilTaram", "2dts" ], "count": 2 }, { "reaction": "🤗", "users": [ "ijohn07" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64da082a79c87e13ca232042/DKGEIffK3kLj7SQ20yYiK.png", "fullname": "Cristian Gutiérrez", "name": "ggcristian", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "🚀🕺🌟 New Research Alert (Avatars Collection)! 🌟💃🚀", "raw": "🚀🕺🌟 New Research Alert (Avatars Collection)! 🌟💃🚀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Title: PhysAvatar: Learning the Physics of Dressed 3D Avatars from Visual Observations 🔝", "raw": "📄 Title: PhysAvatar: Learning the Physics of Dressed 3D Avatars from Visual Observations 🔝", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📝 Description: PhysAvatar is a novel framework that uses inverse rendering and physics to autonomously reconstruct the shape, appearance, and physical properties of clothed human avatars from multi-view video data.", "raw": "📝 Description: PhysAvatar is a novel framework that uses inverse rendering and physics to autonomously reconstruct the shape, appearance, and physical properties of clothed human avatars from multi-view video data.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👥 Authors: Yang Zheng et al.", "raw": "👥 Authors: Yang Zheng et al.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔗 Paper: ", "raw": "🔗 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.04421", "href": null, "resource": { "type": "paper", "id": "2404.04421", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.04421", "code": null, "user": null, "label": "PhysAvatar: Learning the Physics of Dressed 3D Avatars from Visual\n Observations (2404.04421)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌐 GitHub Page: ", "raw": "🌐 GitHub Page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://qingqing-zhao.github.io/PhysAvatar", "href": "https://qingqing-zhao.github.io/PhysAvatar", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "raw": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "href": null, "resource": { "type": "space", "id": "DmitryRyumin/NewEraAI-Papers", "discussionNum": null }, "url": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " curated by ", "raw": " curated by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DmitryRyumin", "href": null, "resource": null, "url": null, "code": null, "user": "DmitryRyumin", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Added to the Avatars Collection: ", "raw": "🚀 Added to the Avatars Collection: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "href": null, "resource": { "type": "collection", "id": "DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "discussionNum": null }, "url": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Keywords: #PhysAvatar #DigitalHumans #InverseRendering #PhysicsSimulation #AvatarModeling #ClothSimulation #PhotorealisticRendering #ComputerVision #DeepLearning #Innovation", "raw": "🔍 Keywords: #PhysAvatar #DigitalHumans #InverseRendering #PhysicsSimulation #AvatarModeling #ClothSimulation #PhotorealisticRendering #ComputerVision #DeepLearning #Innovation", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/5CrDS5Gi9vGXnmOu8JES8.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/N83HO3AqvALR-CCXDCyXF.gif" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/2mYwu1qmMZ77CGRKpvfJq.gif" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/XzE09iAKL6h8E9o_pMZRu.gif" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/TsxEbCAzYfILvRodXdCJJ.gif" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/Z983PGVuPAK88ZKx_EcEG.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/gvOvLkV9NNaz4Yd7L-gvA.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/BYBeQyd-BRBSRVMhruYIM.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/a8LHzGKNptVW0IFJI9Lv8.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/Y-SlM1IPEoS45fr9jcz6Z.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/SnedwFnC9lVCQLaQZJHnz.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/ZvZn5ERhmiYEWJZyaJ9Y9.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/nRCxbVng_PPBqKd-Z3KVc.jpeg", "fullname": "Dmitry Ryumin", "name": "DmitryRyumin", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 377 } ]

[ { "reaction": "🔥", "users": [ "DmitryRyumin", "samusenps", "guandao" ], "count": 3 }, { "reaction": "❤️", "users": [ "samusenps", "guandao", "bbossam" ], "count": 3 }, { "reaction": "🤗", "users": [ "DmitryRyumin", "guandao" ], "count": 2 } ]

[]

[ { "type": "text", "value": "**Some updates on GLiNER**", "raw": "**Some updates on GLiNER**", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🆕 A new commercially permissible multilingual version is available urchade/gliner_multiv2.1", "raw": "🆕 A new commercially permissible multilingual version is available urchade/gliner_multiv2.1", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🐛 A subtle bug that causes performance degradation on some models has been corrected. Thanks to ", "raw": "🐛 A subtle bug that causes performance degradation on some models has been corrected. Thanks to ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@yyDing1", "href": null, "resource": null, "url": null, "code": null, "user": "yyDing1", "label": null, "lang": null }, { "type": "text", "value": " for raising the issue.", "raw": " for raising the issue.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "code_fence", "value": null, "raw": "```\nfrom gliner import GLiNER\n\n# Initialize GLiNER\nmodel = GLiNER.from_pretrained(\"urchade/gliner_multiv2.1\")\n\ntext = \"This is a text about Bill Gates and Microsoft.\"\n\n# Labels for entity prediction\nlabels = [\"person\", \"organization\", \"email\"]\n\nentities = model.predict_entities(text, labels, threshold=0.5)\n\nfor entity in entities:\n print(entity[\"text\"], \"=>\", entity[\"label\"])\n```", "href": null, "resource": null, "url": null, "code": "from gliner import GLiNER\n\n# Initialize GLiNER\nmodel = GLiNER.from_pretrained(\"urchade/gliner_multiv2.1\")\n\ntext = \"This is a text about Bill Gates and Microsoft.\"\n\n# Labels for entity prediction\nlabels = [\"person\", \"organization\", \"email\"]\n\nentities = model.predict_entities(text, labels, threshold=0.5)\n\nfor entity in entities:\n print(entity[\"text\"], \"=>\", entity[\"label\"])", "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "❤️", "users": [ "tomaarsen", "osanseviero", "aari1995", "Bailefan", "erickdp", "mrdbourke" ], "count": 6 }, { "reaction": "🔥", "users": [ "osanseviero", "tomaarsen", "som781" ], "count": 3 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6317233cc92fd6fee317e030/cJHSvvimr1kqgQfHOjO5n.png", "fullname": "Tom Aarsen", "name": "tomaarsen", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1060, "isFollowing": false }, { "avatarUrl": "/avatars/87a827f22fd65236378820e838828101.svg", "fullname": "weiguosun", "name": "chuangxinlezhi", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/e13985712c72cb9dbc517a4a293ad006.svg", "fullname": "abdata", "name": "abdar1925", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/62111fdbe1d974ee5bcbfa27/YUzX6lBvW8pbxDorx1kgV.png", "fullname": "Urchade Zaratiana", "name": "urchade", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 149, "isFollowing": false }, { "avatarUrl": "/avatars/f6e6a24f7d3068fedfbaff75493e4daf.svg", "fullname": "Krishna Meduri", "name": "meduri30", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/91672c1354c7eb06d4b3f83a4f6d3458.svg", "fullname": "Rahul", "name": "Rjp97", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]

[ { "type": "text", "value": "SwapAnything is a new method that allows swapping any object in an image with personalized concepts given by a reference image.", "raw": "SwapAnything is a new method that allows swapping any object in an image with personalized concepts given by a reference image.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Key points: ", "raw": "Key points: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "1️⃣ It uses pre-trained diffusion models to enable precise and high-fidelity object swapping in images.", "raw": "1️⃣ It uses pre-trained diffusion models to enable precise and high-fidelity object swapping in images.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "2️⃣Targeted variable swapping ensures perfect background preservation while swapping specific areas. ", "raw": "2️⃣Targeted variable swapping ensures perfect background preservation while swapping specific areas. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3️⃣SwapAnything achieves good results in single-object, multi-object, partial-object, and cross-domain swapping tasks.", "raw": "3️⃣SwapAnything achieves good results in single-object, multi-object, partial-object, and cross-domain swapping tasks.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Paper: ", "raw": "Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.05717", "href": null, "resource": { "type": "paper", "id": "2404.05717", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.05717", "code": null, "user": null, "label": "SwapAnything: Enabling Arbitrary Object Swapping in Personalized Visual\n Editing (2404.05717)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Project page: ", "raw": "Project page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://swap-anything.github.io", "href": "https://swap-anything.github.io", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Congrats to the authors for their work!", "raw": "Congrats to the authors for their work!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "👍", "users": [ "bsenftner", "samusenps" ], "count": 2 } ]

[]

[ { "type": "text", "value": "Thrilled to unveil DS-MoE: a dense training and sparse inference scheme for enhanced computational and memory efficiency in your MoE models! 🚀🚀🚀", "raw": "Thrilled to unveil DS-MoE: a dense training and sparse inference scheme for enhanced computational and memory efficiency in your MoE models! 🚀🚀🚀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Discover more in our blog: ", "raw": "Discover more in our blog: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/bpan/ds-moe", "href": "https://huggingface.co/blog/bpan/ds-moe", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " and dive into the details with our paper: ", "raw": " and dive into the details with our paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.05567", "href": null, "resource": { "type": "paper", "id": "2404.05567", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.05567", "code": null, "user": null, "label": "Dense Training, Sparse Inference: Rethinking Training of\n Mixture-of-Experts Language Models (2404.05567)", "lang": null } ]

[]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "umuthopeyildirim", "zoratrox", "osanseviero", "victor", "ybelkada", "mayank-mishra" ], "count": 7 }, { "reaction": "🚀", "users": [ "samusenps", "ybelkada" ], "count": 2 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/5f17f0a0925b9863e28ad517/X7QKoiXbUtEZSG9jyvfk3.jpeg", "fullname": "Victor Mustar", "name": "victor", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 2607, "isFollowing": false } ]

[ { "type": "text", "value": "More open research updates 🧵", "raw": "More open research updates 🧵", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Performance estimation is currently the best way to quantify the impact of data drift on model performance. 💡", "raw": "Performance estimation is currently the best way to quantify the impact of data drift on model performance. 💡", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I've been benchmarking performance estimation methods (CBPE and M-CBPE) against data drift signals.", "raw": "I've been benchmarking performance estimation methods (CBPE and M-CBPE) against data drift signals.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I'm using drift results as features for many regression algorithms, and then I'm taking those to estimate the model's performance. Finally, I'm measuring the Mean Absolute Error (MAE) between the regression models' predictions and actual performance.", "raw": "I'm using drift results as features for many regression algorithms, and then I'm taking those to estimate the model's performance. Finally, I'm measuring the Mean Absolute Error (MAE) between the regression models' predictions and actual performance.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "So far, for all my experiments, performance estimation methods do better than drift signals. 👨‍🔬", "raw": "So far, for all my experiments, performance estimation methods do better than drift signals. 👨‍🔬", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Bear in mind that these are some early results, I'm running the flow on more datasets as we speak.", "raw": "Bear in mind that these are some early results, I'm running the flow on more datasets as we speak.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Hopefully, by next week, I will have more results to share 👀", "raw": "Hopefully, by next week, I will have more results to share 👀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/629a173153a72d997d3f57d0/xU7e2gedRifEyyNsoTIia.jpeg" } ]

[]

[ { "reaction": "❤️", "users": [ "afrideva" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Ferret-UI", "raw": "Ferret-UI", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Grounded Mobile UI Understanding with Multimodal LLMs", "raw": "Grounded Mobile UI Understanding with Multimodal LLMs", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.05719", "href": null, "resource": { "type": "paper", "id": "2404.05719", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.05719", "code": null, "user": null, "label": "Ferret-UI: Grounded Mobile UI Understanding with Multimodal LLMs (2404.05719)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Recent advancements in multimodal large language models (MLLMs) have been noteworthy, yet, these general-domain MLLMs often fall short in their ability to comprehend and interact effectively with user interface (UI) screens. In this paper, we present Ferret-UI, a new MLLM tailored for enhanced understanding of mobile UI screens, equipped with referring, grounding, and reasoning capabilities. Given that UI screens typically exhibit a more elongated aspect ratio and contain smaller objects of interest (e.g., icons, texts) than natural images, we incorporate \"any resolution\" on top of Ferret to magnify details and leverage enhanced visual features. Specifically, each screen is divided into 2 sub-images based on the original aspect ratio (i.e., horizontal division for portrait screens and vertical division for landscape screens). Both sub-images are encoded separately before being sent to LLMs. We meticulously gather training samples from an extensive range of elementary UI tasks, such as icon recognition, find text, and widget listing. These samples are formatted for instruction-following with region annotations to facilitate precise referring and grounding. To augment the model's reasoning ability, we further compile a dataset for advanced tasks, including detailed description, perception/interaction conversations, and function inference. After training on the curated datasets, Ferret-UI exhibits outstanding comprehension of UI screens and the capability to execute open-ended instructions. For model evaluation, we establish a comprehensive benchmark encompassing all the aforementioned tasks. Ferret-UI excels not only beyond most open-source UI MLLMs, but also surpasses GPT-4V on all the elementary UI tasks.", "raw": "Recent advancements in multimodal large language models (MLLMs) have been noteworthy, yet, these general-domain MLLMs often fall short in their ability to comprehend and interact effectively with user interface (UI) screens. In this paper, we present Ferret-UI, a new MLLM tailored for enhanced understanding of mobile UI screens, equipped with referring, grounding, and reasoning capabilities. Given that UI screens typically exhibit a more elongated aspect ratio and contain smaller objects of interest (e.g., icons, texts) than natural images, we incorporate \"any resolution\" on top of Ferret to magnify details and leverage enhanced visual features. Specifically, each screen is divided into 2 sub-images based on the original aspect ratio (i.e., horizontal division for portrait screens and vertical division for landscape screens). Both sub-images are encoded separately before being sent to LLMs. We meticulously gather training samples from an extensive range of elementary UI tasks, such as icon recognition, find text, and widget listing. These samples are formatted for instruction-following with region annotations to facilitate precise referring and grounding. To augment the model's reasoning ability, we further compile a dataset for advanced tasks, including detailed description, perception/interaction conversations, and function inference. After training on the curated datasets, Ferret-UI exhibits outstanding comprehension of UI screens and the capability to execute open-ended instructions. For model evaluation, we establish a comprehensive benchmark encompassing all the aforementioned tasks. Ferret-UI excels not only beyond most open-source UI MLLMs, but also surpasses GPT-4V on all the elementary UI tasks.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60f1abe7544c2adfd699860c/kmiT-4tulH422lSFB9_49.png" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "keenyou", "fscheffczyk", "KvrParaskevi" ], "count": 4 }, { "reaction": "🤗", "users": [ "keenyou" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Very excited to share the first two official Gemma variants from Google! Today at Google Cloud Next, we announced cutting-edge models for code and research!", "raw": "Very excited to share the first two official Gemma variants from Google! Today at Google Cloud Next, we announced cutting-edge models for code and research!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "First, ", "raw": "First, ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/google/codegemma-release-66152ac7b683e2667abdee11", "href": null, "resource": { "type": "collection", "id": "google/codegemma-release-66152ac7b683e2667abdee11", "discussionNum": null }, "url": "https://huggingface.co/collections/google/codegemma-release-66152ac7b683e2667abdee11", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " - a new set of code-focused Gemma models at 2B and 7B, in both pretrained and instruction-tuned variants. These exhibit outstanding performance on academic benchmarks and (in my experience) real-life usage. Read more in the excellent HuggingFace blog: ", "raw": " - a new set of code-focused Gemma models at 2B and 7B, in both pretrained and instruction-tuned variants. These exhibit outstanding performance on academic benchmarks and (in my experience) real-life usage. Read more in the excellent HuggingFace blog: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/codegemma", "href": "https://huggingface.co/blog/codegemma", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Second, (", "raw": "Second, (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/google/recurrentgemma-release-66152cbdd2d6619cb1665b7a", "href": null, "resource": { "type": "collection", "id": "google/recurrentgemma-release-66152cbdd2d6619cb1665b7a", "discussionNum": null }, "url": "https://huggingface.co/collections/google/recurrentgemma-release-66152cbdd2d6619cb1665b7a", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "), which is based on the outstanding Google DeepMind research in Griffin: ", "raw": "), which is based on the outstanding Google DeepMind research in Griffin: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://arxiv.org/abs/2402.19427", "href": "https://arxiv.org/abs/2402.19427", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ". RecurrentGemma is a research variant that enables higher throughput and vastly improved memory usage. We are excited about new architectures, especially in the lightweight Gemma sizes, where innovations like RecurrentGemma can scale modern AI to many more use cases.", "raw": ". RecurrentGemma is a research variant that enables higher throughput and vastly improved memory usage. We are excited about new architectures, especially in the lightweight Gemma sizes, where innovations like RecurrentGemma can scale modern AI to many more use cases.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "For details on the launches of these models, check out our launch blog -- and please do not hesitate to send us feedback. We are excited to see what you build with CodeGemma and RecurrentGemma! ", "raw": "For details on the launches of these models, check out our launch blog -- and please do not hesitate to send us feedback. We are excited to see what you build with CodeGemma and RecurrentGemma! ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Huge thanks to the Hugging Face team for helping ensure that these models work flawlessly in the Hugging Face ecosystem at launch!", "raw": "Huge thanks to the Hugging Face team for helping ensure that these models work flawlessly in the Hugging Face ecosystem at launch!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🚀", "users": [ "andysalerno", "kramp", "rmdhirr", "manvinder", "osanseviero", "reach-vb", "philschmid", "eliebak", "victor", "nsarrazin", "BrigitteTousi", "m-ric", "tomaarsen", "lewtun", "VictorSanh", "andrewrreed", "lysandre", "Wauplin", "julien-c", "mcpotato", "yjernite", "lunarflu", "linoyts", "Tonic", "pcuenq", "Sintayew4", "icaio" ], "count": 27 }, { "reaction": "🔥", "users": [ "osanseviero", "reach-vb", "philschmid", "eliebak", "victor", "nsarrazin", "BrigitteTousi", "m-ric", "tomaarsen", "lewtun", "VictorSanh", "andrewrreed", "lysandre", "Wauplin", "brunatrevelin", "julien-c", "yjernite", "lunarflu", "Tonic", "pcuenq", "Sintayew4", "Blane187", "clefourrier" ], "count": 23 }, { "reaction": "❤️", "users": [ "osanseviero", "reach-vb", "philschmid", "victor", "nsarrazin", "BrigitteTousi", "tomaarsen", "lewtun", "andrewrreed", "lysandre", "Wauplin", "julien-c", "lunarflu", "Tonic", "pcuenq", "Blane187", "icaio" ], "count": 17 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1655385361868-61b85ce86eb1f2c5e6233736.jpeg", "fullname": "Vaibhav Srivastav", "name": "reach-vb", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 460, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1583857746553-5df7e9e5da6d0311fd3d53f9.jpeg", "fullname": "Thomas Wolf", "name": "thomwolf", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 704, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6317233cc92fd6fee317e030/cJHSvvimr1kqgQfHOjO5n.png", "fullname": "Tom Aarsen", "name": "tomaarsen", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1060, "isFollowing": false } ]

