modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-07-30 18:29:32
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 538
values | tags
listlengths 1
4.05k
| pipeline_tag
stringclasses 55
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-07-30 18:29:11
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
Shishir17/gemma-3-finetune-SQL-float16
|
Shishir17
| 2025-07-26T11:13:50Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"gemma3n",
"image-text-to-text",
"text-generation-inference",
"unsloth",
"conversational",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2025-07-26T11:09:07Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Shishir17/gemma-3-finetune-SQL-float16/4816b04e9a61653ebab233b36a02f17a3849f29a/README.md?%2FShishir17%2Fgemma-3-finetune-SQL-float16%2Fresolve%2Fmain%2FREADME.md=&etag=%22a3a7b6663f076c2ec945565f818872ffbfb52539%22
|
Not-JaysonYT/Juno
|
Not-JaysonYT
| 2025-07-26T10:56:34Z | 0 | 1 | null |
[
"license:apache-2.0",
"region:us"
] | null | 2025-07-25T14:29:35Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Not-JaysonYT/Juno/9f5c2f2c8c63cf34ad64d7b9cc238a949254e5e4/README.md?%2FNot-JaysonYT%2FJuno%2Fresolve%2Fmain%2FREADME.md=&etag=%22d3571cc81704ecf4af144edeacc82b9ad142d32c%22
|
verba-ai/verba.v1-caramel
|
verba-ai
| 2025-07-26T10:23:26Z | 0 | 0 |
transformers
|
[
"transformers",
"neural-network",
"custom-model",
"wikipedia",
"russian",
"from-scratch",
"text-generation",
"ru",
"dataset:wikimedia/wikipedia",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-26T09:53:46Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/verba-ai/verba.v1-caramel/6c3596530e8f2d93a790b7d9b005173f2196cd98/README.md?%2Fverba-ai%2Fverba.v1-caramel%2Fresolve%2Fmain%2FREADME.md=&etag=%222e27bd28d636ceb6d59acc1060d659d7be93f685%22
|
ACECA/lowMvM_104
|
ACECA
| 2025-07-26T09:42:40Z | 0 | 0 | null |
[
"safetensors",
"any-to-any",
"omega",
"omegalabs",
"bittensor",
"agi",
"license:mit",
"region:us"
] |
any-to-any
| 2025-07-20T08:04:09Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/ACECA/lowMvM_104/13cd13a7b54a25525a2b56b36c04e2c1319a2535/README.md?%2FACECA%2FlowMvM_104%2Fresolve%2Fmain%2FREADME.md=&etag=%2213a6bd2875d0f87525e8d8c5b1cf31c9ccc08b34%22
|
configint/SmolVLM2-256M-Video-Instruct-ActionTokens
|
configint
| 2025-07-26T09:36:48Z | 0 | 0 |
transformers
|
[
"transformers",
"onnx",
"safetensors",
"smolvlm",
"image-text-to-text",
"conversational",
"en",
"dataset:HuggingFaceM4/the_cauldron",
"dataset:HuggingFaceM4/Docmatix",
"dataset:lmms-lab/LLaVA-OneVision-Data",
"dataset:lmms-lab/M4-Instruct-Data",
"dataset:HuggingFaceFV/finevideo",
"dataset:MAmmoTH-VL/MAmmoTH-VL-Instruct-12M",
"dataset:lmms-lab/LLaVA-Video-178K",
"dataset:orrzohar/Video-STaR",
"dataset:Mutonix/Vript",
"dataset:TIGER-Lab/VISTA-400K",
"dataset:Enxin/MovieChat-1K_train",
"dataset:ShareGPT4Video/ShareGPT4Video",
"arxiv:2504.05299",
"base_model:HuggingFaceTB/SmolVLM-256M-Instruct",
"base_model:quantized:HuggingFaceTB/SmolVLM-256M-Instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2025-07-26T09:33:09Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/configint/SmolVLM2-256M-Video-Instruct-ActionTokens/e6e0901170e1823499beccbb8d21dfe09c6860d6/README.md?%2Fconfigint%2FSmolVLM2-256M-Video-Instruct-ActionTokens%2Fresolve%2Fmain%2FREADME.md=&etag=%22769933086ed35f244339acafd5daa76d653a5b62%22
|
adalat-ai/curated-data-whisper-medium-ml-exp2-1
|
adalat-ai
| 2025-07-26T09:14:59Z | 0 | 0 | null |
[
"tensorboard",
"safetensors",
"whisper",
"region:us"
] | null | 2025-07-26T09:12:42Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/adalat-ai/curated-data-whisper-medium-ml-exp2-1/d2d74613f9765debd5b3a86baa258a42a255dff8/README.md?%2Fadalat-ai%2Fcurated-data-whisper-medium-ml-exp2-1%2Fresolve%2Fmain%2FREADME.md=&etag=%22e69de29bb2d1d6434b8b29ae775ad8c2e48c5391%22
|
versaceeros/aed9b9a2-8ace-4beb-ac6a-2763e651f943
|
versaceeros
| 2025-07-26T08:45:03Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-07-26T08:06:19Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/versaceeros/aed9b9a2-8ace-4beb-ac6a-2763e651f943/44ad968257df90bdb5b8d6a425c7e6337a4425d7/README.md?%2Fversaceeros%2Faed9b9a2-8ace-4beb-ac6a-2763e651f943%2Fresolve%2Fmain%2FREADME.md=&etag=%22b7441640181b06c90effb718230b043e35476812%22
|
RichardErkhov/neuralmagic_-_Sparse-Llama-3.1-8B-2of4-4bits
|
RichardErkhov
| 2025-07-26T08:33:39Z | 0 | 0 | null |
[
"safetensors",
"llama",
"arxiv:2301.00774",
"arxiv:2310.06927",
"4-bit",
"bitsandbytes",
"region:us"
] | null | 2025-07-26T08:31:27Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/RichardErkhov/neuralmagic_-_Sparse-Llama-3.1-8B-2of4-4bits/84df14eac786a0dcb4563733a20d7ed0107ebd75/README.md?%2FRichardErkhov%2Fneuralmagic_-_Sparse-Llama-3.1-8B-2of4-4bits%2Fresolve%2Fmain%2FREADME.md=&etag=%22a745d303a97636c639ba83cdfb4a38a17d0f0632%22
|
AgAd2022/pricer-quantile-experiments-50K-Attention-K19-v6
|
AgAd2022
| 2025-07-26T08:12:41Z | 0 | 0 |
transformers
|
[
"transformers",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-07-26T08:12:38Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/AgAd2022/pricer-quantile-experiments-50K-Attention-K19-v6/fb2c80019dcb4a1cef3d25f067456c58fbd4a8a5/README.md?%2FAgAd2022%2Fpricer-quantile-experiments-50K-Attention-K19-v6%2Fresolve%2Fmain%2FREADME.md=&etag=%22bc5f30d6632ac0efdc7be2e9095e9e9579af2e33%22
|
QuixiAI/Kimi-K2-Base-BF16
|
QuixiAI
| 2025-07-26T06:03:20Z | 11 | 0 |
transformers
|
[
"transformers",
"safetensors",
"kimi_k2",
"text-generation",
"conversational",
"custom_code",
"license:other",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-20T00:43:21Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/QuixiAI/Kimi-K2-Base-BF16/f1b1c357b3e4d0266759c3670b3f5a4854f66f07/README.md?%2FQuixiAI%2FKimi-K2-Base-BF16%2Fresolve%2Fmain%2FREADME.md=&etag=%229b9ed58a0a20a1e420c73b9a67585460fc6ee701%22
|
rakib730/vit-base-oxford-iiit-pets
|
rakib730
| 2025-07-26T05:32:09Z | 4 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"vit",
"image-classification",
"generated_from_trainer",
"vision",
"fine-tuned",
"oxford-iiit-pets",
"pytorch",
"base_model:google/vit-base-patch16-224",
"base_model:finetune:google/vit-base-patch16-224",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2025-06-08T13:34:58Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/rakib730/vit-base-oxford-iiit-pets/1756354b3aa3170ede5610f3022c32e99849954c/README.md?%2Frakib730%2Fvit-base-oxford-iiit-pets%2Fresolve%2Fmain%2FREADME.md=&etag=%224a4db5995297382f90b9e1abba73bb180a135446%22
|
hfcreator11/rochbuchon-lora
|
hfcreator11
| 2025-07-26T04:26:02Z | 0 | 0 |
diffusers
|
[
"diffusers",
"text-to-image",
"lora",
"template:diffusion-lora",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"region:us"
] |
text-to-image
| 2025-07-26T03:40:06Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/hfcreator11/rochbuchon-lora/b70b2ff6d08317753a7894d6fcb1eb3630619d2d/README.md?%2Fhfcreator11%2Frochbuchon-lora%2Fresolve%2Fmain%2FREADME.md=&etag=%22bc9fb481c1e95fa7e2f115ec4b155ea0761e5173%22
|
Qwen/Qwen3-4B-Base
|
Qwen
| 2025-07-26T03:45:37Z | 7,528,814 | 37 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:2505.09388",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-04-28T05:04:27Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Qwen/Qwen3-4B-Base/906bfd4b4dc7f14ee4320094d8b41684abff8539/README.md?%2FQwen%2FQwen3-4B-Base%2Fresolve%2Fmain%2FREADME.md=&etag=%2283a7f3264b60636c4c59377ac2024165f33bbb8f%22
|
Thireus/Qwen3-Coder-480B-A35B-Instruct-THIREUS-Q3_K-SPECIAL_SPLIT
|
Thireus
| 2025-07-26T03:10:56Z | 0 | 0 | null |
[
"gguf",
"license:mit",
"endpoints_compatible",
"region:us",
"imatrix",
"conversational"
] | null | 2025-07-26T02:27:09Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Thireus/Qwen3-Coder-480B-A35B-Instruct-THIREUS-Q3_K-SPECIAL_SPLIT/11a3a20deacf9fc4dd943c3ebb830721181caedc/README.md?%2FThireus%2FQwen3-Coder-480B-A35B-Instruct-THIREUS-Q3_K-SPECIAL_SPLIT%2Fresolve%2Fmain%2FREADME.md=&etag=%227be5fc7f47d5db027d120b8024982df93db95b74%22
|
Nanjing-Sister-Hong-Telegram-leaks/sister.hong.twitter.video.X.Link
|
Nanjing-Sister-Hong-Telegram-leaks
| 2025-07-26T02:09:15Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-07-26T02:08:58Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Nanjing-Sister-Hong-Telegram-leaks/sister.hong.twitter.video.X.Link/df4943447f218d6792e37bd553c405f41550988f/README.md?%2FNanjing-Sister-Hong-Telegram-leaks%2Fsister.hong.twitter.video.X.Link%2Fresolve%2Fmain%2FREADME.md=&etag=%22f6967369640aa20ac0553229cc909895ab0a04f2%22