[ { "type": "text", "value": "🎉 Super ", "raw": "🎉 Super ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DeepMount00", "href": null, "resource": null, "url": null, "code": null, "user": "DeepMount00", "label": null, "lang": null }, { "type": "text", "value": " just released 𝗚𝗲𝗺𝗺𝗮_𝗤𝗔_𝗜𝗧𝗔_𝘃𝟯 𝗹𝗲𝗮𝗱𝗶𝗻𝗴 the 𝗥𝗔𝗚 𝘁𝗮𝘀𝗸 on the Italian 𝗟𝗟𝗠_𝗜𝗧𝗔_𝗟𝗘𝗔𝗗𝗘𝗥𝗕𝗢𝗔𝗥𝗗. The model is a fine tuned version of Gemma 2B.", "raw": " just released 𝗚𝗲𝗺𝗺𝗮_𝗤𝗔_𝗜𝗧𝗔_𝘃𝟯 𝗹𝗲𝗮𝗱𝗶𝗻𝗴 the 𝗥𝗔𝗚 𝘁𝗮𝘀𝗸 on the Italian 𝗟𝗟𝗠_𝗜𝗧𝗔_𝗟𝗘𝗔𝗗𝗘𝗥𝗕𝗢𝗔𝗥𝗗. The model is a fine tuned version of Gemma 2B.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Model details: ", "raw": "Model details: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/DeepMount00/Gemma_QA_ITA_v3", "href": null, "resource": { "type": "model", "id": "DeepMount00/Gemma_QA_ITA_v3", "discussionNum": null }, "url": "https://huggingface.co/DeepMount00/Gemma_QA_ITA_v3", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Explore the full RAG section rankings here: ", "raw": "Explore the full RAG section rankings here: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/spaces/FinancialSupport/open_ita_llm_leaderboard", "href": "https://huggingface.co/spaces/FinancialSupport/open_ita_llm_leaderboard", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " on section Classifica RAG", "raw": " on section Classifica RAG", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64f1bf6a8b550e875926a590/xdZHPQGdI2jISWcKhWTMQ.png", "fullname": "Michele Montebovi", "name": "DeepMount00", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 123 } ]

[ { "reaction": "❤️", "users": [ "trisfromgoogle", "DeepMount00", "dillfrescott", "ashah0052", "AtAndDev", "giux78", "clefourrier", "samusenps" ], "count": 8 } ]

[]

[ { "type": "text", "value": "🔁 AutoMerger created the best 7B model on the Open LLM Leaderboard", "raw": "🔁 AutoMerger created the best 7B model on the Open LLM Leaderboard", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "By randomly combining top models from the Open LLM Leaderboard, AutoMerger created YamshadowExperiment28-7B. The model is three weeks old and has been at the top of the leaderboard for a week now. It was created through a simple SLERP merge of:", "raw": "By randomly combining top models from the Open LLM Leaderboard, AutoMerger created YamshadowExperiment28-7B. The model is three weeks old and has been at the top of the leaderboard for a week now. It was created through a simple SLERP merge of:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- ", "raw": "- ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/automerger/YamShadow-7B", "href": null, "resource": { "type": "model", "id": "automerger/YamShadow-7B", "discussionNum": null }, "url": "https://huggingface.co/automerger/YamShadow-7B", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " (another top model created by AutoMerger)", "raw": " (another top model created by AutoMerger)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- ", "raw": "- ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/yam-peleg/Experiment28-7B", "href": null, "resource": { "type": "model", "id": "yam-peleg/Experiment28-7B", "discussionNum": null }, "url": "https://huggingface.co/yam-peleg/Experiment28-7B", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " (a top model from ", "raw": " (a top model from ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@yam-peleg", "href": null, "resource": null, "url": null, "code": null, "user": "yam-peleg", "label": null, "lang": null }, { "type": "text", "value": ")", "raw": ")", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "1/ On the Open LLM Leaderboard, it managed to outperform the excellent M7-7b model, which has been the #1 7B model for a while now.", "raw": "1/ On the Open LLM Leaderboard, it managed to outperform the excellent M7-7b model, which has been the #1 7B model for a while now.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "2/ On the YALL leaderboard, YamshadowExperiment28-7B is ranked as the 9th best-performing automerge (but note that the scores are very close to each other). Compared to others, it does not perform particularly well on AGIEval or Bigbench.", "raw": "2/ On the YALL leaderboard, YamshadowExperiment28-7B is ranked as the 9th best-performing automerge (but note that the scores are very close to each other). Compared to others, it does not perform particularly well on AGIEval or Bigbench.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3/ Thanks to ", "raw": "3/ Thanks to ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@sam-paech", "href": null, "resource": null, "url": null, "code": null, "user": "sam-paech", "label": null, "lang": null }, { "type": "text", "value": " , I have scores on EQ-Bench, where it managed to outperform all of my previous models. It even surpasses recent models such as DBRX instruct, Qwen1.5 32B Chat, and Cohere's Command R+.", "raw": " , I have scores on EQ-Bench, where it managed to outperform all of my previous models. It even surpasses recent models such as DBRX instruct, Qwen1.5 32B Chat, and Cohere's Command R+.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Surprisingly, it does not support ChatML or Mistral Instruct, unlike my other merges (which are part of its family tree). Alpaca works well 99% of the time, but the model can sometimes produce a lot of \"INST\" tokens for no reason.", "raw": "Surprisingly, it does not support ChatML or Mistral Instruct, unlike my other merges (which are part of its family tree). Alpaca works well 99% of the time, but the model can sometimes produce a lot of \"INST\" tokens for no reason.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "In my experiments, YamshadowExperiment28-7B doesn't seem smarter than other successful merges like AlphaMonarch. On the contrary, I found several mathematical or reasoning problems where it fails.", "raw": "In my experiments, YamshadowExperiment28-7B doesn't seem smarter than other successful merges like AlphaMonarch. On the contrary, I found several mathematical or reasoning problems where it fails.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Considering these results, it looks like it might overfit the Open LLM Leaderboard. I guess it's anything but surprising when you randomly merge 156 models.", "raw": "Considering these results, it looks like it might overfit the Open LLM Leaderboard. I guess it's anything but surprising when you randomly merge 156 models.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤗 Model: ", "raw": "🤗 Model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/automerger/YamshadowExperiment28-7B", "href": null, "resource": { "type": "model", "id": "automerger/YamshadowExperiment28-7B", "discussionNum": null }, "url": "https://huggingface.co/automerger/YamshadowExperiment28-7B", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔁 AutoMerger: ", "raw": "🔁 AutoMerger: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/mlabonne/AutoMerger", "href": null, "resource": { "type": "space", "id": "mlabonne/AutoMerger", "discussionNum": null }, "url": "https://huggingface.co/spaces/mlabonne/AutoMerger", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/61b8e2ba285851687028d395/2SJcr8-iq17_A6X-C9g0J.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/65ad56b4c2eef2ba1154618c/CRD6nLAX-34Dmqn1D4xbg.png", "fullname": "Sam Paech", "name": "sam-paech", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 40 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/626bccd70b0b7652fd3100f0/aNNzWbvdTvIhRMa0dtKTZ.jpeg", "fullname": "Yam", "name": "yam-peleg", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 184 } ]

[ { "reaction": "🔥", "users": [ "victor", "AtAndDev", "Tonic", "vladbogo", "rmdhirr", "samusenps", "Gargaz", "osanseviero", "kgourgou", "Kukedlc", "Ramikan-BR" ], "count": 11 }, { "reaction": "👍", "users": [ "neeravkumar", "Heng666", "Gargaz", "osanseviero", "mylesgoose", "kgourgou", "eliolio", "Muhammad2003", "Ramikan-BR" ], "count": 9 }, { "reaction": "🚀", "users": [ "Ramikan-BR" ], "count": 1 }, { "reaction": "👀", "users": [ "Ramikan-BR" ], "count": 1 }, { "reaction": "❤️", "users": [ "Ramikan-BR" ], "count": 1 } ]

[]

[ { "type": "text", "value": "🔍 Today's pick in Interpretability & Analysis of LMs: Do language models plan ahead for future tokens? by W. Wu ", "raw": "🔍 Today's pick in Interpretability & Analysis of LMs: Do language models plan ahead for future tokens? by W. Wu ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@jxm", "href": null, "resource": null, "url": null, "code": null, "user": "jxm", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@lionellevine", "href": null, "resource": null, "url": null, "code": null, "user": "lionellevine", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This work aims to evaluate whether language models exhibit implicit planning during generation.", "raw": "This work aims to evaluate whether language models exhibit implicit planning during generation.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Authors propose two hypotheses that could result in planning-like behavior:
", "raw": "Authors propose two hypotheses that could result in planning-like behavior:
", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Pre-caching: the model engages in computation that is functional to future, but not current, predictions.
", "raw": "- Pre-caching: the model engages in computation that is functional to future, but not current, predictions.
", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Breadcrumbs: Features contributing to the current prediction happen to also be the ones improving future ones.", "raw": "- Breadcrumbs: Features contributing to the current prediction happen to also be the ones improving future ones.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "To validate which behavior is observed in practice, authors note that off-diagonal gradients for weight matrices across the model are the ones responsible for pre-caching, and craft a variant of gradient descent (myopic descent) to remove such terms from the optimization procedure.", "raw": "To validate which behavior is observed in practice, authors note that off-diagonal gradients for weight matrices across the model are the ones responsible for pre-caching, and craft a variant of gradient descent (myopic descent) to remove such terms from the optimization procedure.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Using a synthetic dataset, authors demonstrate that pre-caching occurs in Transformers language models. However, for natural language settings the LM is observed to leverage breadcrumbs from previous passes even in the case of myopic training, rendering the latter hypothesis more plausible to account for model behavior.", "raw": "Using a synthetic dataset, authors demonstrate that pre-caching occurs in Transformers language models. However, for natural language settings the LM is observed to leverage breadcrumbs from previous passes even in the case of myopic training, rendering the latter hypothesis more plausible to account for model behavior.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Paper: ", "raw": "📄 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.00859", "href": null, "resource": { "type": "paper", "id": "2404.00859", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.00859", "code": null, "user": null, "label": "Do language models plan ahead for future tokens? (2404.00859)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 All daily picks: ", "raw": "🔍 All daily picks: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-ofc-lms-65ae3339949c5675d25de2f9", "href": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-ofc-lms-65ae3339949c5675d25de2f9", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/H3OVa0iMrNsAsKGaV14x5.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/HWE443hGoxZmVHZ-xVUO2.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1652114014657-5edc0500968f6028e0559ff8.jpeg", "fullname": "Jack Morris", "name": "jxm", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 29 }, { "avatarUrl": "/avatars/58530a08007f1951ec5a3a277dc962bd.svg", "fullname": "Lionel Levine", "name": "lionellevine", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null } ]

[ { "reaction": "❤️", "users": [ "samusenps", "jxm", "zoratrox", "lordvader31", "NisargBhavsar25" ], "count": 5 } ]

[]

[ { "type": "text", "value": "LegalKit Retrieval, a binary Search with Scalar (int8) Rescoring through French legal codes is now available as a 🤗 Space.", "raw": "LegalKit Retrieval, a binary Search with Scalar (int8) Rescoring through French legal codes is now available as a 🤗 Space.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This process is designed to be memory efficient and fast, with the binary index being small enough to fit in memory and the int8 index being loaded as a view. Additionally, the binary index is much faster (up to 32x) to search than the float32 index, while the rescoring is also extremely efficient.", "raw": "This process is designed to be memory efficient and fast, with the binary index being small enough to fit in memory and the int8 index being loaded as a view. Additionally, the binary index is much faster (up to 32x) to search than the float32 index, while the rescoring is also extremely efficient.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This space also showcases the tsdae-lemone-mbert-base, a sentence embedding model based on BERT fitted using Transformer-based Sequential Denoising Auto-Encoder for unsupervised sentence embedding learning with one objective : french legal domain adaptation.", "raw": "This space also showcases the tsdae-lemone-mbert-base, a sentence embedding model based on BERT fitted using Transformer-based Sequential Denoising Auto-Encoder for unsupervised sentence embedding learning with one objective : french legal domain adaptation.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Link to the 🤗 Space : ", "raw": "Link to the 🤗 Space : ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/louisbrulenaudet/legalkit-retrieval", "href": null, "resource": { "type": "space", "id": "louisbrulenaudet/legalkit-retrieval", "discussionNum": null }, "url": "https://huggingface.co/spaces/louisbrulenaudet/legalkit-retrieval", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Notes:", "raw": "Notes:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The SentenceTransformer model currently in use is in beta and may not be suitable for direct use in production.", "raw": "The SentenceTransformer model currently in use is in beta and may not be suitable for direct use in production.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6459fa0f5b3111fbe83286e1/o0kyKDBQW7e-cxoJk9KGb.jpeg" } ]

[]

[ { "reaction": "❤️", "users": [ "colin-r-carter", "marksmithh", "osanseviero", "tomaarsen", "samusenps", "ethanker" ], "count": 6 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6032802e1f993496bc14d9e3/w6hr-DEQot4VVkoyRIBiy.png", "fullname": "Omar Sanseviero", "name": "osanseviero", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2868, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6317233cc92fd6fee317e030/cJHSvvimr1kqgQfHOjO5n.png", "fullname": "Tom Aarsen", "name": "tomaarsen", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1060, "isFollowing": false } ]

[ { "type": "text", "value": "🚀💃🌟 New Research Alert (Avatars Collection)! 🌟🕺🚀", "raw": "🚀💃🌟 New Research Alert (Avatars Collection)! 🌟🕺🚀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Title: InstructHumans: Editing Animated 3D Human Textures with Instructions", "raw": "📄 Title: InstructHumans: Editing Animated 3D Human Textures with Instructions", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📝 Description: InstructHumans is a novel framework for text-instructed editing of 3D human textures that employs a modified Score Distillation Sampling (SDS-E) method along with spatial smoothness regularization and gradient-based viewpoint sampling to achieve high-quality, consistent, and instruction-true edits.", "raw": "📝 Description: InstructHumans is a novel framework for text-instructed editing of 3D human textures that employs a modified Score Distillation Sampling (SDS-E) method along with spatial smoothness regularization and gradient-based viewpoint sampling to achieve high-quality, consistent, and instruction-true edits.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👥 Authors: Jiayin Zhu, Linlin Yang, Angela Yao", "raw": "👥 Authors: Jiayin Zhu, Linlin Yang, Angela Yao", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔗 Paper: ", "raw": "🔗 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.04037", "href": null, "resource": { "type": "paper", "id": "2404.04037", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.04037", "code": null, "user": null, "label": "InstructHumans: Editing Animated 3D Human Textures with Instructions (2404.04037)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌐 Web Page: ", "raw": "🌐 Web Page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://jyzhu.top/instruct-humans", "href": "https://jyzhu.top/instruct-humans", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📁 Repository: ", "raw": "📁 Repository: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/viridityzhu/InstructHumans", "href": "https://github.com/viridityzhu/InstructHumans", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "raw": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "href": null, "resource": { "type": "space", "id": "DmitryRyumin/NewEraAI-Papers", "discussionNum": null }, "url": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " curated by ", "raw": " curated by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DmitryRyumin", "href": null, "resource": null, "url": null, "code": null, "user": "DmitryRyumin", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Added to the Avatars Collection: ", "raw": "🚀 Added to the Avatars Collection: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "href": null, "resource": { "type": "collection", "id": "DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "discussionNum": null }, "url": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Keywords: #InstructHumans #3DTextureEditing #TextInstructions #ScoreDistillationSampling #SDS-E #SpatialSmoothnessRegularization #3DEditing #AvatarEditing #DeepLearning #Innovation", "raw": "🔍 Keywords: #InstructHumans #3DTextureEditing #TextInstructions #ScoreDistillationSampling #SDS-E #SpatialSmoothnessRegularization #3DEditing #AvatarEditing #DeepLearning #Innovation", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/lwwQLGKy55F_OE478nPy2.gif" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/JTY1GgAvJAExLjZdicPeB.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/2bBaTjMh340N3NHtbEaap.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/-z80_RVdxCaikXmVVjiPo.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/oRa9hdNiQqblh_PPiH-Ij.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/qEgBml6gwSiunJnhMc6i_.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/wak4DHY66IsA_n_u7UNTh.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/aY3LkTZJsNs3KAlfzsdJL.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/nRCxbVng_PPBqKd-Z3KVc.jpeg", "fullname": "Dmitry Ryumin", "name": "DmitryRyumin", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 377 } ]

[ { "reaction": "👍", "users": [ "DmitryRyumin", "xpshn", "samusenps", "tonyassi" ], "count": 4 }, { "reaction": "🔥", "users": [ "samusenps", "impactframes" ], "count": 2 }, { "reaction": "🚀", "users": [ "samusenps" ], "count": 1 }, { "reaction": "😎", "users": [ "eddie292" ], "count": 1 } ]

[]

[ { "type": "text", "value": "We just released gradio version 4.26.0 ! We *highly* recommend you upgrade your apps to this version to bring in these nice changes:", "raw": "We just released gradio version 4.26.0 ! We *highly* recommend you upgrade your apps to this version to bring in these nice changes:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🎥 Introducing the API recorder. Any gradio app running 4.26.0 and above will have an \"API Recorder\" that will record your interactions with the app and auto-generate the corresponding python or js code needed to recreate those actions programmatically. It's very neat!", "raw": "🎥 Introducing the API recorder. Any gradio app running 4.26.0 and above will have an \"API Recorder\" that will record your interactions with the app and auto-generate the corresponding python or js code needed to recreate those actions programmatically. It's very neat!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📝 Enhanced markdown rendering in gr.Chatbot", "raw": "📝 Enhanced markdown rendering in gr.Chatbot", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🐢 Fix for slow load times on spaces as well as the UI locking up on rapid generations", "raw": "🐢 Fix for slow load times on spaces as well as the UI locking up on rapid generations", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "See the full changelog of goodies here: ", "raw": "See the full changelog of goodies here: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://www.gradio.app/changelog#4-26-0", "href": "https://www.gradio.app/changelog#4-26-0", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "DmitryRyumin", "dillfrescott", "gsarti", "osanseviero", "aceeee", "ysharma", "on1onmangoes" ], "count": 8 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1654278567459-626a9bfa03e2e2796f24ca11.jpeg", "fullname": "Freddy Boulton", "name": "freddyaboulton", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 164, "isFollowing": false } ]

[ { "type": "text", "value": "Fun fact about evaluation, part 2!", "raw": "Fun fact about evaluation, part 2!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "How much do scores change depending on prompt format choice?", "raw": "How much do scores change depending on prompt format choice?", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Using different prompts (all present in the literature, from ", "raw": "Using different prompts (all present in the literature, from ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "inline_code", "value": null, "raw": "`Prompt question?`", "href": null, "resource": null, "url": null, "code": "Prompt question?", "user": null, "label": null, "lang": null }, { "type": "text", "value": " to ", "raw": " to ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "inline_code", "value": null, "raw": "`Question: prompt question?\\nChoices: enumeration of all choices\\nAnswer: `", "href": null, "resource": null, "url": null, "code": "Question: prompt question?\\nChoices: enumeration of all choices\\nAnswer: ", "user": null, "label": null, "lang": null }, { "type": "text", "value": "), we get a score range of...", "raw": "), we get a score range of...", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "10 points for a single model! ", "raw": "10 points for a single model! ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Keep in mind that we only changed the prompt, not the evaluation subsets, etc.", "raw": "Keep in mind that we only changed the prompt, not the evaluation subsets, etc.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Again, this confirms that evaluation results reported without their details are basically bullshit.", "raw": "Again, this confirms that evaluation results reported without their details are basically bullshit.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Prompt format on the x axis, all these evals look at the logprob of either \"choice A/choice B...\" or \"A/B...\".", "raw": "Prompt format on the x axis, all these evals look at the logprob of either \"choice A/choice B...\" or \"A/B...\".", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Incidentally, it also changes model rankings - so a \"best\" model might only be best on one type of prompt... ", "raw": "Incidentally, it also changes model rankings - so a \"best\" model might only be best on one type of prompt... ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6202a599216215a22221dea9/TnblLIUyIpvLZnFHMoKdy.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6202a599216215a22221dea9/C9iI8MWUBfAFLhOcDPiQJ.png" } ]