|
dgambettaphd/M_llm2_run0_gen6_S_doc1000_synt64_lr1e-04_acm_SYNLAST
|
dgambettaphd
| 2025-07-26T02:06:12Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-07-26T02:05:52Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/dgambettaphd/M_llm2_run0_gen6_S_doc1000_synt64_lr1e-04_acm_SYNLAST/999528f1e7693a96cc7f9e7628fe7e94bf739722/README.md?%2Fdgambettaphd%2FM_llm2_run0_gen6_S_doc1000_synt64_lr1e-04_acm_SYNLAST%2Fresolve%2Fmain%2FREADME.md=&etag=%22b7441640181b06c90effb718230b043e35476812%22
|
maidacundo/annie-lite-v0.2-website-planning-qwen3-8b
|
maidacundo
| 2025-07-26T01:24:16Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"text-generation-inference",
"unsloth",
"conversational",
"en",
"base_model:unsloth/Qwen3-8B-unsloth-bnb-4bit",
"base_model:finetune:unsloth/Qwen3-8B-unsloth-bnb-4bit",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-26T01:16:56Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/maidacundo/annie-lite-v0.2-website-planning-qwen3-8b/34695b97702297ce31cf917e8653ef4e34d5b21e/README.md?%2Fmaidacundo%2Fannie-lite-v0.2-website-planning-qwen3-8b%2Fresolve%2Fmain%2FREADME.md=&etag=%2275d8748319b51895659e0ac1b842f49c68648aa4%22
|
mradermacher/Edens-Fall-L3.3-70b-0.3b-i1-GGUF
|
mradermacher
| 2025-07-26T00:38:16Z | 0 | 0 | null |
[
"gguf",
"endpoints_compatible",
"region:us",
"imatrix",
"conversational"
] | null | 2025-07-25T08:02:01Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/mradermacher/Edens-Fall-L3.3-70b-0.3b-i1-GGUF/1f7be8d313d8b9d8f6ac834cbca1dc99da67f5c3/README.md?%2Fmradermacher%2FEdens-Fall-L3.3-70b-0.3b-i1-GGUF%2Fresolve%2Fmain%2FREADME.md=&etag=%225b596dbee39e2858a55aef36387b663f49ffa53b%22
|
AnthonyVezina/apocalypse-city-scene
|
AnthonyVezina
| 2025-07-26T00:16:36Z | 0 | 0 | null |
[
"license:mit",
"region:us"
] | null | 2025-07-26T00:16:36Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/AnthonyVezina/apocalypse-city-scene/41f687d66810cb447bbd1471bca07b06c857247c/README.md?%2FAnthonyVezina%2Fapocalypse-city-scene%2Fresolve%2Fmain%2FREADME.md=&etag=%227be5fc7f47d5db027d120b8024982df93db95b74%22
|
sii-research/InnoSpark-72B-0710
|
sii-research
| 2025-07-25T23:27:21Z | 35 | 2 | null |
[
"safetensors",
"qwen2",
"region:us"
] | null | 2025-07-20T07:31:08Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/sii-research/InnoSpark-72B-0710/fdd5550589e75b0a771501613ecbb057d657a190/README.md?%2Fsii-research%2FInnoSpark-72B-0710%2Fresolve%2Fmain%2FREADME.md=&etag=%2240c24b04cd2ecc7b40dc482d2bc3c662c4bdeb4a%22
|
versaceeros/66df970d-29ac-4d5c-8f42-16a3abb8fdf5
|
versaceeros
| 2025-07-25T23:12:31Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-07-25T22:33:38Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/versaceeros/66df970d-29ac-4d5c-8f42-16a3abb8fdf5/27afcaff71ee43fceb3c9ba90ba2b0dc492235ef/README.md?%2Fversaceeros%2F66df970d-29ac-4d5c-8f42-16a3abb8fdf5%2Fresolve%2Fmain%2FREADME.md=&etag=%22b7441640181b06c90effb718230b043e35476812%22
|
Lordkun/lora_model_Qwen_7b
|
Lordkun
| 2025-07-25T22:52:04Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"qwen2",
"trl",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-07-25T22:51:19Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Lordkun/lora_model_Qwen_7b/9b58d301d182eb1f0cea01b2c42d126bd2b19762/README.md?%2FLordkun%2Flora_model_Qwen_7b%2Fresolve%2Fmain%2FREADME.md=&etag=%22a0044930a750301600fdecc24c698375e97f3968%22
|
dadsdasdsa/vpd-rft
|
dadsdasdsa
| 2025-07-25T20:05:02Z | 0 | 0 | null |
[
"license:apache-2.0",
"region:us"
] | null | 2025-07-25T20:04:43Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/dadsdasdsa/vpd-rft/7a6d56806289447f8b5fe565135f96a05ac64155/README.md?%2Fdadsdasdsa%2Fvpd-rft%2Fresolve%2Fmain%2FREADME.md=&etag=%227b95401dc46245ac339fc25059d4a56d90b4cde5%22
|
VIDEOS-19-Kenya-Pastor-Daughter-viral-link/NEW.FULL.VIDEOS.Kenya.Pastor.Daughter.Viral.Video.Official.Tutorial
|
VIDEOS-19-Kenya-Pastor-Daughter-viral-link
| 2025-07-25T17:42:39Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-07-25T17:42:21Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/VIDEOS-19-Kenya-Pastor-Daughter-viral-link/NEW.FULL.VIDEOS.Kenya.Pastor.Daughter.Viral.Video.Official.Tutorial/4c8c63a5fe72a0c37814d2dcc21355af8194ec2c/README.md?%2FVIDEOS-19-Kenya-Pastor-Daughter-viral-link%2FNEW.FULL.VIDEOS.Kenya.Pastor.Daughter.Viral.Video.Official.Tutorial%2Fresolve%2Fmain%2FREADME.md=&etag=%2207dc5cfc031d072d725e73c4293cb19b5e9b5e73%22
|
najmharani/gemma-3-4b-it-finetune-224
|
najmharani
| 2025-07-25T17:17:36Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"gemma3",
"trl",
"en",
"base_model:unsloth/gemma-3-4b-it",
"base_model:finetune:unsloth/gemma-3-4b-it",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-07-25T17:17:13Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/najmharani/gemma-3-4b-it-finetune-224/bd6c066d6df56930a5faf7a4ded0af7baeb50854/README.md?%2Fnajmharani%2Fgemma-3-4b-it-finetune-224%2Fresolve%2Fmain%2FREADME.md=&etag=%22310fce4a4a3acc922441718e43aad17272468f4a%22
|
Mourad1919/DistilBert_FineTuned_ToxicData
|
Mourad1919
| 2025-07-25T17:16:47Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-07-25T17:16:44Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Mourad1919/DistilBert_FineTuned_ToxicData/d994c1b81c417ba30ce511125c5a08fd8e15fd86/README.md?%2FMourad1919%2FDistilBert_FineTuned_ToxicData%2Fresolve%2Fmain%2FREADME.md=&etag=%22bc5f30d6632ac0efdc7be2e9095e9e9579af2e33%22
|
Triangle104/EXAONE-Deep-32B-Q4_K_M-GGUF
|
Triangle104
| 2025-07-25T16:56:00Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"lg-ai",
"exaone",
"exaone-deep",
"llama-cpp",
"gguf-my-repo",
"text-generation",
"en",
"ko",
"base_model:LGAI-EXAONE/EXAONE-Deep-32B",
"base_model:finetune:LGAI-EXAONE/EXAONE-Deep-32B",
"license:other",
"endpoints_compatible",
"region:us",
"conversational"
] |
text-generation
| 2025-07-25T16:53:43Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Triangle104/EXAONE-Deep-32B-Q4_K_M-GGUF/ed18a7c23f282389d4be2b5680c50c2b22812e6a/README.md?%2FTriangle104%2FEXAONE-Deep-32B-Q4_K_M-GGUF%2Fresolve%2Fmain%2FREADME.md=&etag=%22bd4f9c447909a9a84965be5e25b7962688790109%22
|
Triangle104/Qwen3-4B-ShiningValiant3-Q6_K-GGUF
|
Triangle104
| 2025-07-25T15:38:12Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"shining-valiant",
"shining-valiant-3",
"valiant",
"valiant-labs",
"qwen",
"qwen-3",
"qwen-3-4b",
"4b",
"reasoning",
"code",
"code-reasoning",
"science",
"science-reasoning",
"physics",
"biology",
"chemistry",
"earth-science",
"astronomy",
"machine-learning",
"artificial-intelligence",
"compsci",
"computer-science",
"information-theory",
"ML-Ops",
"math",
"cuda",
"deep-learning",
"agentic",
"LLM",
"neuromorphic",
"self-improvement",
"complex-systems",
"cognition",
"linguistics",
"philosophy",
"logic",
"epistemology",
"simulation",
"game-theory",
"knowledge-management",
"creativity",
"problem-solving",
"architect",
"engineer",
"developer",
"creative",
"analytical",
"expert",
"rationality",
"conversational",
"chat",
"instruct",
"llama-cpp",
"gguf-my-repo",
"text-generation",
"en",
"dataset:sequelbox/Celestia3-DeepSeek-R1-0528",
"dataset:sequelbox/Mitakihara-DeepSeek-R1-0528",
"dataset:sequelbox/Raiden-DeepSeek-R1",
"base_model:ValiantLabs/Qwen3-4B-ShiningValiant3",
"base_model:quantized:ValiantLabs/Qwen3-4B-ShiningValiant3",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-25T15:33:18Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Triangle104/Qwen3-4B-ShiningValiant3-Q6_K-GGUF/167ff5035285dd1f859b08bfe97047b148ba8aae/README.md?%2FTriangle104%2FQwen3-4B-ShiningValiant3-Q6_K-GGUF%2Fresolve%2Fmain%2FREADME.md=&etag=%22dd0a487085213c41364e4412a3808cee9678da35%22
|
Triangle104/Qwen3-Esper3-Reasoning-CODER-Instruct-12B-Brainstorm20x-Q4_K_S-GGUF
|
Triangle104
| 2025-07-25T14:37:36Z | 0 | 0 |
transformers
|
[
"transformers",
"gguf",
"merge",
"programming",
"code generation",
"code",
"coding",
"coder",
"chat",
"brainstorm",
"qwen",
"qwen3",
"qwencoder",
"brainstorm20x",
"esper",
"esper-3",
"valiant",
"valiant-labs",
"qwen-3",
"qwen-3-8b",
"8b",
"reasoning",
"code-instruct",
"python",
"javascript",
"dev-ops",
"jenkins",
"terraform",
"scripting",
"powershell",
"azure",
"aws",
"gcp",
"cloud",
"problem-solving",
"architect",