[]

[ { "reaction": "❤️", "users": [ "eduagarcia", "SivilTaram", "andysalerno", "samusenps", "osanseviero", "EricNarro" ], "count": 6 }, { "reaction": "🔥", "users": [ "ancs21", "dark-pen" ], "count": 2 }, { "reaction": "🤯", "users": [ "seth-zou", "jaahas" ], "count": 2 } ]

[]

[ { "type": "text", "value": "text-generation-inference (TGI) is now fully open-source again!", "raw": "text-generation-inference (TGI) is now fully open-source again!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Along with text-embeddings-inference.", "raw": "Along with text-embeddings-inference.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We just switched both of those repos' license back to Apache 2. 🔥", "raw": "We just switched both of those repos' license back to Apache 2. 🔥", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5dd96eb166059660ed1ee413/3MFJ3Ws30B9WK3oiufbvw.png" } ]

[]

[ { "reaction": "❤️", "users": [ "clefourrier", "Saugatkafley", "samusenps", "jcudit", "flozi00", "osanseviero", "joennlae", "Joseph717171", "clem", "freddyaboulton" ], "count": 10 }, { "reaction": "👍", "users": [ "sugatoray", "gsarti", "kramp", "osanseviero", "Joseph717171", "clem", "timestop1" ], "count": 7 }, { "reaction": "🚀", "users": [ "Sylvestre", "clefourrier", "ehcalabres", "osanseviero", "Joseph717171", "clem" ], "count": 6 }, { "reaction": "🔥", "users": [ "seanmiranda" ], "count": 1 } ]

[]

[ { "type": "text", "value": "No \"Zero-Shot\" Without Exponential Data", "raw": "No \"Zero-Shot\" Without Exponential Data", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Pretraining Concept Frequency Determines Multimodal Model Performance", "raw": "Pretraining Concept Frequency Determines Multimodal Model Performance", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.04125", "href": null, "resource": { "type": "paper", "id": "2404.04125", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.04125", "code": null, "user": null, "label": "No \"Zero-Shot\" Without Exponential Data: Pretraining Concept Frequency\n Determines Multimodal Model Performance (2404.04125)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Web-crawled pretraining datasets underlie the impressive \"zero-shot\" evaluation performance of multimodal models, such as CLIP for classification/retrieval and Stable-Diffusion for image generation. However, it is unclear how meaningful the notion of \"zero-shot\" generalization is for such multimodal models, as it is not known to what extent their pretraining datasets encompass the downstream concepts targeted for during \"zero-shot\" evaluation. In this work, we ask: How is the performance of multimodal models on downstream concepts influenced by the frequency of these concepts in their pretraining datasets? We comprehensively investigate this question across 34 models and five standard pretraining datasets (CC-3M, CC-12M, YFCC-15M, LAION-400M, LAION-Aesthetics), generating over 300GB of data artifacts. We consistently find that, far from exhibiting \"zero-shot\" generalization, multimodal models require exponentially more data to achieve linear improvements in downstream \"zero-shot\" performance, following a sample inefficient log-linear scaling trend. This trend persists even when controlling for sample-level similarity between pretraining and downstream datasets, and testing on purely synthetic data distributions. Furthermore, upon benchmarking models on long-tailed data sampled based on our analysis, we demonstrate that multimodal models across the board perform poorly. We contribute this long-tail test set as the \"Let it Wag!\" benchmark to further research in this direction. Taken together, our study reveals an exponential need for training data which implies that the key to \"zero-shot\" generalization capabilities under large-scale training paradigms remains to be found.", "raw": "Web-crawled pretraining datasets underlie the impressive \"zero-shot\" evaluation performance of multimodal models, such as CLIP for classification/retrieval and Stable-Diffusion for image generation. However, it is unclear how meaningful the notion of \"zero-shot\" generalization is for such multimodal models, as it is not known to what extent their pretraining datasets encompass the downstream concepts targeted for during \"zero-shot\" evaluation. In this work, we ask: How is the performance of multimodal models on downstream concepts influenced by the frequency of these concepts in their pretraining datasets? We comprehensively investigate this question across 34 models and five standard pretraining datasets (CC-3M, CC-12M, YFCC-15M, LAION-400M, LAION-Aesthetics), generating over 300GB of data artifacts. We consistently find that, far from exhibiting \"zero-shot\" generalization, multimodal models require exponentially more data to achieve linear improvements in downstream \"zero-shot\" performance, following a sample inefficient log-linear scaling trend. This trend persists even when controlling for sample-level similarity between pretraining and downstream datasets, and testing on purely synthetic data distributions. Furthermore, upon benchmarking models on long-tailed data sampled based on our analysis, we demonstrate that multimodal models across the board perform poorly. We contribute this long-tail test set as the \"Let it Wag!\" benchmark to further research in this direction. Taken together, our study reveals an exponential need for training data which implies that the key to \"zero-shot\" generalization capabilities under large-scale training paradigms remains to be found.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60f1abe7544c2adfd699860c/WS8NWFFVRKIS8d1UWiLQu.png" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "vicgalle", "vladbogo", "adhirajghosh", "KvrParaskevi" ], "count": 5 }, { "reaction": "🚀", "users": [ "AmeyaPrabhu", "vladbogo" ], "count": 2 } ]

[]

[ { "type": "text", "value": "✨ Easy Synthetic Dataset File Generation using LLM DataGen ! Link: ", "raw": "✨ Easy Synthetic Dataset File Generation using LLM DataGen ! Link: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/spaces/lhoestq/LLM_DataGen", "href": "https://huggingface.co/spaces/lhoestq/LLM_DataGen", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "features + how it works:", "raw": "features + how it works:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "✍️ Generate the dataset content you want just by entering a file name", "raw": "✍️ Generate the dataset content you want just by entering a file name", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💡 Optionally specify the column names you need", "raw": "💡 Optionally specify the column names you need", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💨 The dataset is streamed and generated on-the-fly in JSON Lines format", "raw": "💨 The dataset is streamed and generated on-the-fly in JSON Lines format", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "✅ Generation is constrained to always output valid JSON", "raw": "✅ Generation is constrained to always output valid JSON", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "How does this work ?", "raw": "How does this work ?", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "1/ Enter a file name", "raw": "1/ Enter a file name", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "2/ The model generates column names for such a file. Using structured generation, it can generate 2 to 5 column names using lower case characters and underscores. I use a prompt that asks to generate column names for a realistic dataset and low temperature.", "raw": "2/ The model generates column names for such a file. Using structured generation, it can generate 2 to 5 column names using lower case characters and underscores. I use a prompt that asks to generate column names for a realistic dataset and low temperature.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3/ The columns are used to update the Finite State Machine for the dataset content structured generation, so that it is used to generate JSON objects using those columns", "raw": "3/ The columns are used to update the Finite State Machine for the dataset content structured generation, so that it is used to generate JSON objects using those columns", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "4/ The model generates JSON objects using structured generation again, using the updated Finite State Machine. I use a prompt that asks for realistic data and a temperature of 1.", "raw": "4/ The model generates JSON objects using structured generation again, using the updated Finite State Machine. I use a prompt that asks for realistic data and a temperature of 1.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> Why update a Finite State Machine instead of re-creating one ?", "raw": "> Why update a Finite State Machine instead of re-creating one ?", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Creating one can take up to 30sec, while updating one takes 0.1s (though it requires to manipulate a graph which is not easy to implement)", "raw": "Creating one can take up to 30sec, while updating one takes 0.1s (though it requires to manipulate a graph which is not easy to implement)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> Batched generation is faster, why not use it ?", "raw": "> Batched generation is faster, why not use it ?", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Generate in batches is faster but tends to generate duplicates for this demo.", "raw": "Generate in batches is faster but tends to generate duplicates for this demo.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Further work can be to provide different prompts (one per sequence in the batch) to end up with a different distribution of sequences in each batch. Or implement a custom sampler that would forbid generating the same data in sequences of the same batch.", "raw": "Further work can be to provide different prompts (one per sequence in the batch) to end up with a different distribution of sequences in each batch. Or implement a custom sampler that would forbid generating the same data in sequences of the same batch.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "> How does structured generation work ?", "raw": "> How does structured generation work ?", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I used the ", "raw": "I used the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "inline_code", "value": null, "raw": "`outlines`", "href": null, "resource": null, "url": null, "code": "outlines", "user": null, "label": null, "lang": null }, { "type": "text", "value": " library with ", "raw": " library with ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "inline_code", "value": null, "raw": "`transformers`", "href": null, "resource": null, "url": null, "code": "transformers", "user": null, "label": null, "lang": null }, { "type": "text", "value": " to to define a JSON schema that the generation has to follow. It uses a Finite State Machine with ", "raw": " to to define a JSON schema that the generation has to follow. It uses a Finite State Machine with ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "inline_code", "value": null, "raw": "`token_id`", "href": null, "resource": null, "url": null, "code": "token_id", "user": null, "label": null, "lang": null }, { "type": "text", "value": " as transitions.", "raw": " as transitions.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Let me know what you think ! And feel free to duplicate/modify it to try other ", "raw": "Let me know what you think ! And feel free to duplicate/modify it to try other ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "models/prompts", "raw": "models/prompts", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " or sampling methods :)", "raw": " or sampling methods :)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e9ecfc04957053f60648a3e/RfhtBaHy-7UPWKWc-8In8.png" } ]

[]

[ { "reaction": "🚀", "users": [ "victor", "julien-c", "samusenps", "Azamat1k", "osanseviero", "Taylor658", "awacke1" ], "count": 7 }, { "reaction": "🧠", "users": [ "Sylvestre", "musfiqdehan" ], "count": 2 }, { "reaction": "❤️", "users": [ "samusenps" ], "count": 1 }, { "reaction": "➕", "users": [ "srisree" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Hello world!", "raw": "Hello world!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I'd like to share with you all today some specific Research about the Brain & some surface level thoughts", "raw": "I'd like to share with you all today some specific Research about the Brain & some surface level thoughts", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Music (Frequencies)", "raw": "Music (Frequencies)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2307.11078", "href": null, "resource": { "type": "paper", "id": "2307.11078", "discussionNum": null }, "url": "https://huggingface.co/papers/2307.11078", "code": null, "user": null, "label": "Brain2Music: Reconstructing Music from Human Brain Activity (2307.11078)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Speech", "raw": "Speech", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2208.12266", "href": null, "resource": { "type": "paper", "id": "2208.12266", "discussionNum": null }, "url": "https://huggingface.co/papers/2208.12266", "code": null, "user": null, "label": "Decoding speech from non-invasive brain recordings (2208.12266)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Image ", "raw": "Image ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2308.02510", "href": null, "resource": { "type": "paper", "id": "2308.02510", "discussionNum": null }, "url": "https://huggingface.co/papers/2308.02510", "code": null, "user": null, "label": "Seeing through the Brain: Image Reconstruction of Visual Perception from\n Human Brain Signals (2308.02510)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2306.16934", "href": null, "resource": { "type": "paper", "id": "2306.16934", "discussionNum": null }, "url": "https://huggingface.co/papers/2306.16934", "code": null, "user": null, "label": "DreamDiffusion: Generating High-Quality Images from Brain EEG Signals (2306.16934)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2403.18211", "href": null, "resource": { "type": "paper", "id": "2403.18211", "discussionNum": null }, "url": "https://huggingface.co/papers/2403.18211", "code": null, "user": null, "label": "NeuroPictor: Refining fMRI-to-Image Reconstruction via Multi-individual\n Pretraining and Multi-level Modulation (2403.18211)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Video", "raw": "Video", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2402.01590", "href": null, "resource": { "type": "paper", "id": "2402.01590", "discussionNum": null }, "url": "https://huggingface.co/papers/2402.01590", "code": null, "user": null, "label": "NeuroCine: Decoding Vivid Video Sequences from Human Brain Activties (2402.01590)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2305.11675", "href": null, "resource": { "type": "paper", "id": "2305.11675", "discussionNum": null }, "url": "https://huggingface.co/papers/2305.11675", "code": null, "user": null, "label": "Cinematic Mindscapes: High-quality Video Reconstruction from Brain\n Activity (2305.11675)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3D", "raw": "3D", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2312.07485", "href": null, "resource": { "type": "paper", "id": "2312.07485", "discussionNum": null }, "url": "https://huggingface.co/papers/2312.07485", "code": null, "user": null, "label": "MinD-3D: Reconstruct High-quality 3D objects in Human Brain (2312.07485)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Potential Opportunities For BCI", "raw": "Potential Opportunities For BCI", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "4D ", "raw": "4D ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2312.17142", "href": null, "resource": { "type": "paper", "id": "2312.17142", "discussionNum": null }, "url": "https://huggingface.co/papers/2312.17142", "code": null, "user": null, "label": "DreamGaussian4D: Generative 4D Gaussian Splatting (2312.17142)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Realistic High Quality Video ", "raw": "Realistic High Quality Video ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2402.17177", "href": null, "resource": { "type": "paper", "id": "2402.17177", "discussionNum": null }, "url": "https://huggingface.co/papers/2402.17177", "code": null, "user": null, "label": "Sora: A Review on Background, Technology, Limitations, and Opportunities\n of Large Vision Models (2402.17177)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/samusenps/bci-661206b642da659656474db2", "href": null, "resource": { "type": "collection", "id": "samusenps/bci-661206b642da659656474db2", "discussionNum": null }, "url": "https://huggingface.co/collections/samusenps/bci-661206b642da659656474db2", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Reading minds is cool & useful but could be utilized for many things other than thought interrogation 👀", "raw": "Reading minds is cool & useful but could be utilized for many things other than thought interrogation 👀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Current Co pilots are Boring! & have much untapped potential & also people don't seem to want autonomous agents replacing them although it is inevitable for some cases", "raw": "Current Co pilots are Boring! & have much untapped potential & also people don't seem to want autonomous agents replacing them although it is inevitable for some cases", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I believe if humans want to become an interplanetary species that can utilize our accumulative research we need to extend our brain with technology in order to be smarter. Imagine a Co-pilot for your head, Adding extra ‘RAM’ to the brain, or even Processing external data within the brain.", "raw": "I believe if humans want to become an interplanetary species that can utilize our accumulative research we need to extend our brain with technology in order to be smarter. Imagine a Co-pilot for your head, Adding extra ‘RAM’ to the brain, or even Processing external data within the brain.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Ok people are afraid of implanting computer chips within their brains, what if someone hacks it ? , the invasive possibilities are crazy!", "raw": "Ok people are afraid of implanting computer chips within their brains, what if someone hacks it ? , the invasive possibilities are crazy!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "How can we ensure Safety & Interpretability in brain computer interfaces", "raw": "How can we ensure Safety & Interpretability in brain computer interfaces", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "1. External Non Invasive Brain Computer interfaces [ think similar to ", "raw": "1. External Non Invasive Brain Computer interfaces [ think similar to ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://neurosity.co/crown", "href": "https://neurosity.co/crown", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " (overpriced IMO & Hardware is closed source proprietary, who knows what they’re doing 👁️) ] ", "raw": " (overpriced IMO & Hardware is closed source proprietary, who knows what they’re doing 👁️) ] ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "2. Full Reproducible Open-Source Stack Brain computer interface down to the hardware, operating system, and application level.", "raw": "2. Full Reproducible Open-Source Stack Brain computer interface down to the hardware, operating system, and application level.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "3. Maybe you can't, there may always be a risk of danger, though not as consequential with 1 & 2 ", "raw": "3. Maybe you can't, there may always be a risk of danger, though not as consequential with 1 & 2 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6538119803519fddb4a17e10/Jc8Pnp2bEAhT1D1zfO3SZ.png" } ]

[]

[ { "reaction": "🧠", "users": [ "samusenps", "INMYEYES", "DmitryRyumin", "Jigya", "gsarti", "diwank", "osanseviero", "lunarflu", "victor", "medmac01", "not-lain", "s3nh", "efecelik", "samusenpsismydaddy" ], "count": 14 }, { "reaction": "❤️", "users": [ "digiplay", "osanseviero", "flyineye", "lunarflu", "medmac01", "not-lain", "s3nh", "samusenpsismydaddy" ], "count": 8 }, { "reaction": "➕", "users": [ "samusenps", "Jigya", "osanseviero", "lunarflu" ], "count": 4 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/646c83c871d0c8a6e4455854/b5Act86SZ7zuUXmLOIqYj.png", "fullname": "digiplay", "name": "digiplay", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 166, "isFollowing": false } ]

[ { "type": "text", "value": "🚀🕺🌟 New Research Alert - CVPR 2024 (Avatars Collection)! 🌟💃🚀", "raw": "🚀🕺🌟 New Research Alert - CVPR 2024 (Avatars Collection)! 🌟💃🚀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Title: 3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting 🔝", "raw": "📄 Title: 3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting 🔝", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📝 Description: 3DGS-Avatar is a novel method for creating animatable human avatars from monocular videos using 3D Gaussian Splatting (3DGS). By using a non-rigid deformation network and as-isometric-as-possible regularizations, the method achieves comparable or better performance than SOTA methods while being 400x faster in training and 250x faster in inference, allowing real-time rendering at 50+ FPS.", "raw": "📝 Description: 3DGS-Avatar is a novel method for creating animatable human avatars from monocular videos using 3D Gaussian Splatting (3DGS). By using a non-rigid deformation network and as-isometric-as-possible regularizations, the method achieves comparable or better performance than SOTA methods while being 400x faster in training and 250x faster in inference, allowing real-time rendering at 50+ FPS.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👥 Authors: Zhiyin Qian, Shaofei Wang, Marko Mihajlovic, Andreas Geiger, Siyu Tang", "raw": "👥 Authors: Zhiyin Qian, Shaofei Wang, Marko Mihajlovic, Andreas Geiger, Siyu Tang", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📅 Conference: CVPR, Jun 17-21, 2024 | Seattle WA, USA 🇺🇸", "raw": "📅 Conference: CVPR, Jun 17-21, 2024 | Seattle WA, USA 🇺🇸", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔗 Paper: ", "raw": "🔗 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2312.09228", "href": null, "resource": { "type": "paper", "id": "2312.09228", "discussionNum": null }, "url": "https://huggingface.co/papers/2312.09228", "code": null, "user": null, "label": "3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting (2312.09228)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌐 Github Page: ", "raw": "🌐 Github Page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://neuralbodies.github.io/3DGS-Avatar/", "href": "https://neuralbodies.github.io/3DGS-Avatar/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📁 Repository: ", "raw": "📁 Repository: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/mikeqzy/3dgs-avatar-release", "href": "https://github.com/mikeqzy/3dgs-avatar-release", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📺 Video: ", "raw": "📺 Video: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://www.youtube.com/watch?v=FJ29U9OkmmU", "href": "https://www.youtube.com/watch?v=FJ29U9OkmmU", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "raw": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "href": null, "resource": { "type": "space", "id": "DmitryRyumin/NewEraAI-Papers", "discussionNum": null }, "url": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " curated by ", "raw": " curated by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DmitryRyumin", "href": null, "resource": null, "url": null, "code": null, "user": "DmitryRyumin", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Added to the Avatars Collection: ", "raw": "🚀 Added to the Avatars Collection: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "href": null, "resource": { "type": "collection", "id": "DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "discussionNum": null }, "url": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Keywords: #3DGSAvatar #3DAvatar #3DGaussianSplatting #AnimatableAvatars #MonocularVideo #RealTimeRendering #FastTraining #EfficientInference #CVPR2024 #DeepLearning #Innovation", "raw": "🔍 Keywords: #3DGSAvatar #3DAvatar #3DGaussianSplatting #AnimatableAvatars #MonocularVideo #RealTimeRendering #FastTraining #EfficientInference #CVPR2024 #DeepLearning #Innovation", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/xVYvnGTrB0ekf5RdaOXvW.mp4" }, { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/whCZH5v2tS_efVuuCaKkG.mp4" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/uJJW7AXPgNGUYnrwbpPJL.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/ECnLxeuQibEUdCQ1dnbTY.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/i21qlEQLaEr2wF1MbZuZs.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/-Zn5fRaJ9dRkzvlIApjYf.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/Ip-PFyS0I6BexLXmraqdK.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/L8ZLnnQZ9A97Sn0Mn1_Zf.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/N1wg47YxEoG9KSBMI8Cbh.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/nRCxbVng_PPBqKd-Z3KVc.jpeg", "fullname": "Dmitry Ryumin", "name": "DmitryRyumin", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 377 } ]