"engineer",
"developer",
"creative",
"analytical",
"expert",
"rationality",
"conversational",
"instruct",
"llama-cpp",
"gguf-my-repo",
"text-generation",
"en",
"dataset:sequelbox/Titanium2.1-DeepSeek-R1",
"dataset:sequelbox/Tachibana2-DeepSeek-R1",
"dataset:sequelbox/Raiden-DeepSeek-R1",
"base_model:DavidAU/Qwen3-Esper3-Reasoning-CODER-Instruct-12B-Brainstorm20x",
"base_model:quantized:DavidAU/Qwen3-Esper3-Reasoning-CODER-Instruct-12B-Brainstorm20x",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-25T14:26:25Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/Triangle104/Qwen3-Esper3-Reasoning-CODER-Instruct-12B-Brainstorm20x-Q4_K_S-GGUF/3c43c627abdb93563748baa88e65e6fdfc645db4/README.md?%2FTriangle104%2FQwen3-Esper3-Reasoning-CODER-Instruct-12B-Brainstorm20x-Q4_K_S-GGUF%2Fresolve%2Fmain%2FREADME.md=&etag=%22862f8ee41e39b4e5eaba3ee2cade93f5a6fe842f%22
|
RichardErkhov/jan-hq_-_Ichigo-llama3.1-s-instruct-v0.4-step-8000-gguf
|
RichardErkhov
| 2025-07-25T14:34:51Z | 0 | 0 | null |
[
"gguf",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-07-25T13:19:33Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/RichardErkhov/jan-hq_-_Ichigo-llama3.1-s-instruct-v0.4-step-8000-gguf/27094cc42e4596063769f72f6e11387c3779d005/README.md?%2FRichardErkhov%2Fjan-hq_-_Ichigo-llama3.1-s-instruct-v0.4-step-8000-gguf%2Fresolve%2Fmain%2FREADME.md=&etag=%221f53bcd8ec842c2c98e4c43215d7cce052241171%22
|
amanfor18/Kriti
|
amanfor18
| 2025-07-25T14:07:54Z | 0 | 0 |
diffusers
|
[
"diffusers",
"text-to-image",
"lora",
"template:diffusion-lora",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:unknown",
"region:us"
] |
text-to-image
| 2025-07-25T14:07:46Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/amanfor18/Kriti/a6ffa2f38ec1f4b96ad6206c4d09dcc84553d94d/README.md?%2Famanfor18%2FKriti%2Fresolve%2Fmain%2FREADME.md=&etag=%227bb4cdaa21119d006f59340c678c479cbc59bb8a%22
|
mradermacher/C2S-Pythia-410m-diverse-single-and-multi-cell-tasks-GGUF
|
mradermacher
| 2025-07-25T12:49:09Z | 438 | 0 |
transformers
|
[
"transformers",
"gguf",
"biology",
"scRNAseq",
"en",
"base_model:vandijklab/C2S-Pythia-410m-diverse-single-and-multi-cell-tasks",
"base_model:quantized:vandijklab/C2S-Pythia-410m-diverse-single-and-multi-cell-tasks",
"license:cc0-1.0",
"endpoints_compatible",
"region:us"
] | null | 2025-03-14T16:40:40Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/mradermacher/C2S-Pythia-410m-diverse-single-and-multi-cell-tasks-GGUF/7768c3fcd8801334aee1788147788b27597ebdc8/README.md?%2Fmradermacher%2FC2S-Pythia-410m-diverse-single-and-multi-cell-tasks-GGUF%2Fresolve%2Fmain%2FREADME.md=&etag=%222424a8a9e593441396ba613690b9f415e4306ab8%22
|
aravindpai/Qwen3-8B-W4A16-G128
|
aravindpai
| 2025-07-25T12:37:39Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"compressed-tensors",
"region:us"
] |
text-generation
| 2025-07-25T12:35:30Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/aravindpai/Qwen3-8B-W4A16-G128/7986216f4529aed98e66423f28bfa8f97b1c64d6/README.md?%2Faravindpai%2FQwen3-8B-W4A16-G128%2Fresolve%2Fmain%2FREADME.md=&etag=%22bc5f30d6632ac0efdc7be2e9095e9e9579af2e33%22
|
lynn-mikami/wan-testing
|
lynn-mikami
| 2025-07-25T05:54:42Z | 0 | 0 | null |
[
"license:apache-2.0",
"region:us"
] | null | 2025-07-18T10:20:30Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/lynn-mikami/wan-testing/f7ae6ec34cf8b3ddd05f36b563055728fa9356be/README.md?%2Flynn-mikami%2Fwan-testing%2Fresolve%2Fmain%2FREADME.md=&etag=%227b95401dc46245ac339fc25059d4a56d90b4cde5%22
|
phucnd220104/Taxi
|
phucnd220104
| 2025-07-25T03:44:16Z | 0 | 0 | null |
[
"Taxi-v3",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
] |
reinforcement-learning
| 2025-07-25T03:43:11Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/phucnd220104/Taxi/6bf451b52f44a1d88d7ed4a1107c456c7edfbb5a/README.md?%2Fphucnd220104%2FTaxi%2Fresolve%2Fmain%2FREADME.md=&etag=%221f497ad8c5befe1ffb9fae0bcd619c0bca2f6d36%22
|
dgambettaphd/M_llm2_run0_gen7_W_doc1000_synt64_lr1e-04_acm_SYNLAST
|
dgambettaphd
| 2025-07-25T00:47:31Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-07-25T00:47:20Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/dgambettaphd/M_llm2_run0_gen7_W_doc1000_synt64_lr1e-04_acm_SYNLAST/09d4da1cf30b5e71d6bc9500dcec4069943272b8/README.md?%2Fdgambettaphd%2FM_llm2_run0_gen7_W_doc1000_synt64_lr1e-04_acm_SYNLAST%2Fresolve%2Fmain%2FREADME.md=&etag=%22b7441640181b06c90effb718230b043e35476812%22
|
neural-interactive-proofs/finetune_dpo_qwen2_5-1_5b-instruct_cv_qwen2.5-1.5B_verifier_nip_lr_4o_mini_1_1_iter_4_verifier_1
|
neural-interactive-proofs
| 2025-07-25T00:14:11Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"generated_from_trainer",
"trl",
"dpo",
"arxiv:2305.18290",
"base_model:Qwen/Qwen2.5-1.5B-Instruct",
"base_model:finetune:Qwen/Qwen2.5-1.5B-Instruct",
"endpoints_compatible",
"region:us"
] | null | 2025-07-25T00:14:01Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/neural-interactive-proofs/finetune_dpo_qwen2_5-1_5b-instruct_cv_qwen2.5-1.5B_verifier_nip_lr_4o_mini_1_1_iter_4_verifier_1/e3bc509c47b7e9b7db8e89f29633a572c755ea3a/README.md?%2Fneural-interactive-proofs%2Ffinetune_dpo_qwen2_5-1_5b-instruct_cv_qwen2.5-1.5B_verifier_nip_lr_4o_mini_1_1_iter_4_verifier_1%2Fresolve%2Fmain%2FREADME.md=&etag=%22f89f7fcac4549d9077fa930b298f95f3f3ae841c%22
|
zzrtpsh/Zz
|
zzrtpsh
| 2025-07-24T20:57:43Z | 0 | 0 | null |
[
"license:bigscience-openrail-m",
"region:us"
] | null | 2025-07-24T20:57:43Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/zzrtpsh/Zz/80f563b3f106255c22b09a1bb4a9c31c4449a83f/README.md?%2Fzzrtpsh%2FZz%2Fresolve%2Fmain%2FREADME.md=&etag=%2238827d1feedc893c118e9dcf83f860e8a95be77f%22
|
ArtusDev/ReadyArt_MS3.2-The-Omega-Directive-24B-Unslop-v2.0-W8A8
|
ArtusDev
| 2025-07-24T20:36:55Z | 0 | 0 | null |
[
"safetensors",
"mistral",
"nsfw",
"explicit",
"roleplay",
"unaligned",
"ERP",
"Erotic",
"Horror",
"Violence",
"w8a8",
"text-generation",
"conversational",
"en",
"base_model:ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0",
"base_model:quantized:ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0",
"license:apache-2.0",
"8-bit",
"compressed-tensors",
"region:us"
] |
text-generation
| 2025-07-24T20:30:49Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/ArtusDev/ReadyArt_MS3.2-The-Omega-Directive-24B-Unslop-v2.0-W8A8/ef766f06261580dafd336c16f02a02816738f40d/README.md?%2FArtusDev%2FReadyArt_MS3.2-The-Omega-Directive-24B-Unslop-v2.0-W8A8%2Fresolve%2Fmain%2FREADME.md=&etag=%22ad1c329043cbfb6d32cd1666aabab8abb7e6e289%22
|
dennohpeter/wav2vec2-large-xlsr-53-sw-tokenizer
|
dennohpeter
| 2025-07-24T20:08:13Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"dataset:common_voice_17_0",
"base_model:facebook/wav2vec2-large-xlsr-53",
"base_model:finetune:facebook/wav2vec2-large-xlsr-53",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2025-07-19T01:53:21Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/dennohpeter/wav2vec2-large-xlsr-53-sw-tokenizer/ddb5662605e49a80dde583eccf1499e13f58526b/README.md?%2Fdennohpeter%2Fwav2vec2-large-xlsr-53-sw-tokenizer%2Fresolve%2Fmain%2FREADME.md=&etag=%225deac5ed87f64b090cf73f046ed02c261a808c4b%22
|
gpol13/codet5_qlora
|
gpol13
| 2025-07-24T20:07:12Z | 2 | 0 |
peft
|
[
"peft",
"safetensors",
"arxiv:1910.09700",
"base_model:Salesforce/codet5-base",
"base_model:adapter:Salesforce/codet5-base",
"region:us"
] | null | 2025-05-20T18:02:01Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/gpol13/codet5_qlora/70116e406647d1f3c5b35426c0ea9adb810531fc/README.md?%2Fgpol13%2Fcodet5_qlora%2Fresolve%2Fmain%2FREADME.md=&etag=%224aa88c044abef44339b87722094d46730a06bf8d%22
|
AINovice2005/Voxtral-Mini-3B-2507-smashed
|
AINovice2005
| 2025-07-24T14:49:52Z | 0 | 0 |
transformers
|
[
"transformers",
"pruna-ai",
"int8",
"quantized",
"voxtral",
"audio-text-to-text",
"base_model:mistralai/Voxtral-Mini-3B-2507",
"base_model:finetune:mistralai/Voxtral-Mini-3B-2507",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
audio-text-to-text
| 2025-07-24T10:08:02Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/AINovice2005/Voxtral-Mini-3B-2507-smashed/fb27446519c52ccd7f32a54b871065e452fe95d3/README.md?%2FAINovice2005%2FVoxtral-Mini-3B-2507-smashed%2Fresolve%2Fmain%2FREADME.md=&etag=%2240f0e1eec5db7a1e4310ee6fad4763d80f850619%22
|
tamewild/4b_v32_merged_e8
|
tamewild