[ { "reaction": "🔥", "users": [ "DmitryRyumin", "AlekseiPravdin", "Jigya", "YoungMetal", "yehiaserag", "dillfrescott", "osanseviero", "Lei-bw", "Redcalx34B", "mariasandu", "lognat0704" ], "count": 11 }, { "reaction": "❤️", "users": [ "samusenps", "AlekseiPravdin", "Jigya", "YoungMetal", "dillfrescott", "osanseviero", "Lei-bw", "Redcalx34B", "lognat0704" ], "count": 9 }, { "reaction": "🤗", "users": [ "DmitryRyumin", "Jigya", "YoungMetal", "dillfrescott" ], "count": 4 } ]

[]

[ { "type": "text", "value": "🔍 Today's pick in Interpretability & Analysis of LMs: ReFT: Representation Finetuning for Language Models by ", "raw": "🔍 Today's pick in Interpretability & Analysis of LMs: ReFT: Representation Finetuning for Language Models by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@zhengxuanzenwu", "href": null, "resource": null, "url": null, "code": null, "user": "zhengxuanzenwu", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@aryaman", "href": null, "resource": null, "url": null, "code": null, "user": "aryaman", "label": null, "lang": null }, { "type": "text", "value": " Z. Wang ", "raw": " Z. Wang ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@atticusg", "href": null, "resource": null, "url": null, "code": null, "user": "atticusg", "label": null, "lang": null }, { "type": "text", "value": " D. Jurafsky ", "raw": " D. Jurafsky ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@manning", "href": null, "resource": null, "url": null, "code": null, "user": "manning", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@cgpotts", "href": null, "resource": null, "url": null, "code": null, "user": "cgpotts", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This work introduces Representation fine-tuning (ReFT), a framework using learned inference-time interventions as efficient yet effective alternatives to PEFT weight adaptation. LoReFT, a ReFT variant intervening linearly on a representation subspaces, is evaluated against several PEFT approaches showing SOTA performances across popular benchmark with 10-50x speedup. The 🤗-compatible pyreft library is introduced to simplify ReFT usage.", "raw": "This work introduces Representation fine-tuning (ReFT), a framework using learned inference-time interventions as efficient yet effective alternatives to PEFT weight adaptation. LoReFT, a ReFT variant intervening linearly on a representation subspaces, is evaluated against several PEFT approaches showing SOTA performances across popular benchmark with 10-50x speedup. The 🤗-compatible pyreft library is introduced to simplify ReFT usage.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This is one of the most convincing practical applications of interpretability methods/insights I've seen in recent years, and I'm looking forward to people combining this with methods to disentangle features like SAEs and Backpack LMs for making interventions more interpretable!", "raw": "This is one of the most convincing practical applications of interpretability methods/insights I've seen in recent years, and I'm looking forward to people combining this with methods to disentangle features like SAEs and Backpack LMs for making interventions more interpretable!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Paper: ", "raw": "📄 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.03592", "href": null, "resource": { "type": "paper", "id": "2404.03592", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.03592", "code": null, "user": null, "label": "ReFT: Representation Finetuning for Language Models (2404.03592)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 All daily picks: ", "raw": "🔍 All daily picks: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-of-lms-65ae3339949c5675d25de2f9", "href": null, "resource": { "type": "collection", "id": "gsarti/daily-picks-in-interpretability-and-analysis-of-lms-65ae3339949c5675d25de2f9", "discussionNum": null }, "url": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-of-lms-65ae3339949c5675d25de2f9", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/ZEJJyC-gDkbTaA3MZJtBG.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/RF2YjqwobmNZcalKyVjkH.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/603803ad1b4a9bee818ab78e/CLWFX4lUD0hxJsV7uiiQE.jpeg", "fullname": "Aryaman Arora", "name": "aryaman", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 9 }, { "avatarUrl": "/avatars/e70381850f5657b54e90f5539f3d74eb.svg", "fullname": "Atticus Geiger", "name": "atticusg", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1650485126923-noauth.png", "fullname": "Chris Potts", "name": "cgpotts", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 5 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1628695407761-5f555a1f8bf55658acfed19e.jpeg", "fullname": "Christopher Manning", "name": "manning", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 15 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1650432433568-604fa06d4f9b833fc41c75cf.png", "fullname": "Zhengxuan Wu", "name": "zhengxuanzenwu", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 7 } ]

[ { "reaction": "❤️", "users": [ "samusenps", "aryaman", "Jigya", "zhengxuanzenwu", "diwank", "srisree" ], "count": 6 } ]

[]

[ { "type": "text", "value": "To date, louisbrulenaudet/Maxine-34B-stock is the \"Best 🤝 base merges and moerges model of around 30B\" on the Open LLM Leaderboard ❤️‍🔥", "raw": "To date, louisbrulenaudet/Maxine-34B-stock is the \"Best 🤝 base merges and moerges model of around 30B\" on the Open LLM Leaderboard ❤️‍🔥", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "It is a practical application of the stock method recently implemented by ", "raw": "It is a practical application of the stock method recently implemented by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@arcee-ai", "href": null, "resource": null, "url": null, "code": null, "user": "arcee-ai", "label": null, "lang": null }, { "type": "text", "value": " in the MergeKit :", "raw": " in the MergeKit :", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "code_fence", "value": null, "raw": "```yaml\nmodels:\n - model: ConvexAI/Luminex-34B-v0.2\n - model: fblgit/UNA-34BeagleSimpleMath-32K-v1\nmerge_method: model_stock\nbase_model: abacusai/Smaug-34B-v0.1\ndtype: bfloat16\n```", "href": null, "resource": null, "url": null, "code": "models:\n - model: ConvexAI/Luminex-34B-v0.2\n - model: fblgit/UNA-34BeagleSimpleMath-32K-v1\nmerge_method: model_stock\nbase_model: abacusai/Smaug-34B-v0.1\ndtype: bfloat16", "user": null, "label": null, "lang": "yaml" }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Model : ", "raw": "Model : ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/louisbrulenaudet/Maxine-34B-stock", "href": null, "resource": { "type": "model", "id": "louisbrulenaudet/Maxine-34B-stock", "discussionNum": null }, "url": "https://huggingface.co/louisbrulenaudet/Maxine-34B-stock", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "LLM Leaderboard best models ❤️‍🔥 Collection : ", "raw": "LLM Leaderboard best models ❤️‍🔥 Collection : ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/collections/open-llm-leaderboard/llm-leaderboard-best-models-652d6c7965a4619fb5c27a03", "href": "https://huggingface.co/collections/open-llm-leaderboard/llm-leaderboard-best-models-652d6c7965a4619fb5c27a03", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "dhjang", "ben-wise", "Jigya", "osanseviero", "not-lain" ], "count": 6 } ]

[]

[ { "type": "text", "value": "Releasing H2O-Danube2-1.8b - a new and improved language model being the current top model on the Open LLM Leaderboard.", "raw": "Releasing H2O-Danube2-1.8b - a new and improved language model being the current top model on the Open LLM Leaderboard.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The model is trained on additional 2T tokens (total of 3T tokens) with various data mix stages as a result of extensive experimentation. Focus is on natural language tasks.", "raw": "The model is trained on additional 2T tokens (total of 3T tokens) with various data mix stages as a result of extensive experimentation. Focus is on natural language tasks.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Base weights: ", "raw": "Base weights: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/h2oai/h2o-danube2-1.8b-base", "href": null, "resource": { "type": "model", "id": "h2oai/h2o-danube2-1.8b-base", "discussionNum": null }, "url": "https://huggingface.co/h2oai/h2o-danube2-1.8b-base", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "SFT weights: ", "raw": "SFT weights: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/h2oai/h2o-danube2-1.8b-sft", "href": null, "resource": { "type": "model", "id": "h2oai/h2o-danube2-1.8b-sft", "discussionNum": null }, "url": "https://huggingface.co/h2oai/h2o-danube2-1.8b-sft", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Chat weights: ", "raw": "Chat weights: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/h2oai/h2o-danube2-1.8b-chat", "href": null, "resource": { "type": "model", "id": "h2oai/h2o-danube2-1.8b-chat", "discussionNum": null }, "url": "https://huggingface.co/h2oai/h2o-danube2-1.8b-chat", "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🤯", "users": [ "victor", "thomwolf", "osanseviero" ], "count": 3 }, { "reaction": "🔥", "users": [ "DmitryRyumin", "a9i", "thomwolf" ], "count": 3 }, { "reaction": "❤️", "users": [ "samusenps" ], "count": 1 }, { "reaction": "👍", "users": [ "Jigya" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Honestly i don't understand how come we as the open source community haven't surpassed GPT-4 yet ? Like for me it looks like everything is out there just need to be exploited! Clearly specialized small models outperforms gpt4 on downstream tasks ! So why haven't we just trained a 1B-2B really strong general model and then continue pertained and/or finetuned it on datasets for downstream tasks like math, code...well structured as Textbooks format or other datasets formats that have been proven to be really efficient and good! Ounce you have 100 finetuned model, just wrap them all into a FrankenMoE and Voila ✨ ", "raw": "Honestly i don't understand how come we as the open source community haven't surpassed GPT-4 yet ? Like for me it looks like everything is out there just need to be exploited! Clearly specialized small models outperforms gpt4 on downstream tasks ! So why haven't we just trained a 1B-2B really strong general model and then continue pertained and/or finetuned it on datasets for downstream tasks like math, code...well structured as Textbooks format or other datasets formats that have been proven to be really efficient and good! Ounce you have 100 finetuned model, just wrap them all into a FrankenMoE and Voila ✨ ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "And that's just what a NOOB like myself had in mind, I'm sure there is better, more efficient ways to do it ! So the question again, why we haven't yet ? I feel I'm missing something... Right?", "raw": "And that's just what a NOOB like myself had in mind, I'm sure there is better, more efficient ways to do it ! So the question again, why we haven't yet ? I feel I'm missing something... Right?", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/626237d9bbcbd1c34f1bb231/9fF2pqFEuobXYCI4SYQa4.png" } ]

[]

[ { "reaction": "🧠", "users": [ "samusenps", "awhsnap", "victor", "m18coppola", "visitthething" ], "count": 5 }, { "reaction": "👍", "users": [ "AIIAR", "JasperV13" ], "count": 2 }, { "reaction": "❤️", "users": [ "samusenps" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/63d9cf5f25575d85ad0ab40f/AJO1V3eyDLlabWsf9IKag.jpeg", "fullname": "Shaleen Bhartiya", "name": "Shaleen123", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 4, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/626237d9bbcbd1c34f1bb231/EJrOjvAL-68qMCYdnvOrq.png", "fullname": "Ali El Filali", "name": "alielfilali01", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 186, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6538119803519fddb4a17e10/ffJMkdx-rM7VvLTCM6ri_.jpeg", "fullname": "samusenps", "name": "samusenps", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 91, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64df20dc22d604b137270864/C-1_EzY0tnrb-Cyn6lh93.jpeg", "fullname": "TA", "name": "AIIAR", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3, "isFollowing": false }, { "avatarUrl": "/avatars/194e6e20374dd6bba04ab167807a0420.svg", "fullname": "Ivan Peshkov", "name": "Erilaz", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false } ]

[ { "type": "text", "value": "Very interesting model just released by MyShell: ", "raw": "Very interesting model just released by MyShell: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/jetmoe/jetmoe-8b", "href": null, "resource": { "type": "model", "id": "jetmoe/jetmoe-8b", "discussionNum": null }, "url": "https://huggingface.co/jetmoe/jetmoe-8b", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " . It's a 8B-parameters MoE LLM so 2.2B active parameters, really efficient.", "raw": " . It's a 8B-parameters MoE LLM so 2.2B active parameters, really efficient.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Main characteristics:", "raw": "Main characteristics:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- impressive performances for its size (beating ", "raw": "- impressive performances for its size (beating ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/meta-llama/Llama-2-7b", "href": null, "resource": { "type": "model", "id": "meta-llama/Llama-2-7b", "discussionNum": null }, "url": "https://huggingface.co/meta-llama/Llama-2-7b", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " and ", "raw": " and ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/huggyllama/llama-13b", "href": null, "resource": { "type": "model", "id": "huggyllama/llama-13b", "discussionNum": null }, "url": "https://huggingface.co/huggyllama/llama-13b", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ")", "raw": ")", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- combine Mixture of Attention heads (MoA) and Mixture of MLP Experts (MoE) – 8 experts with 2 being active for each token", "raw": "- combine Mixture of Attention heads (MoA) and Mixture of MLP Experts (MoE) – 8 experts with 2 being active for each token", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- trained on a rather limited 1.25T tokens from publicly available datasets – training recipe follows the MiniCPM's two-phases training method => first time I see this for a 2B+ model", "raw": "- trained on a rather limited 1.25T tokens from publicly available datasets – training recipe follows the MiniCPM's two-phases training method => first time I see this for a 2B+ model", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- $100k to train", "raw": "- $100k to train", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- open weights - open sharing of recipes - open dataset - open code => ♡", "raw": "- open weights - open sharing of recipes - open dataset - open code => ♡", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- still interesting room to improve performances (be it only by training longer)", "raw": "- still interesting room to improve performances (be it only by training longer)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Links:", "raw": "Links:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- report: ", "raw": "- report: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://research.myshell.ai/jetmoe", "href": "https://research.myshell.ai/jetmoe", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- model: ", "raw": "- model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/jetmoe/jetmoe-8b", "href": null, "resource": { "type": "model", "id": "jetmoe/jetmoe-8b", "discussionNum": null }, "url": "https://huggingface.co/jetmoe/jetmoe-8b", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- code: ", "raw": "- code: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/myshell-ai/JetMoE", "href": "https://github.com/myshell-ai/JetMoE", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Note: I actually detailed all of the MiniCPM schedule, Mixture-of-expert (MoE) and many of the datasets used in this work in my recent little guide to building LLMs in 2024, so feel free to check it out if you want to learn more on these topics: ", "raw": "Note: I actually detailed all of the MiniCPM schedule, Mixture-of-expert (MoE) and many of the datasets used in this work in my recent little guide to building LLMs in 2024, so feel free to check it out if you want to learn more on these topics: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://www.youtube.com/watch?v=2-SPH9hIKT8", "href": "https://www.youtube.com/watch?v=2-SPH9hIKT8", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🚀", "users": [ "samusenps", "a9i", "giux78", "dillfrescott", "clem", "osanseviero" ], "count": 6 }, { "reaction": "❤️", "users": [ "samusenps", "ajibawa-2023", "dillfrescott", "clem", "osanseviero" ], "count": 5 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64aea8ff67511bd3d965697b/Jxn52EmDF5RApJh8antxn.jpeg", "fullname": "Feynman Innovations", "name": "ajibawa-2023", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 138, "isFollowing": false } ]

[ { "type": "text", "value": "🐦 The IBIS Challenge: an open competition in Inferring and predicting transcription factor Binding Specificities: modeling DNA patterns recognized by human regulatory proteins.", "raw": "🐦 The IBIS Challenge: an open competition in Inferring and predicting transcription factor Binding Specificities: modeling DNA patterns recognized by human regulatory proteins.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🧬 Deciphering human gene regulation is a cornerstone of modern molecular biology and biomedicine. Gene activity is controlled by special regulatory proteins, the transcription factors, which recognize DNA sequence patterns. We invite you to join IBIS in our search for the best method to model binding specificities of yet unexplored human regulatory proteins.", "raw": "🧬 Deciphering human gene regulation is a cornerstone of modern molecular biology and biomedicine. Gene activity is controlled by special regulatory proteins, the transcription factors, which recognize DNA sequence patterns. We invite you to join IBIS in our search for the best method to model binding specificities of yet unexplored human regulatory proteins.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "In the challenge, you may use classic methods to represent sequence patterns or any modern approaches", "raw": "In the challenge, you may use classic methods to represent sequence patterns or any modern approaches", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 including decision trees, CNNs, RNNs, LSTMs, and transformers. ", "raw": "🚀 including decision trees, CNNs, RNNs, LSTMs, and transformers. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💡IBIS allows using arbitrary genome or random sequences to pre-train an artificial neural network or to extract features, and use a few existing datasets, see the IBIS documentation (", "raw": "💡IBIS allows using arbitrary genome or random sequences to pre-train an artificial neural network or to extract features, and use a few existing datasets, see the IBIS documentation (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://ibis.autosome.org/docs/technical_details", "href": "https://ibis.autosome.org/docs/technical_details", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ").", "raw": ").", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📊 Yet, the main power and opportunity come with a diverse array of experimental data on 40 human regulatory proteins, many of which remained unexplored until now. ", "raw": "📊 Yet, the main power and opportunity come with a diverse array of experimental data on 40 human regulatory proteins, many of which remained unexplored until now. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🏆 The best methods will be highlighted in the post-challenge high-impact scientific paper 📝, while the winners 🥇of the Primary track of the Final round will be invited to contribute as co-authors. ", "raw": "🏆 The best methods will be highlighted in the post-challenge high-impact scientific paper 📝, while the winners 🥇of the Primary track of the Final round will be invited to contribute as co-authors. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌐 Learn more at ", "raw": "🌐 Learn more at ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://ibis.autosome.org/", "href": "https://ibis.autosome.org/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤗 Our article at HF: ", "raw": "🤗 Our article at HF: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/nikgr/the-ibis-challenge", "href": "https://huggingface.co/blog/nikgr/the-ibis-challenge", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👥 Organizers - GRECO-BIT & Codebook consortia: ", "raw": "👥 Organizers - GRECO-BIT & Codebook consortia: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://ibis.autosome.org/docs/about_us", "href": "https://ibis.autosome.org/docs/about_us", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/66060bf3cb0bd478ab559952/7bxReZ80PzbPeH4esq0Eo.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/66060bf3cb0bd478ab559952/PPekkHOTSHNJN8-t9OIIO.png" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "ivavlakul", "victor" ], "count": 3 }, { "reaction": "🔥", "users": [ "samusenps", "ivavlakul", "osanseviero" ], "count": 3 }, { "reaction": "👀", "users": [ "samusenps", "ivavlakul" ], "count": 2 }, { "reaction": "🚀", "users": [ "ivavlakul" ], "count": 1 } ]