| 2025-07-24T12:35:34Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-24T12:33:39Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/tamewild/4b_v32_merged_e8/065c79310a276a5dfb07eeccd92b1613f99ff728/README.md?%2Ftamewild%2F4b_v32_merged_e8%2Fresolve%2Fmain%2FREADME.md=&etag=%22bc5f30d6632ac0efdc7be2e9095e9e9579af2e33%22
|
french-datasets/lielbin_BabyBERTa-wikipedia1_2.5-with-Masking_run2-finetuned-SQuAD
|
french-datasets
| 2025-07-24T10:25:42Z | 0 | 0 | null |
[
"fra",
"region:us"
] | null | 2025-07-24T10:25:41Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/french-datasets/lielbin_BabyBERTa-wikipedia1_2.5-with-Masking_run2-finetuned-SQuAD/356a41bac6df59532aa2062a0d2ec1ec67d282b8/README.md?%2Ffrench-datasets%2Flielbin_BabyBERTa-wikipedia1_2.5-with-Masking_run2-finetuned-SQuAD%2Fresolve%2Fmain%2FREADME.md=&etag=%224e0c1eca15ebe0f000b24280a28657c7941cac0c%22
|
french-datasets/tatoy_llama2-qlora-finetunined-french
|
french-datasets
| 2025-07-24T09:24:31Z | 0 | 0 | null |
[
"fra",
"region:us"
] | null | 2025-07-24T09:24:31Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/french-datasets/tatoy_llama2-qlora-finetunined-french/43728909b3dcf65b749fd21fb7089cfe5e3f9d23/README.md?%2Ffrench-datasets%2Ftatoy_llama2-qlora-finetunined-french%2Fresolve%2Fmain%2FREADME.md=&etag=%223432d9cb1f7323cb3c469b95890da428e635aeb7%22
|
RoyArkh/Test_wiki-EleutherAI-pythia-160m_client7_round0
|
RoyArkh
| 2025-07-24T09:04:09Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"gpt_neox",
"text-generation",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-24T09:01:41Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/RoyArkh/Test_wiki-EleutherAI-pythia-160m_client7_round0/7e495a9317da7bde7ac8bbb6f60e98eac80842c8/README.md?%2FRoyArkh%2FTest_wiki-EleutherAI-pythia-160m_client7_round0%2Fresolve%2Fmain%2FREADME.md=&etag=%22bc5f30d6632ac0efdc7be2e9095e9e9579af2e33%22
|
french-datasets/EstherYang_llama2-qlora-finetunined-french
|
french-datasets
| 2025-07-24T08:56:08Z | 0 | 0 | null |
[
"fra",
"region:us"
] | null | 2025-07-24T08:56:07Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/french-datasets/EstherYang_llama2-qlora-finetunined-french/ca47004c4864eeb884b89d1e575be6eb88e020f4/README.md?%2Ffrench-datasets%2FEstherYang_llama2-qlora-finetunined-french%2Fresolve%2Fmain%2FREADME.md=&etag=%22a8d7ac77811f2595569722903d252b0c2bce15f4%22
|
french-datasets/mathonvictor_full-ARIA-7B-V3-mistral-french
|
french-datasets
| 2025-07-24T08:25:24Z | 0 | 0 | null |
[
"fra",
"region:us"
] | null | 2025-07-24T08:25:23Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/french-datasets/mathonvictor_full-ARIA-7B-V3-mistral-french/69e6e0878a1b4c2042306ea2935a26ebb90fb02a/README.md?%2Ffrench-datasets%2Fmathonvictor_full-ARIA-7B-V3-mistral-french%2Fresolve%2Fmain%2FREADME.md=&etag=%2234bbdee845ddcc4c2c4de149d08cef2556362441%22
|
bboeun/ko-conversation-LoRA-gpu-model2
|
bboeun
| 2025-07-24T06:04:47Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-07-24T06:04:33Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/bboeun/ko-conversation-LoRA-gpu-model2/2437babe6005834eedb4b92b488d2dbf3b219521/README.md?%2Fbboeun%2Fko-conversation-LoRA-gpu-model2%2Fresolve%2Fmain%2FREADME.md=&etag=%22bc5f30d6632ac0efdc7be2e9095e9e9579af2e33%22
|
brunoyun/Llama-3.1-Amelia-AR-8B-v1
|
brunoyun
| 2025-07-23T13:33:13Z | 4 | 0 | null |
[
"safetensors",
"llama",
"argumentation",
"argument-mining",
"text-generation",
"conversational",
"en",
"base_model:meta-llama/Llama-3.1-8B-Instruct",
"base_model:finetune:meta-llama/Llama-3.1-8B-Instruct",
"license:llama3.1",
"region:us"
] |
text-generation
| 2025-06-17T11:51:07Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/brunoyun/Llama-3.1-Amelia-AR-8B-v1/856dff96148f0c4edd5ec0e047d155f5bb549fe5/README.md?%2Fbrunoyun%2FLlama-3.1-Amelia-AR-8B-v1%2Fresolve%2Fmain%2FREADME.md=&etag=%2204f88805c53c4b7629368bfe9393ae20e4a1c8b2%22
|
LeonGuertler/ScalingLaws-Qwen3-4B-Seven-Env-1-step_000475
|
LeonGuertler
| 2025-07-23T06:53:28Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-23T06:47:26Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/LeonGuertler/ScalingLaws-Qwen3-4B-Seven-Env-1-step_000475/89457d13558db8a0e5d85d152ca62d2788e8ef10/README.md?%2FLeonGuertler%2FScalingLaws-Qwen3-4B-Seven-Env-1-step_000475%2Fresolve%2Fmain%2FREADME.md=&etag=%22bc5f30d6632ac0efdc7be2e9095e9e9579af2e33%22
|
ChrisToukmaji/laft_tha_mpt
|
ChrisToukmaji
| 2025-07-23T00:51:39Z | 13 | 0 | null |
[
"safetensors",
"mpt",
"generated_from_trainer",
"custom_code",
"dataset:mc4",
"arxiv:2506.19187",
"base_model:mosaicml/mpt-7b",
"base_model:finetune:mosaicml/mpt-7b",
"license:apache-2.0",
"region:us"
] | null | 2025-07-19T05:02:12Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/ChrisToukmaji/laft_tha_mpt/006bbe3ea234e430896e3fd361227a9aa45dc470/README.md?%2FChrisToukmaji%2Flaft_tha_mpt%2Fresolve%2Fmain%2FREADME.md=&etag=%22d0e181cf17d795aef49ce21bf4321de248f1812d%22
|
chinna6/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-skilled_hoarse_anaconda
|
chinna6
| 2025-07-22T18:53:11Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2",
"text-generation",
"rl-swarm",
"genrl-swarm",
"grpo",
"gensyn",
"I am skilled_hoarse_anaconda",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-07-22T18:05:27Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/chinna6/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-skilled_hoarse_anaconda/fd719fd5337e4d96731ad4ca3fb321b6f1d70262/README.md?%2Fchinna6%2FQwen2.5-0.5B-Instruct-Gensyn-Swarm-skilled_hoarse_anaconda%2Fresolve%2Fmain%2FREADME.md=&etag=%2278819fcd811abc0c2339e2b78060a11a44953eb2%22
|
ymoslem/wmt25-cs-de-20layers-2e-05-100k-news-sentences
|
ymoslem
| 2025-07-21T21:32:19Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"cohere",
"text-generation",
"translation",
"generated_from_trainer",
"sft",
"trl",
"cs",
"de",
"base_model:CohereLabs/aya-expanse-8b",
"base_model:finetune:CohereLabs/aya-expanse-8b",
"license:cc-by-nc-4.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
translation
| 2025-07-21T14:29:50Z |
Temporary Redirect. Redirecting to /api/resolve-cache/models/ymoslem/wmt25-cs-de-20layers-2e-05-100k-news-sentences/61d12fc100ef0ce8010d29717be21770ef46d856/README.md?%2Fymoslem%2Fwmt25-cs-de-20layers-2e-05-100k-news-sentences%2Fresolve%2Fmain%2FREADME.md=&etag=%222152398920655a9526580c242f5e64d6f1a75413%22
|
Synthyra/DSM_ppi_full
|
Synthyra
| 2025-06-23T15:22:52Z | 202 | 0 |
transformers
|
[
"transformers",
"pytorch",
"dsm",
"custom_code",
"arxiv:2506.08293",
"endpoints_compatible",
"region:us"
] | null | 2025-06-11T15:02:19Z |
---
library_name: transformers
tags: []
---
# DSM: Diffusion Models for Protein Sequence Generation
### Note: This readme is shared between our GitHub and Huggingface pages.
## Table of Contents
- [Introduction](#introduction)
- [Models](#models)
- [Usage](#usage)
- [Demos](#usage)
- [Local installation](#installation)
- [Training](#training)
- [Evaluation](#evaluation)
- [Results](#results)
- [Cite](#cite)
## Introduction
DSM (Diffusion Sequence Model) is a novel Protein Language Model (pLM) developed in collaboration between the [Gleghorn Lab](https://www.gleghornlab.com/) and [Synthyra](https://synthyra.com/). It was trained with masked diffusion to enable both high-quality representation learning and generative protein design. This repository contains the code for training, evaluating, and applying DSM and its variants.
DSM is capable of generating diverse, biomimetic sequences that align with expected amino acid compositions, secondary structures, and predicted functions. Furthermore, DSM's learned representations match or exceed those of comparably sized pLMs on various downstream tasks. DSM is detailed extensively in our [preprint](https://arxiv.org/abs/2506.08293) (which is currently in review). Beyond the base and PPI variants, we are currently training versions to jointly diffuse over sequence and foldseek tokens, as well as [Annotation Vocabulary](https://www.biorxiv.org/content/10.1101/2024.07.30.605924v1) tokens. Since the preprint release, Synthyra has trained [Synthyra/DSM_ppi_full](https://huggingface.co/Synthyra/DSM_ppi_full) which neglects the LoRA PPI training in favor for full finetuning. Additionally, the sequences SeqA and SeqB are jointly masked instead of just SeqB in the original version. We plan on adding the **many** new results to the second version of the preprint and eventual journal article.
## Models
Relevant Huggingface hosted models and datasets
- **Base DSM Models**:
- [GleghornLab/DSM_150](https://huggingface.co/GleghornLab/DSM_150) - 150M parameter DSM model
- [GleghornLab/DSM_650](https://huggingface.co/GleghornLab/DSM_650) - 650M parameter DSM model
- **DSM-ppi Models**:
(LoRA versions - results reported in paper but not recommended for real use)
- [GleghornLab/DSM_150_ppi_lora](https://huggingface.co/GleghornLab/DSM_150_ppi_lora) - 150M parameter LoRA DSM-ppi model
- [GleghornLab/DSM_650_ppi_lora](https://huggingface.co/GleghornLab/DSM_650_ppi_lora) - 650M parameter LoRA DSM-ppi model
- [GleghornLab/DSM_150_ppi_control](https://huggingface.co/GleghornLab/DSM_150_ppi_control) - Control version of LoRA DSM-ppi
(Fully finetuned - recommended for real use)
- [Synthyra/DSM_ppi_full](https://huggingface.co/Synthyra/DSM_ppi_full) - 650M parameter DSM-ppi model
- **Datasets**:
- [Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50) - Open MetaGenomic dataset clustered at 50% identity (207M sequences)
- [GleghornLab/stringv12_modelorgs_9090](https://huggingface.co/datasets/GleghornLab/stringv12_modelorgs_9090) - STRING database model organisms (653k sequences)
- **Utility Models**:
- [GleghornLab/production_ss4_model](https://huggingface.co/GleghornLab/production_ss4_model) - Secondary structure prediction (4-class)
- [GleghornLab/production_ss9_model](https://huggingface.co/GleghornLab/production_ss9_model) - Secondary structure prediction (9-class)
## Usage
This section outlines how to use a trained `DSM` model for common generation tasks. The core generation logic is provided by the `GenerateMixin` class, used by `DSM` models.
First, ensure you have a trained model (either one you trained or a pre-trained one from Hugging Face Hub) and the necessary environment set up.
```python
import torch
from models.modeling_dsm import DSM # Or DSM_ppi for binder generation
# Load a pre-trained model
model_name_or_path = "GleghornLab/DSM_650" # Replace with your model of choice
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = DSM.from_pretrained(model_name_or_path).to(device).eval()
tokenizer = model.tokenizer
```
```console
You are using a model of type esm_diff to instantiate a model of type dsm. This is not supported for all configurations of models and can yield errors.
```
This warning is normal - all good!
### 1. Unconditional Sequence Generation
To generate a novel sequence of a specific length. DSM uses a progressive denoising approach.
```python
### Unconditional generation
length = 100
mask_token = tokenizer.mask_token
# optionally, enforce starting with methionine
input_tokens = tokenizer.encode('M' + ''.join([mask_token] * (length - 1)), add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
generated_sequences = model.decode_output(output)
print(f"Generated sequence: {generated_sequences[0]}")
```
```console
Generated sequence: MFRVDALQVAQQETLAIGRSTAYDKQESPSMAQRQVLTQLAAYGGENDLRQICIPAERRNFLSIANGASYQFVEEDNEANGGYWSPHKAGLPESACKRFI
```
### 2. Mask Filling (Inpainting)
To fill in masked regions of a template sequence:
```python
# Mask Filling / Inpainting
template_sequence = "MA<mask><mask><mask>KEG<mask><mask>STL"
input_tokens = tokenizer.encode(template_sequence, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
generated_sequences = model.decode_output(output)
print(f"Generated sequence: {generated_sequences[0]}")
```
```console
Generated sequence: MAVKFKEGGISTL
```
### 3. Conditional Generation (e.g., Binders - using DSM-ppi)
```python
# from models.modeling_dsm import DSM_ppi
# model_binder = DSM_ppi.from_pretrained("GleghornLab/DSM_650_ppi_lora").to(device).eval()
# The lora version from the paper leads to unreliable outputs
# Synthyra has generously trained a version through full fine tuning
model = DSM.from_pretrained("Synthyra/DSM_ppi_full").to(device).eval()
# BBF-14
target_seq = "MGTPLWALLGGPWRGTATYEDGTKVTLDYRYTRVSPDRLRADVTYTTPDGTTLEATVDLWKDANGVIRYHATYPDGTSADGTLTQLDADTLLATGTYDDGTKYTVTLTRVAPGSGWHHHHHH"
# For binder generation, the 'interactor' (SeqB) part is what gets generated/filled.
# Start with a fully masked interactor of desired length.
interactor_template_len = 256
interactor_template = ''.join([mask_token] * interactor_template_len)
combined_input_str = target_seq + '<eos>' + interactor_template
input_tokens = tokenizer.encode(combined_input_str, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
target, binder = model.decode_dual_input(output, seperator='<eos>')
# Parse out the generated interactor part based on EOS tokens.
# Example: generated_full_seq_str.split(model_binder.tokenizer.eos_token)[1]
print(f"Generated binder {binder[0]}")
```
```console
Generated binder HRHHHRRPTHARETEWLARMRLGIAEHQRIAVPRSDLEPDQMRERAADNQRLVKEYDQVIDHQTEGSTERLFEVLRVWEQVNTEQAHHEASAALEFGRVGYPDDEGGRAFYTQANAHKKDLVEYIGGIDEDAKWDPRIAWLMPEGGQPVKATVIGVSEERINGLKVLDDHWGRERRLWLINLFTALQAYDDPTRPTQVTLTPATDQLTNDVQYLLLSTRYTPPGVTTAVKIRKLDGRTLKVLTTEAPYVVRGATLS
```
Folded with Chai1:
`Synthyra/DSM_ppi_full` was actually trained to fill masks from any part of SeqA and SeqB. That means you can fully hallucinate plausibly interacting protein pairs.
```python
seq_a_length = 128
seq_b_length = 128
seq_a_template = ''.join([mask_token] * seq_a_length)
seq_b_template = ''.join([mask_token] * seq_b_length)
combined_input_str = seq_a_template + '<eos>' + seq_b_template
input_tokens = tokenizer.encode(combined_input_str, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=10, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
seqa, seqb = model.decode_dual_input(output, seperator='<eos>')
# Parse out the generated interactor part based on EOS tokens.
# Example: generated_full_seq_str.split(model_binder.tokenizer.eos_token)[1]
print(f"SeqA: {seqa[0][5:]}") # remove cls token
print(f"SeqB: {seqb[0]}")
```
```console
SeqA: MVNLAKMRQRTEQNLREVSSFVKILFHTVLKFPMKINIGIHVHINMQAAQNAAADQNMQATNVIDLHNFKMGKDIGVDNKASATAHIYDEAHHTFLQLGAIKLLHAIPMIAGPVRCRLPIGFGHRFRG
SeqB: HYKNPMHSLLDSNVLHKDVVEVRLPIKIGMELDVMASAMREFLMPGTQQGDLRVIAEKRPVNKLHTYRRDLVKLLLAGAKLGTEAKSVELDLYRTELGGLVVYIININIATWDIIFAKVKICRGNDKP
```
Folded with Chai1:
## Demos
There are various demos with many more to come. For example, in `demo_dsm_ppi_full.py` (run by `python -m demos.demo_dsm_ppi_full`) we perform a test on DSM-ppi.
We take 1000 protein pairs from BIOGRID (real protein-protein interactions) and 1000 from Negatome (non interacting protein pairs) and mask the second sequence (SeqB) by 50%.
This acts as a sanity check, as we expect the accuracy on reconstructing real positive PPIs to be higher than the accuracy on non-interacting proteins.
Indeed, this is the case:
```console
==================================================
RESULTS COMPARISON
==================================================
Positive examples:
Mean accuracy: 0.495 ± 0.322
Processed: 1000 examples
Negative examples:
Mean accuracy: 0.227 ± 0.231
Processed: 1000 examples
Difference (Positive - Negative): 0.267
T-test: t=21.331, p=0.000
Difference is statistically significant (p < 0.05)
```
## Installation
1. **Clone the repository:**
```bash
git clone <repository-url>
cd <repository-name>
```
2. **Initialize the submodules:**
```bash
git submodule update --init --remote --recursive
```
3. **Set up the Python virtual environment:**
The `setup_bioenv.sh` script creates a virtual environment named `bioenv` in your home directory (`~/bioenv`), installs PyTorch with CUDA 12.6 support, and then installs all other dependencies from `requirements.txt`.
Make the script executable:
```bash
chmod +x setup_bioenv.sh
```
Run the script:
```bash
./setup_bioenv.sh
```
If you are not on a linux machine, you can install the requirements directly
```console
python -m pip install -r requirements.txt
```
4. **Activate the environment:**
Each time you want to work on this project, activate the virtual environment:
```bash
source ~/bioenv/bin/activate
```
5. **To deactivate the environment:**
```bash
deactivate
```
## Training
The primary script for training models is `training/train_dsm.py`. This script further pretrains an ESM2 checkpoint using the DSM objective (masked diffusion based on LLaDA) on a large protein sequence dataset like [OMG-prot50](https://huggingface.co/datasets/Synthyra/omg_prot50).
### Main Training Script: `train_dsm.py`
- **Base Model**: DSM models are extended from pre-trained ESM2 checkpoints (e.g., ESM2-150M, ESM2-650M).
- **Training Objective**: Masked diffusion loss, where the model predicts masked tokens. The loss is scaled by `1/(t + epsilon)` where `t` is the corruption level, penalizing errors more at low mask rates.
- **Language Modeling Head**: Uses a modified head with a soft-logit cap (`tau=30`) and tied output projection weights to the token embeddings.
- **Data Handling**:
- Training data can be streamed from datasets like [Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50) (a version of Open MetaGenomic dataset clustered at 50% identity).
- Uses `data.dataset_classes.SequenceDatasetFromList` for validation/test sets and `data.dataset_classes.IterableDatasetFromHF` for streaming training.
- `data.data_collators.SequenceCollator` is used for batching.
- **Training Process**:
- Utilizes Hugging Face `TrainingArguments`.
- A custom `IterableTrainer` (from `training.iterable_trainer.py`) handles iterable datasets.
- Uses AdamW optimizer and a cosine learning rate scheduler with linear warmup.
- Supports logging to Weights & Biases (wandb).
- The trained model can be pushed to Hugging Face Hub.
- Example checkpoints mentioned in the paper: [DSM-150](https://huggingface.co/GleghornLab/DSM_150) (from ESM2-150M, 100k steps, batch 32, seqlen 512, LR 1e-4) and [DSM-650](https://huggingface.co/GleghornLab/DSM_650) (from ESM2-650M, 100k steps, global batch 128, seqlen 2048, LR 1e-4).
**Usage Example:**
```bash
python -m training.train_dsm \
--model_path facebook/esm2_t33_650M_UR50D \
--save_path GleghornLab/DSM_650 \
--lr 1e-4 \
--batch_size 8 \
--grad_accum 16 \
--max_steps 100000 \
--save_every 1000 \
--fp16 \
--wandb_project "DSM_Training" \
--token <your_hf_token_if_needed_for_private_repo_or_saving>
```
**Key Command-Line Arguments for `train_dsm.py`:**
* `--token`: Hugging Face token.
* `--model_path`: Path to the base ESM2 model to start from.
* `--save_path`: Path to save the trained DSM model on Hugging Face Hub.