[]

[ { "type": "text", "value": "🔍 Today's pick in Interpretability & Analysis of LMs: Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models by ", "raw": "🔍 Today's pick in Interpretability & Analysis of LMs: Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@sammarks", "href": null, "resource": null, "url": null, "code": null, "user": "sammarks", "label": null, "lang": null }, { "type": "text", "value": " C. Rager ", "raw": " C. Rager ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@eircjm", "href": null, "resource": null, "url": null, "code": null, "user": "eircjm", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@belinkov", "href": null, "resource": null, "url": null, "code": null, "user": "belinkov", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@davidbau", "href": null, "resource": null, "url": null, "code": null, "user": "davidbau", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@amueller", "href": null, "resource": null, "url": null, "code": null, "user": "amueller", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This work proposes using features and errors from sparse autoencoders trained to reconstruct LM activations as interpretable units for circuit discovery. The authors then introduce SHIFT, a technique for editing model behavior by ablating interpretable elements from sparse feature circuits. This method is applied alongside unsupervised circuit discovery at scale by means of clustering, showing highly interpretable feature circuits interacting to produce behaviors like predicting sequence increments.", "raw": "This work proposes using features and errors from sparse autoencoders trained to reconstruct LM activations as interpretable units for circuit discovery. The authors then introduce SHIFT, a technique for editing model behavior by ablating interpretable elements from sparse feature circuits. This method is applied alongside unsupervised circuit discovery at scale by means of clustering, showing highly interpretable feature circuits interacting to produce behaviors like predicting sequence increments.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "I found the experiment of Section 4 especially convincing and exciting in terms of downstream applications: authors trained a classifier over a biased dataset, and showcased how SHIFT intervention in feature space leads to performances matching those of the same model trained on an unbiased data distribution!", "raw": "I found the experiment of Section 4 especially convincing and exciting in terms of downstream applications: authors trained a classifier over a biased dataset, and showcased how SHIFT intervention in feature space leads to performances matching those of the same model trained on an unbiased data distribution!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Paper: ", "raw": "📄 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2403.19647", "href": null, "resource": { "type": "paper", "id": "2403.19647", "discussionNum": null }, "url": "https://huggingface.co/papers/2403.19647", "code": null, "user": null, "label": "Sparse Feature Circuits: Discovering and Editing Interpretable Causal\n Graphs in Language Models (2403.19647)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 All daily picks: ", "raw": "🔍 All daily picks: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-of-lms-65ae3339949c5675d25de2f9", "href": null, "resource": { "type": "collection", "id": "gsarti/daily-picks-in-interpretability-and-analysis-of-lms-65ae3339949c5675d25de2f9", "discussionNum": null }, "url": "https://huggingface.co/collections/gsarti/daily-picks-in-interpretability-and-analysis-of-lms-65ae3339949c5675d25de2f9", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/dUdmCBTUZYDOVp0VRX5c_.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/N1oSguoNKfc1Hnjo8WV57.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e7749883d77a72421292d07/C-VPT0qbRqBRkvNrYwnYN.png" } ]

[ { "avatarUrl": "/avatars/eba6c3e61a7a6a12f487d531cf7a9934.svg", "fullname": "Aaron Mueller", "name": "amueller", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1 }, { "avatarUrl": "/avatars/186a9aed84681246f48ed2a012c50def.svg", "fullname": "Yonatan Belinkov", "name": "belinkov", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3 }, { "avatarUrl": "/avatars/ac208cd180b4f3ed1ec367e581facfcf.svg", "fullname": "David Bau", "name": "davidbau", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 4 }, { "avatarUrl": "/avatars/f200b4e78f4c8008689f61b914f23998.svg", "fullname": "Sam Marks", "name": "sammarks", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1 } ]

[ { "reaction": "❤️", "users": [ "samusenps", "nikgr", "thomwolf", "osanseviero", "victor", "toyxyz" ], "count": 6 }, { "reaction": "🔥", "users": [ "NeuralNovel", "thomwolf" ], "count": 2 }, { "reaction": "🚀", "users": [ "NeuralNovel", "thomwolf" ], "count": 2 } ]

[]

[ { "type": "text", "value": "Today we're starting a new initiative: LM Studio Community Models! 🤖", "raw": "Today we're starting a new initiative: LM Studio Community Models! 🤖", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@bartowski", "href": null, "resource": null, "url": null, "code": null, "user": "bartowski", "label": null, "lang": null }, { "type": "text", "value": ", a prolific quantizer (both GGUF and EXL2) will be helping to curate notable new models in LM Studio's Community Models page: ", "raw": ", a prolific quantizer (both GGUF and EXL2) will be helping to curate notable new models in LM Studio's Community Models page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/lmstudio-community", "href": "https://huggingface.co/lmstudio-community", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ". ", "raw": ". ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Our goal is to ensure the community has access to GGUF files for new & noteworthy models as soon as possible. Keep an eye on that page for updates.", "raw": "Our goal is to ensure the community has access to GGUF files for new & noteworthy models as soon as possible. Keep an eye on that page for updates.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "If you're unfamiliar with GGUF, it's the de-facto standard for 'compressed' LLM weights. It is the native format of llama.cpp (", "raw": "If you're unfamiliar with GGUF, it's the de-facto standard for 'compressed' LLM weights. It is the native format of llama.cpp (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/ggerganov/llama.cpp", "href": "https://github.com/ggerganov/llama.cpp", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ", an LLM runtime C/C++ library.) This format is supported in LM Studio.", "raw": ", an LLM runtime C/C++ library.) This format is supported in LM Studio.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We will also be sharing new models on the LM Studio Discord: ", "raw": "We will also be sharing new models on the LM Studio Discord: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://discord.gg/aPQfnNkxGC", "href": "https://discord.gg/aPQfnNkxGC", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6435718aaaef013d1aec3b8b/XKf-8MA47tjVAM6SCX0MP.jpeg", "fullname": "Bartowski", "name": "bartowski", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 2816 } ]

[ { "reaction": "🔥", "users": [ "bartowski", "victor", "samusenps", "Kukedlc", "InferenceIllusionist", "bunnycore", "Severian", "osanseviero", "MexIvanov", "dillfrescott", "nikgr", "thomwolf", "lixin67", "julien-c", "sroecker", "Midgardsormr", "AgentHaswing", "apol", "British-Rat", "tcowley" ], "count": 20 }, { "reaction": "❤️", "users": [ "julien-c", "British-Rat" ], "count": 2 }, { "reaction": "🧠", "users": [ "julien-c", "British-Rat" ], "count": 2 }, { "reaction": "🚀", "users": [ "jik1992" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/5dd96eb166059660ed1ee413/NQtzmrDdbG0H8qkZvRyGk.jpeg", "fullname": "Julien Chaumond", "name": "julien-c", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 1580, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1583857746553-5df7e9e5da6d0311fd3d53f9.jpeg", "fullname": "Thomas Wolf", "name": "thomwolf", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 704, "isFollowing": false }, { "avatarUrl": "/avatars/21d78d21fac28c5c33712f8895bec850.svg", "fullname": "Bulidoge", "name": "DBL0422", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "/avatars/dc348e39cfcda272e95f206dbe8c9de7.svg", "fullname": "Yagil Burowski", "name": "yagilb", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2086, "isFollowing": false } ]

[ { "type": "text", "value": "TIL: since Text Generation Inference supports Messages API, which is compatible with the OpenAI Chat Completion API, you can trace calls made to inference endpoints using Langfuse's OpenAI API integration. ", "raw": "TIL: since Text Generation Inference supports Messages API, which is compatible with the OpenAI Chat Completion API, you can trace calls made to inference endpoints using Langfuse's OpenAI API integration. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "A Hugging Face Pro subscription includes access to many models you want to test when developing an app (", "raw": "A Hugging Face Pro subscription includes access to many models you want to test when developing an app (", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/inference-pro", "href": "https://huggingface.co/blog/inference-pro", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "). Using the endpoint and tracing your generations during this development process is an excellent way for GPU-poor people to bootstrap an initial dataset quickly while prototyping. ", "raw": "). Using the endpoint and tracing your generations during this development process is an excellent way for GPU-poor people to bootstrap an initial dataset quickly while prototyping. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60107b385ac3e86b3ea4fc34/T_PgHvbueF1Vn0m9pRabY.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60107b385ac3e86b3ea4fc34/JOxrRPzipWRXiCq9S5pcN.png" } ]

[]

[ { "reaction": "👍", "users": [ "samusenps", "lunarflu", "osanseviero", "nikgr", "victor", "VictorSanh" ], "count": 6 }, { "reaction": "➕", "users": [ "samusenps", "lunarflu", "osanseviero" ], "count": 3 } ]

[]

[ { "type": "text", "value": "After giving GPU Programming a hands-on try, I have come to appreciate the level of complexity in AI compute:", "raw": "After giving GPU Programming a hands-on try, I have come to appreciate the level of complexity in AI compute:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Existing/leading frameworks (CUDA, OpenCL, DSLs, even Triton), still fall at the mercy of low-level compute that requires deeper understanding and experience.", "raw": "- Existing/leading frameworks (CUDA, OpenCL, DSLs, even Triton), still fall at the mercy of low-level compute that requires deeper understanding and experience.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Ambiguous optimizations methods that will literally drive you mad 🤯", "raw": "- Ambiguous optimizations methods that will literally drive you mad 🤯", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Triton is cool but not cool enough (high level abstractions that fall back to low level compute issues as you build more specialized kernels)", "raw": "- Triton is cool but not cool enough (high level abstractions that fall back to low level compute issues as you build more specialized kernels)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- As for CUDA, optimization requires considering all major components of the GPU (DRAM, SRAM, ALUs) 🤕", "raw": "- As for CUDA, optimization requires considering all major components of the GPU (DRAM, SRAM, ALUs) 🤕", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Models today require stallion written GPU kernels to reduce storage and compute cost.", "raw": "- Models today require stallion written GPU kernels to reduce storage and compute cost.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- GPTQ was a big save 👍🏼", "raw": "- GPTQ was a big save 👍🏼", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@karpathy", "href": null, "resource": null, "url": null, "code": null, "user": "karpathy", "label": null, "lang": null }, { "type": "text", "value": " is right expertise in this area is scarce and the reason is quite obvious - uncertainties: we are still struggling to get peak performance from multi-connected GPUs while maintaining precision and reducing cost. ", "raw": " is right expertise in this area is scarce and the reason is quite obvious - uncertainties: we are still struggling to get peak performance from multi-connected GPUs while maintaining precision and reducing cost. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "May the Scaling Laws favor us lol.", "raw": "May the Scaling Laws favor us lol.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/x1F_eDWUsTTX4x1xxDcZA.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/bNjc7O4yAEvC45V3MpzNy.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/q_l8rikzfPYXMedJpRaOc.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/krrfQNxhwPX8w_SKN1tyg.jpeg" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1660434061546-62f83661fe21cc4875221c0f.jpeg", "fullname": "Andrej K", "name": "karpathy", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 476 } ]

[ { "reaction": "🤗", "users": [ "alielfilali01", "nebiyu29", "victor", "lunarflu", "larryvrh", "nikgr", "str4007", "Abecid", "lixin67" ], "count": 9 }, { "reaction": "❤️", "users": [ "samusenps", "lunarflu", "badayvedat", "Abecid", "axwinters" ], "count": 5 }, { "reaction": "🧠", "users": [ "samusenps", "lunarflu" ], "count": 2 } ]

[ { "avatarUrl": "/avatars/e4eea78b457e9f00b2ccfe17efeaace2.svg", "fullname": "Aaron Lisman", "name": "pwf", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2, "isFollowing": false }, { "avatarUrl": "/avatars/acff4702ebdeb696a185a6b9e58c0558.svg", "fullname": "Matt St. Hilaire", "name": "msthil2", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 2, "isFollowing": false }, { "avatarUrl": "/avatars/613ea7a2b4f8acc06c1def85da8f1640.svg", "fullname": "Clyde Wright", "name": "clydeiii", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6438a9027de34e8ea7e4b257/vib8QSd1AWMr_bR9ig_xJ.jpeg", "fullname": "Jaward Sesay", "name": "Jaward", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 191, "isFollowing": false }, { "avatarUrl": "/avatars/223d964e993126fac680a8dfb71cc291.svg", "fullname": "James K J", "name": "james92", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false } ]

[ { "type": "text", "value": "[𝐍𝐞𝐰 𝐏𝐚𝐩𝐞𝐫] 𝐀𝐥𝐥 𝐭𝐨𝐤𝐞𝐧𝐬 𝐬𝐡𝐨𝐮𝐥𝐝 𝐧𝐨𝐭 𝐫𝐞𝐪𝐮𝐢𝐫𝐞 𝐭𝐡𝐞 𝐬𝐚𝐦𝐞 𝐞𝐟𝐟𝐨𝐫𝐭 𝐭𝐨 𝐜𝐨𝐦𝐩𝐮𝐭𝐞! ⇒ 𝐌𝐢𝐱𝐭𝐮𝐫𝐞 𝐨𝐟 𝐝𝐞𝐩𝐭𝐡𝐬 🫧🐠", "raw": "[𝐍𝐞𝐰 𝐏𝐚𝐩𝐞𝐫] 𝐀𝐥𝐥 𝐭𝐨𝐤𝐞𝐧𝐬 𝐬𝐡𝐨𝐮𝐥𝐝 𝐧𝐨𝐭 𝐫𝐞𝐪𝐮𝐢𝐫𝐞 𝐭𝐡𝐞 𝐬𝐚𝐦𝐞 𝐞𝐟𝐟𝐨𝐫𝐭 𝐭𝐨 𝐜𝐨𝐦𝐩𝐮𝐭𝐞! ⇒ 𝐌𝐢𝐱𝐭𝐮𝐫𝐞 𝐨𝐟 𝐝𝐞𝐩𝐭𝐡𝐬 🫧🐠", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Google Researchers were unhappy with the way current decoding generally works: all tokens go through the same layers, thus requiring exactly the same effort to compute.", "raw": "Google Researchers were unhappy with the way current decoding generally works: all tokens go through the same layers, thus requiring exactly the same effort to compute.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Whereas in reality, completing the answer to a difficult math problem for instance should be more computationally intense than completing the text of the Declaration of Independence: 𝗻𝗼𝘁 𝗮𝗹𝗹 𝘁𝗼𝗸𝗲𝗻𝘀 𝗮𝗿𝗲 𝗰𝗿𝗲𝗮𝘁𝗲𝗱 𝗲𝗾𝘂𝗮𝗹!", "raw": "Whereas in reality, completing the answer to a difficult math problem for instance should be more computationally intense than completing the text of the Declaration of Independence: 𝗻𝗼𝘁 𝗮𝗹𝗹 𝘁𝗼𝗸𝗲𝗻𝘀 𝗮𝗿𝗲 𝗰𝗿𝗲𝗮𝘁𝗲𝗱 𝗲𝗾𝘂𝗮𝗹!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "➡️ 𝗧𝗵𝗲𝘆 𝗵𝗮𝗱 𝘁𝗵𝗶𝘀 𝗴𝗲𝗻𝗶𝘂𝘀 𝗶𝗱𝗲𝗮: 💡 𝗵𝗮𝘃𝗶𝗻𝗴 𝗮 𝘁𝗼𝗸𝗲𝗻 𝗴𝗼 𝘁𝗵𝗿𝗼𝘂𝗴𝗵 𝗮 𝗯𝗹𝗼𝗰𝗸 𝘀𝗵𝗼𝘂𝗹𝗱 𝗯𝗲 𝗼𝗽𝘁𝗶𝗼𝗻𝗮𝗹. The token can go through the block (thus undergoing expensive self-attention computation) or avoid it through a skip connection.", "raw": "➡️ 𝗧𝗵𝗲𝘆 𝗵𝗮𝗱 𝘁𝗵𝗶𝘀 𝗴𝗲𝗻𝗶𝘂𝘀 𝗶𝗱𝗲𝗮: 💡 𝗵𝗮𝘃𝗶𝗻𝗴 𝗮 𝘁𝗼𝗸𝗲𝗻 𝗴𝗼 𝘁𝗵𝗿𝗼𝘂𝗴𝗵 𝗮 𝗯𝗹𝗼𝗰𝗸 𝘀𝗵𝗼𝘂𝗹𝗱 𝗯𝗲 𝗼𝗽𝘁𝗶𝗼𝗻𝗮𝗹. The token can go through the block (thus undergoing expensive self-attention computation) or avoid it through a skip connection.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The routing decision is taken on the block level: each block selects from the total sequence the top-k tokens that will go through it, and the others tokens will skip it. 𝘛𝘩𝘪𝘴 𝘢𝘭𝘭𝘰𝘸𝘴 𝘵𝘰 𝘤𝘩𝘰𝘰𝘴𝘦 𝘵𝘩𝘦 𝘦𝘹𝘢𝘤𝘵 𝙘𝙖𝙥𝙖𝙘𝙞𝙩𝙮 𝘰𝘧 𝘢 𝘣𝘭𝘰𝘤𝘬, 𝘪.𝘦. 𝘵𝘩𝘦 𝘱𝘳𝘰𝘱𝘰𝘳𝘵𝘪𝘰𝘯 𝘰𝘧 𝘵𝘰𝘬𝘦𝘯𝘴 𝘵𝘩𝘢𝘵 𝘨𝘰 𝘵𝘩𝘳𝘰𝘶𝘨𝘩 𝘪𝘵, 𝘸𝘩𝘪𝘤𝘩 𝘥𝘪𝘳𝘦𝘤𝘵𝘭𝘺 𝘪𝘯𝘧𝘭𝘶𝘦𝘯𝘤𝘦𝘴 𝘵𝘩𝘦 𝘤𝘰𝘮𝘱𝘶𝘵𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘪𝘯𝘵𝘦𝘯𝘴𝘪𝘵𝘺 𝘰𝘧 𝘵𝘩𝘦 𝘧𝘰𝘳𝘸𝘢𝘳𝘥 𝘱𝘢𝘴𝘴.", "raw": "The routing decision is taken on the block level: each block selects from the total sequence the top-k tokens that will go through it, and the others tokens will skip it. 𝘛𝘩𝘪𝘴 𝘢𝘭𝘭𝘰𝘸𝘴 𝘵𝘰 𝘤𝘩𝘰𝘰𝘴𝘦 𝘵𝘩𝘦 𝘦𝘹𝘢𝘤𝘵 𝙘𝙖𝙥𝙖𝙘𝙞𝙩𝙮 𝘰𝘧 𝘢 𝘣𝘭𝘰𝘤𝘬, 𝘪.𝘦. 𝘵𝘩𝘦 𝘱𝘳𝘰𝘱𝘰𝘳𝘵𝘪𝘰𝘯 𝘰𝘧 𝘵𝘰𝘬𝘦𝘯𝘴 𝘵𝘩𝘢𝘵 𝘨𝘰 𝘵𝘩𝘳𝘰𝘶𝘨𝘩 𝘪𝘵, 𝘸𝘩𝘪𝘤𝘩 𝘥𝘪𝘳𝘦𝘤𝘵𝘭𝘺 𝘪𝘯𝘧𝘭𝘶𝘦𝘯𝘤𝘦𝘴 𝘵𝘩𝘦 𝘤𝘰𝘮𝘱𝘶𝘵𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘪𝘯𝘵𝘦𝘯𝘴𝘪𝘵𝘺 𝘰𝘧 𝘵𝘩𝘦 𝘧𝘰𝘳𝘸𝘢𝘳𝘥 𝘱𝘢𝘴𝘴.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "This yields Mixture-of-Depths (MoD), with spectacular results.", "raw": "This yields Mixture-of-Depths (MoD), with spectacular results.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "✨ 𝗥𝗲𝘀𝘂𝗹𝘁𝘀:", "raw": "✨ 𝗥𝗲𝘀𝘂𝗹𝘁𝘀:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🎚️ 𝗖𝗮𝗽𝗮𝗰𝗶𝘁𝘆 𝗰𝗮𝗻 𝗯𝗲 𝘁𝘂𝗻𝗲𝗱 𝗮𝗹𝗹 𝘁𝗵𝗲 𝘄𝗮𝘆 𝗱𝗼𝘄𝗻 𝘁𝗼 𝟭𝟮.𝟱% for every second block: thus 87.5% of tokens just skip the block!", "raw": "🎚️ 𝗖𝗮𝗽𝗮𝗰𝗶𝘁𝘆 𝗰𝗮𝗻 𝗯𝗲 𝘁𝘂𝗻𝗲𝗱 𝗮𝗹𝗹 𝘁𝗵𝗲 𝘄𝗮𝘆 𝗱𝗼𝘄𝗻 𝘁𝗼 𝟭𝟮.𝟱% for every second block: thus 87.5% of tokens just skip the block!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 For the same training time and performance, >𝟲𝟬% 𝗶𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝘀𝗽𝗲𝗲𝗱!", "raw": "🚀 For the same training time and performance, >𝟲𝟬% 𝗶𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝘀𝗽𝗲𝗲𝗱!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤝 𝗖𝗮𝗻 𝗯𝗲 𝗰𝗼𝗺𝗯𝗶𝗻𝗲𝗱 𝘄𝗶𝘁𝗵 𝗠𝗶𝘅𝘁𝘂𝗿𝗲-𝗼𝗳-𝗘𝘅𝗽𝗲𝗿𝘁𝘀 for further improvements.", "raw": "🤝 𝗖𝗮𝗻 𝗯𝗲 𝗰𝗼𝗺𝗯𝗶𝗻𝗲𝗱 𝘄𝗶𝘁𝗵 𝗠𝗶𝘅𝘁𝘂𝗿𝗲-𝗼𝗳-𝗘𝘅𝗽𝗲𝗿𝘁𝘀 for further improvements.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 𝗣𝗮𝗽𝗲𝗿 𝗵𝗲𝗿𝗲 👉 ", "raw": "📄 𝗣𝗮𝗽𝗲𝗿 𝗵𝗲𝗿𝗲 👉 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.02258", "href": null, "resource": { "type": "paper", "id": "2404.02258", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.02258", "code": null, "user": null, "label": "Mixture-of-Depths: Dynamically allocating compute in transformer-based\n language models (2404.02258)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📚 I added it to my paper collection 👉 ", "raw": "📚 I added it to my paper collection 👉 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/m-ric/spinning-up-in-llms-659e698f9dd5a71bd3f579a7", "href": null, "resource": { "type": "collection", "id": "m-ric/spinning-up-in-llms-659e698f9dd5a71bd3f579a7", "discussionNum": null }, "url": "https://huggingface.co/collections/m-ric/spinning-up-in-llms-659e698f9dd5a71bd3f579a7", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/63d10d4e8eaa4831005e92b5/rQ7w1wn7wGkGqn-0AU5m2.png" } ]