* `--lr`: Learning rate.
* `--batch_size`: Batch size per device.
* `--grad_accum`: Gradient accumulation steps.
* `--max_steps`: Maximum training steps.
* `--wandb_project`: Wandb project name (default: `DSM`).
* `--max_length`: Maximum sequence length.
* `--save_every`: Save model and evaluate every N steps.
* `--fp16`: Enable mixed-precision training.
* `--bugfix`: Use small batch size and max length for debugging.
### Other Training Scripts (e.g., for DSM-ppi)
The `training/` directory may also contain scripts like `train_dsm_bind.py`.
- DSM-ppi (e.g., [DSM-150-ppi](https://huggingface.co/GleghornLab/DSM_150_ppi_lora), [DSM-650-ppi](https://huggingface.co/GleghornLab/DSM_650_ppi_lora)) is fine-tuned on PPI datasets.
- Training involves conditioning on a target sequence (SeqA) to generate an interactor (SeqB) using the format `[CLS]--SeqA--[EOS]--[MASKED~SeqB]--[EOS]`.
- LoRA (Low-Rank Adaptation) can be applied to attention layers for efficient fine-tuning.
And `training/iterable_trainer.py` provides the `get_iterable_trainer` function used by `train_dsm.py` to enable training with iterable datasets.
## Evaluation
The repository includes a comprehensive suite for evaluating model performance, focusing on:
1. **Sequence Reconstruction (Mask Filling):**
* Evaluated by masking validation/test sets at various corruption rates (5% to 90%) and measuring cross-entropy loss, weighted F1 score, and Alignment Score (ASc) for the masked positions.
* The script `evaluation/mask_filling.py` is central to this.
2. **Unconditional Generation Quality:**
* Generate a corpus of sequences based on lengths from a reference set (e.g., validation data).
* Compare distributions (1-mers, 2-mers, 3-mers) of amino acids and predicted secondary structures between generated and natural sequences using χ² test and Jensen-Shannon (JS) divergence.
* Compare distributions of predicted functional annotations (e.g., using Annotation Vocabulary - AV terms).
* Scripts involved: `evaluation/unconditional_generation_tuning.py` (to find optimal generation parameters like temperature and step divisor `s`), `evaluation/unconditional_generation.py`, `evaluation/ss_pred.py` (using [production_ss4_model](https://huggingface.co/GleghornLab/production_ss4_model) or [production_ss9_model](https://huggingface.co/GleghornLab/production_ss9_model)), `evaluation/annotate_comparisons.py`, `evaluation/compare_distributions.py`, `evaluation/plot_distribution_comparisons.py`.
* The `run_eval_pipeline.py` script automates this workflow.
3. **Representation Quality (Model Probing):**
* Evaluate learned embeddings by training linear probes (or simple transformer blocks) on various downstream tasks (e.g., secondary structure prediction, localization prediction, etc.).
* Performance is compared against random vectors, randomized transformers, and other established pLMs.
* The assessment was done with [Protify](https://github.com/Synthyra/Protify), an open-source framework that can be used for pLM training and evaluation.
4. **Conditional Generation (Binder Design for DSM-ppi):**
* Evaluate DSM-ppi on benchmarks like BenchBB.
* Generate binders for target proteins using template-based masking strategies.
* Assess generated binders using *in-silico* tools like Synteract2 for predicted binding affinity (ppKd).
The `evaluation/` directory also contains a `readme.md` which provides further details on some evaluation workflows. Key metrics used include:
- **Alignment Score (ASc):** A normalized Needleman-Wunsch global alignment score (using BLOSUM62) to measure sequence similarity, robust to length variations. ASc(a, b) = l/(f(a, a) - f(a, b) + l).
- **Jensen-Shannon (JS) Divergence:** To compare distributions of k-mers and functional terms.
**Running the Full Unconditional Evaluation Pipeline:**
```bash
python run_eval_pipeline.py --token YOUR_HF_TOKEN --data_dir ./evaluation_results
```
Refer to `run_eval_pipeline.py --help` for more options, such as `--skip_tuning`.
### Mask Filling Evaluation
The script `evaluation/mask_filling.py` is used to evaluate models on their ability to predict masked tokens in a sequence across various masking rates.
- **Functionality:**
- Evaluates different models (DSM, DPLM, standard ESM models).
- Tests across multiple datasets ([Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50), [GleghornLab/stringv12_modelorgs_9090](https://huggingface.co/datasets/GleghornLab/stringv12_modelorgs_9090)).
- Calculates metrics: loss, perplexity, precision, recall, F1, accuracy, MCC, and alignment score.
- Saves detailed results to CSV files.
- Can generate a summary plot comparing model performance across different mask rates using `evaluation/plot_mask_fill_results.py`.
- **Usage Example:**
```bash
python -m evaluation.mask_filling \
--token YOUR_HF_TOKEN \
--batch_size 4 \
--mask_rates 0.15 0.30 0.50 \
--data_splits valid test \
--results_dir ./results/mask_fill_custom
```
To generate a comparison plot from existing results:
```bash
python -m evaluation.mask_filling --generate_comparison_plot --results_dir ./results/mask_fill_custom --plot_output ./results/mask_fill_custom/comparison.png
```
### Other Evaluation Scripts
The `evaluation/` directory contains additional scripts for more specific analyses. These are typically run independently:
- `evaluation/all_targets_uncond.py` and `evaluation/all_targets_cond.py`: Likely for evaluating generation towards specific targets, unconditionally and conditionally.
- `evaluation/conditional_binder.py` and `evaluation/unconditional_binder.py`: Suggest evaluation focused on generating protein binders.
- `evaluation/unconditional_by_length.py`: May evaluate unconditional generation focusing on sequence length distributions.
- `evaluation/utils.py`: Utility functions for evaluation scripts.
Users should refer to individual scripts (e.g., using `python -m evaluation.<script_name> --help`) for their specific usage and arguments.
The `evaluation/` directory also contains a `readme.md` which provides further details on the unconditional generation evaluation workflow.
## Results
DSM demonstrates strong performance in both protein sequence generation and representation learning, establishing masked diffusion as a powerful paradigm.
- **Biomimetic Sequence Generation**: Unconditionally generated DSM sequences closely mimic natural protein distributions in terms of amino acid k-mers, predicted secondary structures (JS divergence < 0.01 for AA k-mers), and predicted functional annotations (AV terms, JS divergence ~0.1). This suggests DSM captures underlying biological principles.
- **Superior Sequence Reconstruction**: DSM models significantly outperform MLM-based ESM2 models in reconstructing sequences from highly corrupted inputs (up to 90% masking).
- At 90% masking, DSM achieves an Alignment Score (ASc) of ~0.27, considerably higher than random.
- DSM models show higher F1 scores in reconstruction tasks compared to DPLM models, especially at high mask rates.
- **High-Quality Embeddings**: DSM embeddings match or exceed the quality of those from comparably sized pLMs (ESM2, DPLM) and even larger autoregressive models (ProtCLM 1B) on various downstream tasks evaluated by linear probing. [DSM-650](https://huggingface.co/GleghornLab/DSM_650) generally provides the best representations among tested models of similar size.
- **Effective Binder Design (DSM-ppi):**
- DSM-ppi fine-tuned on protein-protein interaction data, demonstrates the ability to generate protein binders conditioned on target sequences.
- On the BenchBB benchmark, DSM-generated binders (both unconditional DSM and conditional DSM-ppi) show promising predicted binding affinities, in some cases superior to known binders. For example, designs for EGFR showed high predicted pKd and good structural metrics (ipTM, pTM with AlphaFold3).
- **Efficiency**: DSM can generate realistic protein sequences from a single forward pass during reconstruction tasks at high mask rates, offering potential efficiency advantages over iterative AR or some discrete diffusion models.
These results highlight DSM's capability to unify high-quality protein representation learning and biologically coherent generative modeling within a single framework.
## Cite
```
@misc{hallee2025diffusionsequencemodelsenhanced,
title={Diffusion Sequence Models for Enhanced Protein Representation and Generation},
author={Logan Hallee and Nikolaos Rafailidis and David B. Bichara and Jason P. Gleghorn},
year={2025},
eprint={2506.08293},
archivePrefix={arXiv},
primaryClass={q-bio.BM},
url={https://arxiv.org/abs/2506.08293},
}
```
|
GleghornLab/DSM_650_ppi_lora
|
GleghornLab
| 2025-06-23T15:22:51Z | 40 | 0 |
transformers
|
[
"transformers",
"safetensors",
"esm_diff",
"custom_code",
"arxiv:2506.08293",
"endpoints_compatible",
"region:us"
] | null | 2025-05-08T19:10:47Z |
---
library_name: transformers
tags: []
---
# DSM: Diffusion Models for Protein Sequence Generation
### Note: This readme is shared between our GitHub and Huggingface pages.
## Table of Contents
- [Introduction](#introduction)
- [Models](#models)
- [Usage](#usage)
- [Demos](#usage)
- [Local installation](#installation)
- [Training](#training)
- [Evaluation](#evaluation)
- [Results](#results)
- [Cite](#cite)
## Introduction
DSM (Diffusion Sequence Model) is a novel Protein Language Model (pLM) developed in collaboration between the [Gleghorn Lab](https://www.gleghornlab.com/) and [Synthyra](https://synthyra.com/). It was trained with masked diffusion to enable both high-quality representation learning and generative protein design. This repository contains the code for training, evaluating, and applying DSM and its variants.
DSM is capable of generating diverse, biomimetic sequences that align with expected amino acid compositions, secondary structures, and predicted functions. Furthermore, DSM's learned representations match or exceed those of comparably sized pLMs on various downstream tasks. DSM is detailed extensively in our [preprint](https://arxiv.org/abs/2506.08293) (which is currently in review). Beyond the base and PPI variants, we are currently training versions to jointly diffuse over sequence and foldseek tokens, as well as [Annotation Vocabulary](https://www.biorxiv.org/content/10.1101/2024.07.30.605924v1) tokens. Since the preprint release, Synthyra has trained [Synthyra/DSM_ppi_full](https://huggingface.co/Synthyra/DSM_ppi_full) which neglects the LoRA PPI training in favor for full finetuning. Additionally, the sequences SeqA and SeqB are jointly masked instead of just SeqB in the original version. We plan on adding the **many** new results to the second version of the preprint and eventual journal article.
## Models
Relevant Huggingface hosted models and datasets
- **Base DSM Models**:
- [GleghornLab/DSM_150](https://huggingface.co/GleghornLab/DSM_150) - 150M parameter DSM model
- [GleghornLab/DSM_650](https://huggingface.co/GleghornLab/DSM_650) - 650M parameter DSM model
- **DSM-ppi Models**:
(LoRA versions - results reported in paper but not recommended for real use)
- [GleghornLab/DSM_150_ppi_lora](https://huggingface.co/GleghornLab/DSM_150_ppi_lora) - 150M parameter LoRA DSM-ppi model
- [GleghornLab/DSM_650_ppi_lora](https://huggingface.co/GleghornLab/DSM_650_ppi_lora) - 650M parameter LoRA DSM-ppi model
- [GleghornLab/DSM_150_ppi_control](https://huggingface.co/GleghornLab/DSM_150_ppi_control) - Control version of LoRA DSM-ppi
(Fully finetuned - recommended for real use)
- [Synthyra/DSM_ppi_full](https://huggingface.co/Synthyra/DSM_ppi_full) - 650M parameter DSM-ppi model
- **Datasets**:
- [Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50) - Open MetaGenomic dataset clustered at 50% identity (207M sequences)
- [GleghornLab/stringv12_modelorgs_9090](https://huggingface.co/datasets/GleghornLab/stringv12_modelorgs_9090) - STRING database model organisms (653k sequences)
- **Utility Models**:
- [GleghornLab/production_ss4_model](https://huggingface.co/GleghornLab/production_ss4_model) - Secondary structure prediction (4-class)
- [GleghornLab/production_ss9_model](https://huggingface.co/GleghornLab/production_ss9_model) - Secondary structure prediction (9-class)
## Usage
This section outlines how to use a trained `DSM` model for common generation tasks. The core generation logic is provided by the `GenerateMixin` class, used by `DSM` models.
First, ensure you have a trained model (either one you trained or a pre-trained one from Hugging Face Hub) and the necessary environment set up.
```python
import torch
from models.modeling_dsm import DSM # Or DSM_ppi for binder generation
# Load a pre-trained model
model_name_or_path = "GleghornLab/DSM_650" # Replace with your model of choice
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = DSM.from_pretrained(model_name_or_path).to(device).eval()
tokenizer = model.tokenizer
```
```console
You are using a model of type esm_diff to instantiate a model of type dsm. This is not supported for all configurations of models and can yield errors.
```
This warning is normal - all good!
### 1. Unconditional Sequence Generation
To generate a novel sequence of a specific length. DSM uses a progressive denoising approach.
```python
### Unconditional generation
length = 100
mask_token = tokenizer.mask_token
# optionally, enforce starting with methionine
input_tokens = tokenizer.encode('M' + ''.join([mask_token] * (length - 1)), add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
generated_sequences = model.decode_output(output)
print(f"Generated sequence: {generated_sequences[0]}")
```
```console
Generated sequence: MFRVDALQVAQQETLAIGRSTAYDKQESPSMAQRQVLTQLAAYGGENDLRQICIPAERRNFLSIANGASYQFVEEDNEANGGYWSPHKAGLPESACKRFI
```
### 2. Mask Filling (Inpainting)
To fill in masked regions of a template sequence:
```python
# Mask Filling / Inpainting
template_sequence = "MA<mask><mask><mask>KEG<mask><mask>STL"
input_tokens = tokenizer.encode(template_sequence, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
generated_sequences = model.decode_output(output)
print(f"Generated sequence: {generated_sequences[0]}")
```
```console
Generated sequence: MAVKFKEGGISTL
```
### 3. Conditional Generation (e.g., Binders - using DSM-ppi)
```python
# from models.modeling_dsm import DSM_ppi
# model_binder = DSM_ppi.from_pretrained("GleghornLab/DSM_650_ppi_lora").to(device).eval()
# The lora version from the paper leads to unreliable outputs
# Synthyra has generously trained a version through full fine tuning
model = DSM.from_pretrained("Synthyra/DSM_ppi_full").to(device).eval()
# BBF-14
target_seq = "MGTPLWALLGGPWRGTATYEDGTKVTLDYRYTRVSPDRLRADVTYTTPDGTTLEATVDLWKDANGVIRYHATYPDGTSADGTLTQLDADTLLATGTYDDGTKYTVTLTRVAPGSGWHHHHHH"
# For binder generation, the 'interactor' (SeqB) part is what gets generated/filled.
# Start with a fully masked interactor of desired length.
interactor_template_len = 256
interactor_template = ''.join([mask_token] * interactor_template_len)
combined_input_str = target_seq + '<eos>' + interactor_template
input_tokens = tokenizer.encode(combined_input_str, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
target, binder = model.decode_dual_input(output, seperator='<eos>')
# Parse out the generated interactor part based on EOS tokens.
# Example: generated_full_seq_str.split(model_binder.tokenizer.eos_token)[1]
print(f"Generated binder {binder[0]}")
```
```console
Generated binder HRHHHRRPTHARETEWLARMRLGIAEHQRIAVPRSDLEPDQMRERAADNQRLVKEYDQVIDHQTEGSTERLFEVLRVWEQVNTEQAHHEASAALEFGRVGYPDDEGGRAFYTQANAHKKDLVEYIGGIDEDAKWDPRIAWLMPEGGQPVKATVIGVSEERINGLKVLDDHWGRERRLWLINLFTALQAYDDPTRPTQVTLTPATDQLTNDVQYLLLSTRYTPPGVTTAVKIRKLDGRTLKVLTTEAPYVVRGATLS
```
Folded with Chai1:
`Synthyra/DSM_ppi_full` was actually trained to fill masks from any part of SeqA and SeqB. That means you can fully hallucinate plausibly interacting protein pairs.
```python
seq_a_length = 128
seq_b_length = 128
seq_a_template = ''.join([mask_token] * seq_a_length)
seq_b_template = ''.join([mask_token] * seq_b_length)
combined_input_str = seq_a_template + '<eos>' + seq_b_template
input_tokens = tokenizer.encode(combined_input_str, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=10, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
seqa, seqb = model.decode_dual_input(output, seperator='<eos>')
# Parse out the generated interactor part based on EOS tokens.
# Example: generated_full_seq_str.split(model_binder.tokenizer.eos_token)[1]
print(f"SeqA: {seqa[0][5:]}") # remove cls token
print(f"SeqB: {seqb[0]}")
```
```console
SeqA: MVNLAKMRQRTEQNLREVSSFVKILFHTVLKFPMKINIGIHVHINMQAAQNAAADQNMQATNVIDLHNFKMGKDIGVDNKASATAHIYDEAHHTFLQLGAIKLLHAIPMIAGPVRCRLPIGFGHRFRG
SeqB: HYKNPMHSLLDSNVLHKDVVEVRLPIKIGMELDVMASAMREFLMPGTQQGDLRVIAEKRPVNKLHTYRRDLVKLLLAGAKLGTEAKSVELDLYRTELGGLVVYIININIATWDIIFAKVKICRGNDKP
```
Folded with Chai1:
## Demos
There are various demos with many more to come. For example, in `demo_dsm_ppi_full.py` (run by `python -m demos.demo_dsm_ppi_full`) we perform a test on DSM-ppi.
We take 1000 protein pairs from BIOGRID (real protein-protein interactions) and 1000 from Negatome (non interacting protein pairs) and mask the second sequence (SeqB) by 50%.
This acts as a sanity check, as we expect the accuracy on reconstructing real positive PPIs to be higher than the accuracy on non-interacting proteins.
Indeed, this is the case:
```console
==================================================
RESULTS COMPARISON
==================================================
Positive examples:
Mean accuracy: 0.495 ± 0.322
Processed: 1000 examples
Negative examples:
Mean accuracy: 0.227 ± 0.231
Processed: 1000 examples
Difference (Positive - Negative): 0.267
T-test: t=21.331, p=0.000
Difference is statistically significant (p < 0.05)
```
## Installation
1. **Clone the repository:**
```bash
git clone <repository-url>
cd <repository-name>
```
2. **Initialize the submodules:**
```bash
git submodule update --init --remote --recursive
```
3. **Set up the Python virtual environment:**
The `setup_bioenv.sh` script creates a virtual environment named `bioenv` in your home directory (`~/bioenv`), installs PyTorch with CUDA 12.6 support, and then installs all other dependencies from `requirements.txt`.
Make the script executable:
```bash
chmod +x setup_bioenv.sh
```
Run the script:
```bash
./setup_bioenv.sh
```
If you are not on a linux machine, you can install the requirements directly
```console
python -m pip install -r requirements.txt
```
4. **Activate the environment:**
Each time you want to work on this project, activate the virtual environment:
```bash
source ~/bioenv/bin/activate
```
5. **To deactivate the environment:**
```bash
deactivate
```
## Training
The primary script for training models is `training/train_dsm.py`. This script further pretrains an ESM2 checkpoint using the DSM objective (masked diffusion based on LLaDA) on a large protein sequence dataset like [OMG-prot50](https://huggingface.co/datasets/Synthyra/omg_prot50).
### Main Training Script: `train_dsm.py`
- **Base Model**: DSM models are extended from pre-trained ESM2 checkpoints (e.g., ESM2-150M, ESM2-650M).
- **Training Objective**: Masked diffusion loss, where the model predicts masked tokens. The loss is scaled by `1/(t + epsilon)` where `t` is the corruption level, penalizing errors more at low mask rates.
- **Language Modeling Head**: Uses a modified head with a soft-logit cap (`tau=30`) and tied output projection weights to the token embeddings.
- **Data Handling**:
- Training data can be streamed from datasets like [Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50) (a version of Open MetaGenomic dataset clustered at 50% identity).
- Uses `data.dataset_classes.SequenceDatasetFromList` for validation/test sets and `data.dataset_classes.IterableDatasetFromHF` for streaming training.
- `data.data_collators.SequenceCollator` is used for batching.
- **Training Process**:
- Utilizes Hugging Face `TrainingArguments`.
- A custom `IterableTrainer` (from `training.iterable_trainer.py`) handles iterable datasets.
- Uses AdamW optimizer and a cosine learning rate scheduler with linear warmup.
- Supports logging to Weights & Biases (wandb).
- The trained model can be pushed to Hugging Face Hub.
- Example checkpoints mentioned in the paper: [DSM-150](https://huggingface.co/GleghornLab/DSM_150) (from ESM2-150M, 100k steps, batch 32, seqlen 512, LR 1e-4) and [DSM-650](https://huggingface.co/GleghornLab/DSM_650) (from ESM2-650M, 100k steps, global batch 128, seqlen 2048, LR 1e-4).
**Usage Example:**
```bash
python -m training.train_dsm \
--model_path facebook/esm2_t33_650M_UR50D \
--save_path GleghornLab/DSM_650 \
--lr 1e-4 \
--batch_size 8 \
--grad_accum 16 \
--max_steps 100000 \
--save_every 1000 \
--fp16 \
--wandb_project "DSM_Training" \
--token <your_hf_token_if_needed_for_private_repo_or_saving>
```
**Key Command-Line Arguments for `train_dsm.py`:**
* `--token`: Hugging Face token.