[]

[ { "reaction": "❤️", "users": [ "clem", "samusenps", "ABX-AI", "mmhamdy", "nikgr", "minjejeon", "qinluo", "MultiTrickFox", "mammour", "danielus", "Joseph717171", "ivanfioravanti" ], "count": 12 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64345a5f4b34368fdb045ef4/NfUgcPA08_Gl40pTrlWfe.jpeg", "fullname": "Ivan Fioravanti", "name": "ivanfioravanti", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 31, "isFollowing": false } ]

[ { "type": "text", "value": "⚓️ Sailor: A New Multilingual Open LLM for South-East Asia 🌏", "raw": "⚓️ Sailor: A New Multilingual Open LLM for South-East Asia 🌏", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Last month we have released a new family of multilingual language models called **Sailor**, ranging from 0.5B to 7B parameters, continually pre-trained from the Qwen1.5 models. Based on our extensive benchmarking, the Sailor models demonstrate exceptional performance on South-East Asian languages, taking us one step closer to multilingual LLMs that can serve the diverse needs of the region and beyond. ", "raw": "Last month we have released a new family of multilingual language models called **Sailor**, ranging from 0.5B to 7B parameters, continually pre-trained from the Qwen1.5 models. Based on our extensive benchmarking, the Sailor models demonstrate exceptional performance on South-East Asian languages, taking us one step closer to multilingual LLMs that can serve the diverse needs of the region and beyond. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Today, we're more than excited to share the key technical details behind the Sailor models! 💪", "raw": "Today, we're more than excited to share the key technical details behind the Sailor models! 💪", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "**Key highlights**:", "raw": "**Key highlights**:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Data curation: Merging short examples, document-level code-switching, aggressive data cleaning and deduplication.", "raw": "🔍 Data curation: Merging short examples, document-level code-switching, aggressive data cleaning and deduplication.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤖 Tokenization Robustness: We find that BPE dropout is really effective to deal with prompt variations.", "raw": "🤖 Tokenization Robustness: We find that BPE dropout is really effective to deal with prompt variations.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Optimizing Data Mixture: We propose a new approach to automatically balance capabilities across different languages! ", "raw": "🔍 Optimizing Data Mixture: We propose a new approach to automatically balance capabilities across different languages! ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌟 Recipe in Continual Pre-training: We discover a powerful metric that can help predict how well the Sailor models will perform on the original domain (e.g., English) after continual pre-training.", "raw": "🌟 Recipe in Continual Pre-training: We discover a powerful metric that can help predict how well the Sailor models will perform on the original domain (e.g., English) after continual pre-training.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We are thrilled to share these technical details with the community and invite you to explore the Sailor models. We hope Sailor models take us one step closer to multilingual LLMs in the world! 🌍✨ ", "raw": "We are thrilled to share these technical details with the community and invite you to explore the Sailor models. We hope Sailor models take us one step closer to multilingual LLMs in the world! 🌍✨ ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "To learn more, please access our research paper or reach out to our team.", "raw": "To learn more, please access our research paper or reach out to our team.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔗 Paper: ", "raw": "🔗 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.03608", "href": null, "resource": { "type": "paper", "id": "2404.03608", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.03608", "code": null, "user": null, "label": "Sailor: Open Language Models for South-East Asia (2404.03608)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🧩 Model: ", "raw": "🧩 Model: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/sail/sailor-language-models-65e19a749f978976f1959825", "href": null, "resource": { "type": "collection", "id": "sail/sailor-language-models-65e19a749f978976f1959825", "discussionNum": null }, "url": "https://huggingface.co/collections/sail/sailor-language-models-65e19a749f978976f1959825", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "💻 Code: ", "raw": "💻 Code: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/sail-sg/sailor-llm", "href": "https://github.com/sail-sg/sailor-llm", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/612ee6a7b960e78c6d2319d4/axEbDm39hzsW-2nE00Rb1.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/612ee6a7b960e78c6d2319d4/7t7ks6f4CcsLNt6qAStjo.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/612ee6a7b960e78c6d2319d4/KHFGz9S9qILPDg5oKPBxD.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/612ee6a7b960e78c6d2319d4/yPSM_wPg8nz4UAZsOvJ6C.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/612ee6a7b960e78c6d2319d4/bjVhZdcLFP47Da-uG-ubx.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/612ee6a7b960e78c6d2319d4/zWIKRvFCtHyyqhMKHj-6E.jpeg" } ]

[]

[ { "reaction": "🔥", "users": [ "dreamerdeo", "Jiiiia", "teddythinh", "chaoscodes", "DmitryRyumin", "taufiqdp", "SivilTaram", "osanseviero", "clem", "victor", "lunarflu", "Kernel", "nikgr", "gabrielchua", "DongfuJiang", "damerajee", "yjernite" ], "count": 17 }, { "reaction": "🚀", "users": [ "dreamerdeo", "Jiiiia", "teddythinh", "aboutyourresearch", "chaoscodes", "monsoon-nlp", "osanseviero", "clem", "SivilTaram", "samusenps", "lunarflu", "damerajee", "yjernite" ], "count": 13 }, { "reaction": "❤️", "users": [ "vitaliim0", "damerajee" ], "count": 2 }, { "reaction": "👀", "users": [ "SimonRAVCARE" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Mixture-of-Depths", "raw": "Mixture-of-Depths", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Dynamically allocating compute in transformer-based language models", "raw": "Dynamically allocating compute in transformer-based language models", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.02258", "href": null, "resource": { "type": "paper", "id": "2404.02258", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.02258", "code": null, "user": null, "label": "Mixture-of-Depths: Dynamically allocating compute in transformer-based\n language models (2404.02258)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Transformer-based language models spread FLOPs uniformly across input sequences. In this work we demonstrate that transformers can instead learn to dynamically allocate FLOPs (or compute) to specific positions in a sequence, optimising the allocation along the sequence for different layers across the model depth. Our method enforces a total compute budget by capping the number of tokens (k) that can participate in the self-attention and MLP computations at a given layer. The tokens to be processed are determined by the network using a top-k routing mechanism. Since k is defined a priori, this simple procedure uses a static computation graph with known tensor sizes, unlike other conditional computation techniques. Nevertheless, since the identities of the k tokens are fluid, this method can expend FLOPs non-uniformly across the time and model depth dimensions. Thus, compute expenditure is entirely predictable in sum total, but dynamic and context-sensitive at the token-level. Not only do models trained in this way learn to dynamically allocate compute, they do so efficiently. These models match baseline performance for equivalent FLOPS and wall-clock times to train, but require a fraction of the FLOPs per forward pass, and can be upwards of 50\\% faster to step during post-training sampling.", "raw": "Transformer-based language models spread FLOPs uniformly across input sequences. In this work we demonstrate that transformers can instead learn to dynamically allocate FLOPs (or compute) to specific positions in a sequence, optimising the allocation along the sequence for different layers across the model depth. Our method enforces a total compute budget by capping the number of tokens (k) that can participate in the self-attention and MLP computations at a given layer. The tokens to be processed are determined by the network using a top-k routing mechanism. Since k is defined a priori, this simple procedure uses a static computation graph with known tensor sizes, unlike other conditional computation techniques. Nevertheless, since the identities of the k tokens are fluid, this method can expend FLOPs non-uniformly across the time and model depth dimensions. Thus, compute expenditure is entirely predictable in sum total, but dynamic and context-sensitive at the token-level. Not only do models trained in this way learn to dynamically allocate compute, they do so efficiently. These models match baseline performance for equivalent FLOPS and wall-clock times to train, but require a fraction of the FLOPs per forward pass, and can be upwards of 50\\% faster to step during post-training sampling.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60f1abe7544c2adfd699860c/tX3kCVjboQNSNcNrMXPC6.png" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "taufiqdp", "clem", "mick-bitovi", "iancos", "csouza", "nikgr", "rupeshs", "sugatoray", "dillfrescott", "ceshine" ], "count": 11 }, { "reaction": "👀", "users": [ "stormchaser" ], "count": 1 }, { "reaction": "👍", "users": [ "dashfunnydashdash" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6382255fcae34727b9cc149e/PYiwi8LVZParYvImmcGez.png", "fullname": "Ram", "name": "0-hero", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 32, "isFollowing": false } ]

[ { "type": "text", "value": "Introducing ", "raw": "Introducing ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/gretelai/synthetic_text_to_sql", "href": null, "resource": { "type": "dataset", "id": "gretelai/synthetic_text_to_sql", "discussionNum": null }, "url": "https://huggingface.co/datasets/gretelai/synthetic_text_to_sql", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " by ", "raw": " by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/gretelai", "href": "https://huggingface.co/gretelai", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "It stands as the largest and most diverse synthetic Text-to-SQL dataset available to-date.", "raw": "It stands as the largest and most diverse synthetic Text-to-SQL dataset available to-date.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The dataset includes:", "raw": "The dataset includes:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- 105,851 records partitioned into 100,000 train and 5,851 test records", "raw": "- 105,851 records partitioned into 100,000 train and 5,851 test records", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "~23M total tokens, including ~12M SQL tokens", "raw": "~23M total tokens, including ~12M SQL tokens", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Coverage across 100 distinct domains/verticals", "raw": "- Coverage across 100 distinct domains/verticals", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Comprehensive array of SQL tasks: data definition, retrieval, manipulation, analytics & reporting", "raw": "- Comprehensive array of SQL tasks: data definition, retrieval, manipulation, analytics & reporting", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Wide range of SQL complexity levels, including subqueries, single joins, multiple joins, aggregations, window functions, set operations", "raw": "- Wide range of SQL complexity levels, including subqueries, single joins, multiple joins, aggregations, window functions, set operations", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Database context, including table and view create statements", "raw": "- Database context, including table and view create statements", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Natural language explanations of what the SQL query is doing", "raw": "- Natural language explanations of what the SQL query is doing", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- Contextual tags to optimize model training", "raw": "- Contextual tags to optimize model training", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Blogpost: ", "raw": "Blogpost: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://gretel.ai/blog/synthetic-text-to-sql-dataset", "href": "https://gretel.ai/blog/synthetic-text-to-sql-dataset", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Dataset: ", "raw": "Dataset: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/gretelai/synthetic_text_to_sql", "href": null, "resource": { "type": "dataset", "id": "gretelai/synthetic_text_to_sql", "discussionNum": null }, "url": "https://huggingface.co/datasets/gretelai/synthetic_text_to_sql", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5e67bdd61009063689407479/LReH7TiCXyLumal3deBBn.webp" } ]

[]

[ { "reaction": "🤗", "users": [ "ichigoberry", "mvaloatto", "Svngoku", "nikgr", "thomwolf", "lhoestq", "brhnsbn" ], "count": 7 }, { "reaction": "❤️", "users": [ "samusenps", "thomwolf", "lhoestq" ], "count": 3 }, { "reaction": "🔥", "users": [ "Jensen-holm" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6389d3c61e8755d777902366/xR50wHQW4ZsFxI8lolk7a.jpeg", "fullname": "Thomas Schranz", "name": "ichigoberry", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 16, "isFollowing": false } ]

[ { "type": "text", "value": "🚀🎭🌟 New Research Alert - CVPR 2024 (Avatars Collection)! 🌟 🎭🚀", "raw": "🚀🎭🌟 New Research Alert - CVPR 2024 (Avatars Collection)! 🌟 🎭🚀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Title: GeneAvatar: Generic Expression-Aware Volumetric Head Avatar Editing from a Single Image 🔝", "raw": "📄 Title: GeneAvatar: Generic Expression-Aware Volumetric Head Avatar Editing from a Single Image 🔝", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📝 Description: GeneAvatar is a generic approach for editing 3D head avatars based on a single 2D image, applicable to different volumetric representations. The novel expression-aware generative modification model delivers high quality and consistent editing results across multiple viewpoints and emotions.", "raw": "📝 Description: GeneAvatar is a generic approach for editing 3D head avatars based on a single 2D image, applicable to different volumetric representations. The novel expression-aware generative modification model delivers high quality and consistent editing results across multiple viewpoints and emotions.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👥 Authors: Chong Bao et al.", "raw": "👥 Authors: Chong Bao et al.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📅 Conference: CVPR, Jun 17-21, 2024 | Seattle WA, USA 🇺🇸", "raw": "📅 Conference: CVPR, Jun 17-21, 2024 | Seattle WA, USA 🇺🇸", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔗 Paper: ", "raw": "🔗 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.02152", "href": null, "resource": { "type": "paper", "id": "2404.02152", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.02152", "code": null, "user": null, "label": "GeneAvatar: Generic Expression-Aware Volumetric Head Avatar Editing from\n a Single Image (2404.02152)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌐 Github Page: ", "raw": "🌐 Github Page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://zju3dv.github.io/geneavatar/", "href": "https://zju3dv.github.io/geneavatar/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📁 Repository: ", "raw": "📁 Repository: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/zju3dv/GeneAvatar", "href": "https://github.com/zju3dv/GeneAvatar", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📺 Video: ", "raw": "📺 Video: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://www.youtube.com/watch?v=4zfbfPivtVU", "href": "https://www.youtube.com/watch?v=4zfbfPivtVU", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "raw": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "href": null, "resource": { "type": "space", "id": "DmitryRyumin/NewEraAI-Papers", "discussionNum": null }, "url": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " curated by ", "raw": " curated by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DmitryRyumin", "href": null, "resource": null, "url": null, "code": null, "user": "DmitryRyumin", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Added to the Avatars Collection: ", "raw": "🚀 Added to the Avatars Collection: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "href": null, "resource": { "type": "collection", "id": "DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "discussionNum": null }, "url": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Keywords: #GeneAvatar #HeadAvatar #3DHeadAvatarEditing #VolumetricHeadAvatar #SingleImageEditing #ExpressionAwareModification #CVPR2024 #DeepLearning #Innovation", "raw": "🔍 Keywords: #GeneAvatar #HeadAvatar #3DHeadAvatarEditing #VolumetricHeadAvatar #SingleImageEditing #ExpressionAwareModification #CVPR2024 #DeepLearning #Innovation", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/ol7FKtAPKKA7ppLdHFCZj.gif" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/6Vz2lsQpf_QYGQEtaviEi.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/FbXT0rS7dBmQ2w5pzIJyR.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/T-x2S0C1K9rQS19w5I4iM.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/qM8npv1vMLpOYRY0PEWSB.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/obK9RJ7J7j0p3CFkgy9jm.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/GhmBxqeyXzJ4GszxWacBy.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/U3lxG5Zdg2Alfs8ZkFNLu.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/YJfZe4ymSWbMTo915ptlS.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/nRCxbVng_PPBqKd-Z3KVc.jpeg", "fullname": "Dmitry Ryumin", "name": "DmitryRyumin", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 377 } ]

[ { "reaction": "👍", "users": [ "dashfunnydashdash", "samusenps", "Limbicnation", "MichaelKwak", "osanseviero", "clem", "introvoyz041", "Benson", "Alkohole" ], "count": 9 }, { "reaction": "🔥", "users": [ "DmitryRyumin", "samusenps", "osanseviero", "Kukedlc", "clem" ], "count": 5 }, { "reaction": "🤗", "users": [ "DmitryRyumin", "Wilhelm", "osanseviero", "nikgr" ], "count": 4 }, { "reaction": "👀", "users": [ "SimonRAVCARE" ], "count": 1 } ]

[]

[ { "type": "text", "value": "Introducing the Gradio API Recorder 🪄", "raw": "Introducing the Gradio API Recorder 🪄", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Every Gradio app now includes an API recorder that lets you reconstruct your interaction in a Gradio app as code using the Python or JS clients! Our goal is to make Gradio the easiest way to build ML APIs, not just UIs 🔥", "raw": "Every Gradio app now includes an API recorder that lets you reconstruct your interaction in a Gradio app as code using the Python or JS clients! Our goal is to make Gradio the easiest way to build ML APIs, not just UIs 🔥", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/608b8bb39d7c9519b4adae19/c3TeW08ESdnICPPxPri8J.qt" } ]

[]

[ { "reaction": "👍", "users": [ "Bagus", "victor", "merve", "gabrielchua", "samusenps", "radames", "WhatIFTheMind", "osanseviero", "ademait", "brianjking", "RocktProductions" ], "count": 11 }, { "reaction": "🤗", "users": [ "DmitryRyumin", "merve", "nikgr", "samusenps", "radames", "jeffboudier", "osanseviero" ], "count": 7 }, { "reaction": "🚀", "users": [ "DmitryRyumin", "victor", "merve", "radames", "osanseviero" ], "count": 5 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1652197428055-625a6e0c535747b1a15be2de.jpeg", "fullname": "Md. Musfiqur Rahaman", "name": "musfiqdehan", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 5, "isFollowing": false }, { "avatarUrl": "/avatars/e0f4ee1c029ab2241dd9b24ae86fb31a.svg", "fullname": "Brian King", "name": "brianjking", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1621947938344-noauth.png", "fullname": "Abubakar Abid", "name": "abidlabs", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 487, "isFollowing": false } ]

[ { "type": "text", "value": "Greeting!", "raw": "Greeting!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We are happy to introduce our InstantStyle, which is a framework that employs straightforward yet potent techniques for achieving effective disentanglement of style and content from reference images.", "raw": "We are happy to introduce our InstantStyle, which is a framework that employs straightforward yet potent techniques for achieving effective disentanglement of style and content from reference images.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Code: ", "raw": "Code: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/InstantStyle/InstantStyle", "href": "https://github.com/InstantStyle/InstantStyle", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Paper: ", "raw": "Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.02733", "href": null, "resource": { "type": "paper", "id": "2404.02733", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.02733", "code": null, "user": null, "label": "InstantStyle: Free Lunch towards Style-Preserving in Text-to-Image\n Generation (2404.02733)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Project Page: ", "raw": "Project Page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://instantstyle.github.io/", "href": "https://instantstyle.github.io/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/637745113a63a2983ffbde13/FK7jtaOPeGq_zLQslSSVs.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/637745113a63a2983ffbde13/3MoFdn2Ckdfw0P6YZ6VaA.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/637745113a63a2983ffbde13/riEozFFvsRDvTxcBo4i5z.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/637745113a63a2983ffbde13/G2G78ZO6-yihHAvNca0Xi.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/637745113a63a2983ffbde13/ertX9VBmlOPPAot5qi_m-.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/637745113a63a2983ffbde13/3-hX8Dr-VMUkNn74_OV5c.png" } ]

[]

[ { "reaction": "🔥", "users": [ "DmitryRyumin", "samusenps", "victor", "merve", "tonyassi", "namelysweet", "thomwolf", "badayvedat", "dillfrescott", "penouc" ], "count": 10 }, { "reaction": "❤️", "users": [ "penouc" ], "count": 1 } ]

[]

[ { "type": "text", "value": "ARABIC CHINESE FRENCH GERMAN RUSSIAN SPANISH TURKISH", "raw": "ARABIC CHINESE FRENCH GERMAN RUSSIAN SPANISH TURKISH", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "mLLM - first release:", "raw": "mLLM - first release:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "orca_dpo_pairs by Intel (translated into 7 languages)", "raw": "orca_dpo_pairs by Intel (translated into 7 languages)", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "ARABIC CHINESE FRENCH GERMAN RUSSIAN SPANISH TURKISH", "raw": "ARABIC CHINESE FRENCH GERMAN RUSSIAN SPANISH TURKISH", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Upcoming:", "raw": "Upcoming:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- more datasets", "raw": "- more datasets", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- cleaning steps", "raw": "- cleaning steps", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- a blogpost", "raw": "- a blogpost", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "- stay updated at ", "raw": "- stay updated at ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://hf.co/multilingual", "href": "https://hf.co/multilingual", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/datasets/multilingual/orca_dpo_pairs", "href": null, "resource": { "type": "dataset", "id": "multilingual/orca_dpo_pairs", "discussionNum": null }, "url": "https://huggingface.co/datasets/multilingual/orca_dpo_pairs", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5f3801ab7e583543386217ac/yVoK7ZaFJTIarQq5Fwrf4.png" } ]

[]

[ { "reaction": "🚀", "users": [ "samusenps", "monsoon-nlp", "seyf1elislam", "aari1995", "osanseviero", "victor", "merve", "nikgr", "Kukedlc" ], "count": 9 }, { "reaction": "❤️", "users": [ "samusenps", "kristaller486", "merve" ], "count": 3 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/630920925a5c889aaedc7f33/w00N19M21l2FXe6ZasSYc.jpeg", "fullname": "Kristaller486", "name": "kristaller486", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 1, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/5f3801ab7e583543386217ac/4xMdDV1gws7nxCJrU321H.jpeg", "fullname": "Aaron Chibb", "name": "aari1995", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 32, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64dbf5901855ce11cd6b9524/ex4dCcI32Zac8IzxNBJY7.jpeg", "fullname": "Felix", "name": "felix-red-panda", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 25, "isFollowing": false }, { "avatarUrl": "/avatars/9af5e9e03b06efced757a62445519545.svg", "fullname": "Polipogu", "name": "Manikyarao", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/60b23b069c978cce68723b25/rFi5BkB8XepyiQ-j4ipdO.jpeg", "fullname": "Wortega", "name": "AlexWortega", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 55, "isFollowing": false } ]

[ { "type": "text", "value": "🆕 Releasing a new series of 8 zeroshot classifiers: better performance, fully commercially useable thanks to synthetic data, up to 8192 tokens, run on any hardware.", "raw": "🆕 Releasing a new series of 8 zeroshot classifiers: better performance, fully commercially useable thanks to synthetic data, up to 8192 tokens, run on any hardware.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Summary: ", "raw": "Summary: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤖 The zeroshot-v2.0-c series replaces commercially restrictive training data with synthetic data generated with mistralai/Mixtral-8x7B-Instruct-v0.1 (Apache 2.0). All models are released under the MIT license.", "raw": "🤖 The zeroshot-v2.0-c series replaces commercially restrictive training data with synthetic data generated with mistralai/Mixtral-8x7B-Instruct-v0.1 (Apache 2.0). All models are released under the MIT license.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🦾 The best model performs 17%-points better across 28 tasks vs. facebook/bart-large-mnli (the most downloaded commercially-friendly baseline).", "raw": "🦾 The best model performs 17%-points better across 28 tasks vs. facebook/bart-large-mnli (the most downloaded commercially-friendly baseline).", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌏 The series includes a multilingual variant fine-tuned from BAAI/bge-m3 for zeroshot classification in 100+ languages and with a context window of 8192 tokens ", "raw": "🌏 The series includes a multilingual variant fine-tuned from BAAI/bge-m3 for zeroshot classification in 100+ languages and with a context window of 8192 tokens ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🪶 The models are 0.2 - 0.6 B parameters small, so they run on any hardware. The base-size models are +2x faster than bart-large-mnli while performing significantly better. ", "raw": "🪶 The models are 0.2 - 0.6 B parameters small, so they run on any hardware. The base-size models are +2x faster than bart-large-mnli while performing significantly better. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤏 The models are not generative LLMs, they are efficient encoder-only models specialized in zeroshot classification through the universal NLI task. ", "raw": "🤏 The models are not generative LLMs, they are efficient encoder-only models specialized in zeroshot classification through the universal NLI task. ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤑 For users where commercially restrictive training data is not an issue, I've also trained variants with even more human data for improved performance.", "raw": "🤑 For users where commercially restrictive training data is not an issue, I've also trained variants with even more human data for improved performance.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Next steps: ", "raw": "Next steps: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "✍️ I'll publish a blog post with more details soon", "raw": "✍️ I'll publish a blog post with more details soon", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔮 There are several improvements I'm planning for v2.1. Especially the multilingual model has room for improvement.", "raw": "🔮 There are several improvements I'm planning for v2.1. Especially the multilingual model has room for improvement.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "All models are available for download in this Hugging Face collection: ", "raw": "All models are available for download in this Hugging Face collection: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f", "href": null, "resource": { "type": "collection", "id": "MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f", "discussionNum": null }, "url": "https://huggingface.co/collections/MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "These models are an extension of the approach explained in this paper, but with additional synthetic data: ", "raw": "These models are an extension of the approach explained in this paper, but with additional synthetic data: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://arxiv.org/abs/2312.17543", "href": "https://arxiv.org/abs/2312.17543", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/5fb15d1e84389b139cf3b508/Q_BDehoaByuTDEO1rpWIJ.png" } ]

[]

[ { "reaction": "🔥", "users": [ "victor", "tomaarsen", "samusenps", "clem", "vladbogo", "jeffboudier", "mmhamdy", "osanseviero", "peter2000", "merve", "peycke" ], "count": 11 }, { "reaction": "😎", "users": [ "samusenps", "clem", "osanseviero", "merve", "M1cler" ], "count": 5 }, { "reaction": "🚀", "users": [ "Benjoyo" ], "count": 1 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6317233cc92fd6fee317e030/cJHSvvimr1kqgQfHOjO5n.png", "fullname": "Tom Aarsen", "name": "tomaarsen", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 1060, "isFollowing": false }, { "avatarUrl": "/avatars/c848a4d037da8d5773098e2cdc70ac39.svg", "fullname": "Yong Hye Ryeon", "name": "HAMRONI", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1613511937628-5fb15d1e84389b139cf3b508.jpeg", "fullname": "Moritz Laurer", "name": "MoritzLaurer", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 236, "isFollowing": false } ]

[ { "type": "text", "value": "👋 👋 👋 In a recent AI Time Journal Interview, I was asked: How has HackerNoon leveraged AI to enhance its platform? and what lessons have you learned about AI’s role in spam creation and prevention? ", "raw": "👋 👋 👋 In a recent AI Time Journal Interview, I was asked: How has HackerNoon leveraged AI to enhance its platform? and what lessons have you learned about AI’s role in spam creation and prevention? ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://HackerNoon.com", "href": "https://HackerNoon.com", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " uses AI and machine learning to improve grammar, translate stories, generate featured images, suggest possible headlines, (and note, I would love to hear what else you think we should use AI for!) . Maybe Clippy was ahead of its time? Writers have and will continue to have smart AI assistants. AI can not and should not replace the writer, but it should and can assist the writer and editor throughout the publishing process. We are documenting our approach to editorial at ", "raw": " uses AI and machine learning to improve grammar, translate stories, generate featured images, suggest possible headlines, (and note, I would love to hear what else you think we should use AI for!) . Maybe Clippy was ahead of its time? Writers have and will continue to have smart AI assistants. AI can not and should not replace the writer, but it should and can assist the writer and editor throughout the publishing process. We are documenting our approach to editorial at ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://EditingProtocol.com", "href": "https://EditingProtocol.com", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": ". With the rise of LLMs, we’ve seen an influx of AI story spam submissions. Other blogging platforms are hosting so many of these types of machine-generated stories right now. We are actively taking steps not to become a wasteland of AI’s musings. Because every story is reviewed before publication (with about half being rejected), I think our average story quality is higher than other blogging platforms. Every story submission automatically enters into plagiarism and AI writing detection. We highlight what sections are likely, possibly, and unlikely to be written by AI. It’s all about confidence levels. Ultimately, human editors make the final call in interpreting the machines’ reports.", "raw": ". With the rise of LLMs, we’ve seen an influx of AI story spam submissions. Other blogging platforms are hosting so many of these types of machine-generated stories right now. We are actively taking steps not to become a wasteland of AI’s musings. Because every story is reviewed before publication (with about half being rejected), I think our average story quality is higher than other blogging platforms. Every story submission automatically enters into plagiarism and AI writing detection. We highlight what sections are likely, possibly, and unlikely to be written by AI. It’s all about confidence levels. Ultimately, human editors make the final call in interpreting the machines’ reports.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🔥", "users": [ "julien-c", "clem", "osanseviero", "merve" ], "count": 4 }, { "reaction": "❤️", "users": [ "samusenps", "kramp", "osanseviero", "merve" ], "count": 4 }, { "reaction": "➕", "users": [ "samusenps", "clem", "merve" ], "count": 3 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64862a25cf5ad5e1f0482ef2/61qPUtw9jIl7zpPYmi0VW.jpeg", "fullname": "David Smooke", "name": "Smooke", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 43, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6538119803519fddb4a17e10/ffJMkdx-rM7VvLTCM6ri_.jpeg", "fullname": "samusenps", "name": "samusenps", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 91, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1583857146757-5e67bdd61009063689407479.jpeg", "fullname": "Clem 🤗", "name": "clem", "type": "user", "isPro": true, "isHf": true, "isMod": false, "followerCount": 1763, "isFollowing": false } ]

[ { "type": "text", "value": "Advancing LLM Reasoning Generalists with Preference Trees", "raw": "Advancing LLM Reasoning Generalists with Preference Trees", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.02078", "href": null, "resource": { "type": "paper", "id": "2404.02078", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.02078", "code": null, "user": null, "label": "Advancing LLM Reasoning Generalists with Preference Trees (2404.02078)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We introduce Eurus, a suite of large language models (LLMs) optimized for reasoning. Finetuned from Mistral-7B and CodeLlama-70B, Eurus models achieve state-of-the-art results among open-source models on a diverse set of benchmarks covering mathematics, code generation, and logical reasoning problems. Notably, Eurus-70B beats GPT-3.5 Turbo in reasoning through a comprehensive benchmarking across 12 tests covering five tasks, and achieves a 33.3% pass@1 accuracy on LeetCode and 32.6% on TheoremQA, two challenging benchmarks, substantially outperforming existing open-source models by margins more than 13.3%. The strong performance of Eurus can be primarily attributed to UltraInteract, our newly-curated large-scale, high-quality alignment dataset specifically designed for complex reasoning tasks. UltraInteract can be used in both supervised fine-tuning and preference learning. For each instruction, it includes a preference tree consisting of (1) reasoning chains with diverse planning strategies in a unified format, (2) multi-turn interaction trajectories with the environment and the critique, and (3) pairwise data to facilitate preference learning. UltraInteract allows us to conduct an in-depth exploration of preference learning for reasoning tasks. Our investigation reveals that some well-established preference learning algorithms may be less suitable for reasoning tasks compared to their effectiveness in general conversations. Inspired by this, we derive a novel reward modeling objective which, together with UltraInteract, leads to a strong reward model.", "raw": "We introduce Eurus, a suite of large language models (LLMs) optimized for reasoning. Finetuned from Mistral-7B and CodeLlama-70B, Eurus models achieve state-of-the-art results among open-source models on a diverse set of benchmarks covering mathematics, code generation, and logical reasoning problems. Notably, Eurus-70B beats GPT-3.5 Turbo in reasoning through a comprehensive benchmarking across 12 tests covering five tasks, and achieves a 33.3% pass@1 accuracy on LeetCode and 32.6% on TheoremQA, two challenging benchmarks, substantially outperforming existing open-source models by margins more than 13.3%. The strong performance of Eurus can be primarily attributed to UltraInteract, our newly-curated large-scale, high-quality alignment dataset specifically designed for complex reasoning tasks. UltraInteract can be used in both supervised fine-tuning and preference learning. For each instruction, it includes a preference tree consisting of (1) reasoning chains with diverse planning strategies in a unified format, (2) multi-turn interaction trajectories with the environment and the critique, and (3) pairwise data to facilitate preference learning. UltraInteract allows us to conduct an in-depth exploration of preference learning for reasoning tasks. Our investigation reveals that some well-established preference learning algorithms may be less suitable for reasoning tasks compared to their effectiveness in general conversations. Inspired by this, we derive a novel reward modeling objective which, together with UltraInteract, leads to a strong reward model.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/60f1abe7544c2adfd699860c/rKblhohR_WI-BCQ0XM6lr.png" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "clem", "vladbogo", "merve", "nikgr", "Bruno" ], "count": 6 } ]

[]

[ { "type": "text", "value": "Introducing VLRM: Vision-Language Models act as Reward Models for Image Captioning", "raw": "Introducing VLRM: Vision-Language Models act as Reward Models for Image Captioning", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Paper:", "raw": "Paper:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://arxiv.org/abs/2404.01911", "href": "https://arxiv.org/abs/2404.01911", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Model weights and training logs:", "raw": "Model weights and training logs:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/sashakunitsyn/vlrm-blip2-opt-2.7b", "href": null, "resource": { "type": "model", "id": "sashakunitsyn/vlrm-blip2-opt-2.7b", "discussionNum": null }, "url": "https://huggingface.co/sashakunitsyn/vlrm-blip2-opt-2.7b", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "We propose VLRM, a method for fine-tuning an existing image captioning model using reinforcement learning and vision-language models as reward models.", "raw": "We propose VLRM, a method for fine-tuning an existing image captioning model using reinforcement learning and vision-language models as reward models.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The method manages to significantly improve the generation quality without human-labeled data and is applicable to any image captioning model.", "raw": "The method manages to significantly improve the generation quality without human-labeled data and is applicable to any image captioning model.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Our model reaches impressive 0.90 R@1 CLIP Recall score on MS-COCO Carpathy Test Split.", "raw": "Our model reaches impressive 0.90 R@1 CLIP Recall score on MS-COCO Carpathy Test Split.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/635799e5265b76ad9edef8d9/6YKgoME5ZiRUmc_dyGlcB.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/635799e5265b76ad9edef8d9/7h45RN8KQp6nCw0dmeZCH.jpeg" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/635799e5265b76ad9edef8d9/jYnzcNiAkq32avwd4DA_z.jpeg" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "nikgr", "victor", "clem", "osanseviero", "merve", "Kukedlc", "AlkaouDembele", "xziayro" ], "count": 9 }, { "reaction": "➕", "users": [ "samusenps", "merve" ], "count": 2 } ]

[]

[ { "type": "text", "value": "🚀🎭🌟 New Research Alert - CVPR 2024 (Avatars Collection)! 🌟 🎭🚀", "raw": "🚀🎭🌟 New Research Alert - CVPR 2024 (Avatars Collection)! 🌟 🎭🚀", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📄 Title: MonoAvatar++: Efficient 3D Implicit Head Avatar with Mesh-anchored Hash Table Blendshapes 🔝", "raw": "📄 Title: MonoAvatar++: Efficient 3D Implicit Head Avatar with Mesh-anchored Hash Table Blendshapes 🔝", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📝 Description: MonoAvatar++ is a real-time neural implicit 3D head avatar model with high quality and fine-grained control over facial expressions. It uses local hash table blendshapes attached to a parametric facial model for efficient rendering, achieving SOTA results even for challenging expressions.", "raw": "📝 Description: MonoAvatar++ is a real-time neural implicit 3D head avatar model with high quality and fine-grained control over facial expressions. It uses local hash table blendshapes attached to a parametric facial model for efficient rendering, achieving SOTA results even for challenging expressions.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "👥 Authors: Ziqian Bai, Feitong Tan, Sean Fanello, Rohit Pandey, Mingsong Dou, Shichen Liu, Ping Tan, Yinda Zhang", "raw": "👥 Authors: Ziqian Bai, Feitong Tan, Sean Fanello, Rohit Pandey, Mingsong Dou, Shichen Liu, Ping Tan, Yinda Zhang", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📅 Conference: CVPR, Jun 17-21, 2024 | Seattle WA, USA 🇺🇸", "raw": "📅 Conference: CVPR, Jun 17-21, 2024 | Seattle WA, USA 🇺🇸", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔗 Paper: ", "raw": "🔗 Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/papers/2404.01543", "href": null, "resource": { "type": "paper", "id": "2404.01543", "discussionNum": null }, "url": "https://huggingface.co/papers/2404.01543", "code": null, "user": null, "label": "Efficient 3D Implicit Head Avatar with Mesh-anchored Hash Table\n Blendshapes (2404.01543)", "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🌐 Github Page: ", "raw": "🌐 Github Page: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://augmentedperception.github.io/monoavatar-plus", "href": "https://augmentedperception.github.io/monoavatar-plus", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "raw": "📚 More Papers: more cutting-edge research presented at other conferences in the ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "href": null, "resource": { "type": "space", "id": "DmitryRyumin/NewEraAI-Papers", "discussionNum": null }, "url": "https://huggingface.co/spaces/DmitryRyumin/NewEraAI-Papers", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " curated by ", "raw": " curated by ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@DmitryRyumin", "href": null, "resource": null, "url": null, "code": null, "user": "DmitryRyumin", "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🚀 Added to the Avatars Collection: ", "raw": "🚀 Added to the Avatars Collection: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "href": null, "resource": { "type": "collection", "id": "DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "discussionNum": null }, "url": "https://huggingface.co/collections/DmitryRyumin/avatars-65df37cdf81fec13d4dbac36", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔍 Keywords: #MonoAvatar++ #HeadAvatar #3DModeling #AvatarGeneration #NeuralImplicitAvatar #EfficientRendering #CVPR2024 #DeepLearning #Innovation", "raw": "🔍 Keywords: #MonoAvatar++ #HeadAvatar #3DModeling #AvatarGeneration #NeuralImplicitAvatar #EfficientRendering #CVPR2024 #DeepLearning #Innovation", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/nNG75t6Rc_Coj7Fs658Sw.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/UtYqifxl_uzO07ARNqyWt.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/A0NILX7xpqRhTwWViF9_j.png" }, { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/MhRNn1SLzXBJglXyF4FR4.mp4" }, { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/JUDuoxySx8QylHpvnRjwD.mp4" }, { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/j9vYHcLIBA_GGOinDLwuc.mp4" }, { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/uRKakPAUDxuBE8QjwDC3H.mp4" }, { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/XWUUsfSTSBnGAcYJTFOwI.mp4" }, { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/ahyjDg-c9G19hmM5rLrM-.mp4" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/eWBeOl0oPw1G6V-TOxrUv.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/D1fN9eq3dOY8EjIbQQKR1.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/-hvBLUL04WlLSSgBqtIAO.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/oc3KUiYGc0xNUj7I8jLRo.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6493306970d925ae80523a53/ROssTa6e0Dx11xgOJWeeI.png" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/nRCxbVng_PPBqKd-Z3KVc.jpeg", "fullname": "Dmitry Ryumin", "name": "DmitryRyumin", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 377 } ]

[ { "reaction": "🔥", "users": [ "DmitryRyumin", "samusenps", "nikgr", "victor", "clem", "merve" ], "count": 6 }, { "reaction": "🤯", "users": [ "DmitryRyumin", "clem", "merve" ], "count": 3 }, { "reaction": "🚀", "users": [ "samusenps", "merve" ], "count": 2 }, { "reaction": "❤️", "users": [ "samusenps" ], "count": 1 } ]

[]

[ { "type": "text", "value": "ChemLLM version 1.5 has been open sourced. It has also released a version that has undergone human feedback reinforcement learning. In particular, the front-end and back-end codes have been open sourced simultaneously to facilitate deployment. Everyone is welcome to pay attention.", "raw": "ChemLLM version 1.5 has been open sourced. It has also released a version that has undergone human feedback reinforcement learning. In particular, the front-end and back-end codes have been open sourced simultaneously to facilitate deployment. Everyone is welcome to pay attention.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "The evaluation set and part of the training set will also be contributed to the community in days, so stay tuned ✧٩(ˊωˋ*)و✧", "raw": "The evaluation set and part of the training set will also be contributed to the community in days, so stay tuned ✧٩(ˊωˋ*)و✧", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "ChemLLM-1.5", "raw": "ChemLLM-1.5", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/AI4Chem/ChemLLM-7B-Chat-1.5-DPO", "href": "https://huggingface.co/AI4Chem/ChemLLM-7B-Chat-1.5-DPO", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "ChemLLM-1.5-dpo ", "raw": "ChemLLM-1.5-dpo ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/AI4Chem/ChemLLM-7B-Chat-1.5-SFT", "href": "https://huggingface.co/AI4Chem/ChemLLM-7B-Chat-1.5-SFT", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Front-end ", "raw": "Front-end ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/AI4Chem/NextChat", "href": "https://github.com/AI4Chem/NextChat", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "RAG application example ", "raw": "RAG application example ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/AI4Chem/NextChat/blob/main/chatpdf.py", "href": "https://github.com/AI4Chem/NextChat/blob/main/chatpdf.py", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Backend ", "raw": "Backend ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/AI4Chem/api_backend", "href": "https://github.com/AI4Chem/api_backend", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Front-end Demo ", "raw": "Front-end Demo ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://chemllm.org/", "href": "https://chemllm.org/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "RAG application example Demo ", "raw": "RAG application example Demo ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://chatpdf.chemllm.org/", "href": "https://chatpdf.chemllm.org/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Backend Demo ", "raw": "Backend Demo ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://api.chemllm.org/", "href": "https://api.chemllm.org/", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "🤗", "users": [ "qq8933", "osanseviero", "ajibawa-2023", "samusenps", "nikgr", "clem", "merve", "mmhamdy", "dillfrescott" ], "count": 9 }, { "reaction": "🚀", "users": [ "samusenps", "clem", "merve", "dillfrescott" ], "count": 4 }, { "reaction": "😎", "users": [ "samusenps", "clem", "merve", "dillfrescott" ], "count": 4 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64aea8ff67511bd3d965697b/Jxn52EmDF5RApJh8antxn.jpeg", "fullname": "Feynman Innovations", "name": "ajibawa-2023", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 138, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64bce15bafd1e46c5504ad38/bQFX1iFbXEBXcQvUNL811.png", "fullname": "Di Zhang", "name": "qq8933", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 108, "isFollowing": false }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1652197428055-625a6e0c535747b1a15be2de.jpeg", "fullname": "Md. Musfiqur Rahaman", "name": "musfiqdehan", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 5, "isFollowing": false } ]

[ { "type": "text", "value": "Following up on ", "raw": "Following up on ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@vikhyatk", "href": null, "resource": null, "url": null, "code": null, "user": "vikhyatk", "label": null, "lang": null }, { "type": "text", "value": "'s Moondream2 update and ", "raw": "'s Moondream2 update and ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@santiagomed", "href": null, "resource": null, "url": null, "code": null, "user": "santiagomed", "label": null, "lang": null }, { "type": "text", "value": "'s implementation on Candle, I quickly put togheter the WASM module so that you could try running the ~1.5GB quantized model in the browser. Perhaps the next step is to rewrite it using ", "raw": "'s implementation on Candle, I quickly put togheter the WASM module so that you could try running the ~1.5GB quantized model in the browser. Perhaps the next step is to rewrite it using ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://github.com/huggingface/ratchet", "href": "https://github.com/huggingface/ratchet", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " and run it even faster with WebGPU, ", "raw": " and run it even faster with WebGPU, ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@FL33TW00D-HF", "href": null, "resource": null, "url": null, "code": null, "user": "FL33TW00D-HF", "label": null, "lang": null }, { "type": "text", "value": ".", "raw": ".", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/spaces/radames/Candle-Moondream-2", "href": null, "resource": { "type": "space", "id": "radames/Candle-Moondream-2", "discussionNum": null }, "url": "https://huggingface.co/spaces/radames/Candle-Moondream-2", "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "ps: I have a collection of all Candle WASM demos here ", "raw": "ps: I have a collection of all Candle WASM demos here ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/radames/candle-wasm-examples-650898dee13ff96230ce3e1f", "href": null, "resource": { "type": "collection", "id": "radames/candle-wasm-examples-650898dee13ff96230ce3e1f", "discussionNum": null }, "url": "https://huggingface.co/collections/radames/candle-wasm-examples-650898dee13ff96230ce3e1f", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/6064e095abd8d3692e3e2ed6/_fO2jLfq4l_37kTcxxbMf.jpeg" } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/6597e9f42235d4056bc6980a/6N_Eira5Rj5e8ZdgekKPQ.jpeg", "fullname": "Christopher Fleetwood", "name": "FL33TW00D-HF", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 54 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64f2c90c1ae35aedb8df2b27/ng37Jo7jZJ6Ze9JsY9PoN.jpeg", "fullname": "Santiago Medina", "name": "santiagomed", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 13 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/63117568fa95534e218da163/8h9zN8aKRxPLBnXW7sqY9.jpeg", "fullname": "Vik Korrapati", "name": "vikhyatk", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 375 } ]

[ { "reaction": "❤️", "users": [ "vikhyatk", "FL33TW00D-HF", "samusenps", "nikgr", "clem", "merve", "Felladrin", "osanseviero" ], "count": 8 }, { "reaction": "🔥", "users": [ "vikhyatk", "FL33TW00D-HF", "clem", "merve", "kylen", "osanseviero" ], "count": 6 } ]

[]

[ { "type": "text", "value": "Gradio 4.25.0 is out with some nice improvements and bug fixes!", "raw": "Gradio 4.25.0 is out with some nice improvements and bug fixes!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🧹 Automatic deletion of gr.State variables stored in the server. Never run out of RAM again. Also adds an unload event you can run when a user closes their browser tab.", "raw": "🧹 Automatic deletion of gr.State variables stored in the server. Never run out of RAM again. Also adds an unload event you can run when a user closes their browser tab.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "😴 Lazy example caching. You can set cache_examples=\"lazy\" to cache examples when they're first requested as opposed to before the server launches. This can cut down the server's start-up time drastically.", "raw": "😴 Lazy example caching. You can set cache_examples=\"lazy\" to cache examples when they're first requested as opposed to before the server launches. This can cut down the server's start-up time drastically.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🔊 Fixes a bug with streaming audio outputs", "raw": "🔊 Fixes a bug with streaming audio outputs", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "🤖 Improvements to gr.ChatInterface like pasting images directly from the clipboard.", "raw": "🤖 Improvements to gr.ChatInterface like pasting images directly from the clipboard.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "See the rest of the changelog here: ", "raw": "See the rest of the changelog here: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://www.gradio.app/changelog#4-25-0", "href": "https://www.gradio.app/changelog#4-25-0", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "ysharma", "not-lain", "ajibawa-2023", "radames", "DmitryRyumin", "nikgr", "clem", "clefourrier", "akhaliq", "on1onmangoes" ], "count": 11 }, { "reaction": "➕", "users": [ "samusenps", "ysharma", "clem", "abidlabs", "akhaliq" ], "count": 5 }, { "reaction": "🚀", "users": [ "ysharma", "radames", "clem", "abidlabs", "akhaliq" ], "count": 5 } ]

[]

[ { "type": "text", "value": "Released a new version of ", "raw": "Released a new version of ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/vikhyatk/moondream2", "href": null, "resource": { "type": "model", "id": "vikhyatk/moondream2", "discussionNum": null }, "url": "https://huggingface.co/vikhyatk/moondream2", "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": " today! Primarily focused on improving OCR and captioning (e.g. \"Describe this image\", \"Describe this image in one sentence\"), but also seeing general improvement across all benchmarks.", "raw": " today! Primarily focused on improving OCR and captioning (e.g. \"Describe this image\", \"Describe this image in one sentence\"), but also seeing general improvement across all benchmarks.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/63117568fa95534e218da163/WvKqH1EoqmKvlGD3XFtlo.png" } ]

[]

[ { "reaction": "🔥", "users": [ "not-lain", "Corianas", "ajibawa-2023", "radames", "DmitryRyumin", "nikgr", "clem", "xingren23", "dillfrescott", "aaryanverma", "linz", "Jorgegarciairazabal", "Vlansu" ], "count": 13 }, { "reaction": "🚀", "users": [ "not-lain", "clem", "sooperDime", "Sylvestre", "dillfrescott", "omaryshchenko", "Jorgegarciairazabal" ], "count": 7 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/64aea8ff67511bd3d965697b/Jxn52EmDF5RApJh8antxn.jpeg", "fullname": "Feynman Innovations", "name": "ajibawa-2023", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 138, "isFollowing": false } ]

[ { "type": "text", "value": "These 15 open models are available for serverless inference on Cloudflare Workers AI, powered by GPUs distributed in 150 datacenters globally - 👏 ", "raw": "These 15 open models are available for serverless inference on Cloudflare Workers AI, powered by GPUs distributed in 150 datacenters globally - 👏 ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@rita3ko", "href": null, "resource": null, "url": null, "code": null, "user": "rita3ko", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@mchenco", "href": null, "resource": null, "url": null, "code": null, "user": "mchenco", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@jtkipp", "href": null, "resource": null, "url": null, "code": null, "user": "jtkipp", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@nkothariCF", "href": null, "resource": null, "url": null, "code": null, "user": "nkothariCF", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "mention", "value": null, "raw": "@philschmid", "href": null, "resource": null, "url": null, "code": null, "user": "philschmid", "label": null, "lang": null }, { "type": "text", "value": " ", "raw": " ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "resource", "value": null, "raw": "https://huggingface.co/collections/Cloudflare/hf-curated-models-available-on-workers-ai-66036e7ad5064318b3e45db6", "href": null, "resource": { "type": "collection", "id": "Cloudflare/hf-curated-models-available-on-workers-ai-66036e7ad5064318b3e45db6", "discussionNum": null }, "url": "https://huggingface.co/collections/Cloudflare/hf-curated-models-available-on-workers-ai-66036e7ad5064318b3e45db6", "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "video", "url": "https://cdn-uploads.huggingface.co/production/uploads/5fac18fb5eec0323e9470ba2/Dd5igw7sDsKjo75l4gLDr.mp4" } ]

[ { "avatarUrl": "/avatars/55bc85ffea21b0a46fcb3bce77bb4779.svg", "fullname": "Jesse Kipp", "name": "jtkipp", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 3 }, { "avatarUrl": "/avatars/0c96b358e25d8834c086e3b9b5d543eb.svg", "fullname": "Michelle Chen", "name": "mchenco", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 5 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/65416f1741676ceaa2c14c58/PJMka1bOP20Jk3ZnnG7us.jpeg", "fullname": "Nikhil Kothari", "name": "nkothariCF", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 4 }, { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/1624629516652-5ff5d596f244529b3ec0fb89.png", "fullname": "Philipp Schmid", "name": "philschmid", "type": "user", "isPro": false, "isHf": true, "isMod": false, "followerCount": 657 }, { "avatarUrl": "/avatars/a022460e9db28ddf363e65ce3171453b.svg", "fullname": "Rita Kozlov", "name": "rita3ko", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": 4 } ]

[ { "reaction": "🔥", "users": [ "rwightman", "nkothariCF", "samusenps", "ajibawa-2023", "philschmid", "radames", "victor", "nikgr", "kramp", "clem", "Tonic" ], "count": 11 }, { "reaction": "❤️", "users": [ "samusenps", "philschmid", "radames", "victor", "clem", "Tonic" ], "count": 6 } ]

[]

[ { "type": "text", "value": "Anthropic introduces \"Many-shot Jailbreaking\" (MSJ), a new attack on large language models! MSJ exploits long context windows to override safety constraints.", "raw": "Anthropic introduces \"Many-shot Jailbreaking\" (MSJ), a new attack on large language models! MSJ exploits long context windows to override safety constraints.", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Key Points:", "raw": "Key Points:", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "* Prompts LLMs with hundreds of examples of harmful behavior formatted as a dialogue", "raw": "* Prompts LLMs with hundreds of examples of harmful behavior formatted as a dialogue", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "* Generates malicious examples using an uninhibited \"helpful-only\" model", "raw": "* Generates malicious examples using an uninhibited \"helpful-only\" model", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "* Effective at jailbreaking models like Claude 2.0, GPT-3.5, GPT-4", "raw": "* Effective at jailbreaking models like Claude 2.0, GPT-3.5, GPT-4", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "* Standard alignment techniques provide limited protection against long context attacks", "raw": "* Standard alignment techniques provide limited protection against long context attacks", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Paper: ", "raw": "Paper: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://www.anthropic.com/research/many-shot-jailbreaking", "href": "https://www.anthropic.com/research/many-shot-jailbreaking", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "More details in my blog: ", "raw": "More details in my blog: ", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "link", "value": null, "raw": "https://huggingface.co/blog/vladbogo/many-shot-jailbreaking", "href": "https://huggingface.co/blog/vladbogo/many-shot-jailbreaking", "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "new_line", "value": null, "raw": "\n", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null }, { "type": "text", "value": "Congrats to the authors for their work!", "raw": "Congrats to the authors for their work!", "href": null, "resource": null, "url": null, "code": null, "user": null, "label": null, "lang": null } ]

[ { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/657217faabb25ed8aedd5e48/W94Ey2-cITOHk--mtzt1J.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/657217faabb25ed8aedd5e48/od61V0DfaYOJmFS8GUX8T.png" }, { "type": "image", "url": "https://cdn-uploads.huggingface.co/production/uploads/657217faabb25ed8aedd5e48/6_80HmNX86HA5ZCY13HHz.png" } ]

[]

[ { "reaction": "❤️", "users": [ "samusenps", "nikgr", "gruber", "sauravssss" ], "count": 4 }, { "reaction": "👀", "users": [ "monsoon-nlp", "gruber" ], "count": 2 } ]

[ { "avatarUrl": "https://cdn-avatars.huggingface.co/v1/production/uploads/634262af8d8089ebaefd410e/pr6KcEebXTo5V2XAlpQNw.png", "fullname": "Fizz 🏳️‍⚧️", "name": "Fizzarolli", "type": "user", "isPro": true, "isHf": false, "isMod": false, "followerCount": 49, "isFollowing": false }, { "avatarUrl": "/avatars/eac1b6903a19bf24057e914b03796f46.svg", "fullname": "Saurav singh", "name": "sauravssss", "type": "user", "isPro": false, "isHf": false, "isMod": false, "followerCount": null, "isFollowing": false } ]