* `--model_path`: Path to the base ESM2 model to start from.
* `--save_path`: Path to save the trained DSM model on Hugging Face Hub.
* `--lr`: Learning rate.
* `--batch_size`: Batch size per device.
* `--grad_accum`: Gradient accumulation steps.
* `--max_steps`: Maximum training steps.
* `--wandb_project`: Wandb project name (default: `DSM`).
* `--max_length`: Maximum sequence length.
* `--save_every`: Save model and evaluate every N steps.
* `--fp16`: Enable mixed-precision training.
* `--bugfix`: Use small batch size and max length for debugging.
### Other Training Scripts (e.g., for DSM-ppi)
The `training/` directory may also contain scripts like `train_dsm_bind.py`.
- DSM-ppi (e.g., [DSM-150-ppi](https://huggingface.co/GleghornLab/DSM_150_ppi_lora), [DSM-650-ppi](https://huggingface.co/GleghornLab/DSM_650_ppi_lora)) is fine-tuned on PPI datasets.
- Training involves conditioning on a target sequence (SeqA) to generate an interactor (SeqB) using the format `[CLS]--SeqA--[EOS]--[MASKED~SeqB]--[EOS]`.
- LoRA (Low-Rank Adaptation) can be applied to attention layers for efficient fine-tuning.
And `training/iterable_trainer.py` provides the `get_iterable_trainer` function used by `train_dsm.py` to enable training with iterable datasets.
## Evaluation
The repository includes a comprehensive suite for evaluating model performance, focusing on:
1. **Sequence Reconstruction (Mask Filling):**
* Evaluated by masking validation/test sets at various corruption rates (5% to 90%) and measuring cross-entropy loss, weighted F1 score, and Alignment Score (ASc) for the masked positions.
* The script `evaluation/mask_filling.py` is central to this.
2. **Unconditional Generation Quality:**
* Generate a corpus of sequences based on lengths from a reference set (e.g., validation data).
* Compare distributions (1-mers, 2-mers, 3-mers) of amino acids and predicted secondary structures between generated and natural sequences using χ² test and Jensen-Shannon (JS) divergence.
* Compare distributions of predicted functional annotations (e.g., using Annotation Vocabulary - AV terms).
* Scripts involved: `evaluation/unconditional_generation_tuning.py` (to find optimal generation parameters like temperature and step divisor `s`), `evaluation/unconditional_generation.py`, `evaluation/ss_pred.py` (using [production_ss4_model](https://huggingface.co/GleghornLab/production_ss4_model) or [production_ss9_model](https://huggingface.co/GleghornLab/production_ss9_model)), `evaluation/annotate_comparisons.py`, `evaluation/compare_distributions.py`, `evaluation/plot_distribution_comparisons.py`.
* The `run_eval_pipeline.py` script automates this workflow.
3. **Representation Quality (Model Probing):**
* Evaluate learned embeddings by training linear probes (or simple transformer blocks) on various downstream tasks (e.g., secondary structure prediction, localization prediction, etc.).
* Performance is compared against random vectors, randomized transformers, and other established pLMs.
* The assessment was done with [Protify](https://github.com/Synthyra/Protify), an open-source framework that can be used for pLM training and evaluation.
4. **Conditional Generation (Binder Design for DSM-ppi):**
* Evaluate DSM-ppi on benchmarks like BenchBB.
* Generate binders for target proteins using template-based masking strategies.
* Assess generated binders using *in-silico* tools like Synteract2 for predicted binding affinity (ppKd).
The `evaluation/` directory also contains a `readme.md` which provides further details on some evaluation workflows. Key metrics used include:
- **Alignment Score (ASc):** A normalized Needleman-Wunsch global alignment score (using BLOSUM62) to measure sequence similarity, robust to length variations. ASc(a, b) = l/(f(a, a) - f(a, b) + l).
- **Jensen-Shannon (JS) Divergence:** To compare distributions of k-mers and functional terms.
**Running the Full Unconditional Evaluation Pipeline:**
```bash
python run_eval_pipeline.py --token YOUR_HF_TOKEN --data_dir ./evaluation_results
```
Refer to `run_eval_pipeline.py --help` for more options, such as `--skip_tuning`.
### Mask Filling Evaluation
The script `evaluation/mask_filling.py` is used to evaluate models on their ability to predict masked tokens in a sequence across various masking rates.
- **Functionality:**
- Evaluates different models (DSM, DPLM, standard ESM models).
- Tests across multiple datasets ([Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50), [GleghornLab/stringv12_modelorgs_9090](https://huggingface.co/datasets/GleghornLab/stringv12_modelorgs_9090)).
- Calculates metrics: loss, perplexity, precision, recall, F1, accuracy, MCC, and alignment score.
- Saves detailed results to CSV files.
- Can generate a summary plot comparing model performance across different mask rates using `evaluation/plot_mask_fill_results.py`.
- **Usage Example:**
```bash
python -m evaluation.mask_filling \
--token YOUR_HF_TOKEN \
--batch_size 4 \
--mask_rates 0.15 0.30 0.50 \
--data_splits valid test \
--results_dir ./results/mask_fill_custom
```
To generate a comparison plot from existing results:
```bash
python -m evaluation.mask_filling --generate_comparison_plot --results_dir ./results/mask_fill_custom --plot_output ./results/mask_fill_custom/comparison.png
```
### Other Evaluation Scripts
The `evaluation/` directory contains additional scripts for more specific analyses. These are typically run independently:
- `evaluation/all_targets_uncond.py` and `evaluation/all_targets_cond.py`: Likely for evaluating generation towards specific targets, unconditionally and conditionally.
- `evaluation/conditional_binder.py` and `evaluation/unconditional_binder.py`: Suggest evaluation focused on generating protein binders.
- `evaluation/unconditional_by_length.py`: May evaluate unconditional generation focusing on sequence length distributions.
- `evaluation/utils.py`: Utility functions for evaluation scripts.
Users should refer to individual scripts (e.g., using `python -m evaluation.<script_name> --help`) for their specific usage and arguments.
The `evaluation/` directory also contains a `readme.md` which provides further details on the unconditional generation evaluation workflow.
## Results
DSM demonstrates strong performance in both protein sequence generation and representation learning, establishing masked diffusion as a powerful paradigm.
- **Biomimetic Sequence Generation**: Unconditionally generated DSM sequences closely mimic natural protein distributions in terms of amino acid k-mers, predicted secondary structures (JS divergence < 0.01 for AA k-mers), and predicted functional annotations (AV terms, JS divergence ~0.1). This suggests DSM captures underlying biological principles.
- **Superior Sequence Reconstruction**: DSM models significantly outperform MLM-based ESM2 models in reconstructing sequences from highly corrupted inputs (up to 90% masking).
- At 90% masking, DSM achieves an Alignment Score (ASc) of ~0.27, considerably higher than random.
- DSM models show higher F1 scores in reconstruction tasks compared to DPLM models, especially at high mask rates.
- **High-Quality Embeddings**: DSM embeddings match or exceed the quality of those from comparably sized pLMs (ESM2, DPLM) and even larger autoregressive models (ProtCLM 1B) on various downstream tasks evaluated by linear probing. [DSM-650](https://huggingface.co/GleghornLab/DSM_650) generally provides the best representations among tested models of similar size.
- **Effective Binder Design (DSM-ppi):**
- DSM-ppi fine-tuned on protein-protein interaction data, demonstrates the ability to generate protein binders conditioned on target sequences.
- On the BenchBB benchmark, DSM-generated binders (both unconditional DSM and conditional DSM-ppi) show promising predicted binding affinities, in some cases superior to known binders. For example, designs for EGFR showed high predicted pKd and good structural metrics (ipTM, pTM with AlphaFold3).
- **Efficiency**: DSM can generate realistic protein sequences from a single forward pass during reconstruction tasks at high mask rates, offering potential efficiency advantages over iterative AR or some discrete diffusion models.
These results highlight DSM's capability to unify high-quality protein representation learning and biologically coherent generative modeling within a single framework.
## Cite
```
@misc{hallee2025diffusionsequencemodelsenhanced,
title={Diffusion Sequence Models for Enhanced Protein Representation and Generation},
author={Logan Hallee and Nikolaos Rafailidis and David B. Bichara and Jason P. Gleghorn},
year={2025},
eprint={2506.08293},
archivePrefix={arXiv},
primaryClass={q-bio.BM},
url={https://arxiv.org/abs/2506.08293},
}
```
|
GleghornLab/DSM_650
|
GleghornLab
| 2025-06-23T15:22:49Z | 259 | 0 |
transformers
|
[
"transformers",
"safetensors",
"esm_diff",
"custom_code",
"arxiv:2506.08293",
"endpoints_compatible",
"region:us"
] | null | 2025-05-05T05:58:30Z |
---
library_name: transformers
tags: []
---
# DSM: Diffusion Models for Protein Sequence Generation
### Note: This readme is shared between our GitHub and Huggingface pages.
## Table of Contents
- [Introduction](#introduction)
- [Models](#models)
- [Usage](#usage)
- [Demos](#usage)
- [Local installation](#installation)
- [Training](#training)
- [Evaluation](#evaluation)
- [Results](#results)
- [Cite](#cite)
## Introduction
DSM (Diffusion Sequence Model) is a novel Protein Language Model (pLM) developed in collaboration between the [Gleghorn Lab](https://www.gleghornlab.com/) and [Synthyra](https://synthyra.com/). It was trained with masked diffusion to enable both high-quality representation learning and generative protein design. This repository contains the code for training, evaluating, and applying DSM and its variants.
DSM is capable of generating diverse, biomimetic sequences that align with expected amino acid compositions, secondary structures, and predicted functions. Furthermore, DSM's learned representations match or exceed those of comparably sized pLMs on various downstream tasks. DSM is detailed extensively in our [preprint](https://arxiv.org/abs/2506.08293) (which is currently in review). Beyond the base and PPI variants, we are currently training versions to jointly diffuse over sequence and foldseek tokens, as well as [Annotation Vocabulary](https://www.biorxiv.org/content/10.1101/2024.07.30.605924v1) tokens. Since the preprint release, Synthyra has trained [Synthyra/DSM_ppi_full](https://huggingface.co/Synthyra/DSM_ppi_full) which neglects the LoRA PPI training in favor for full finetuning. Additionally, the sequences SeqA and SeqB are jointly masked instead of just SeqB in the original version. We plan on adding the **many** new results to the second version of the preprint and eventual journal article.
## Models
Relevant Huggingface hosted models and datasets
- **Base DSM Models**:
- [GleghornLab/DSM_150](https://huggingface.co/GleghornLab/DSM_150) - 150M parameter DSM model
- [GleghornLab/DSM_650](https://huggingface.co/GleghornLab/DSM_650) - 650M parameter DSM model
- **DSM-ppi Models**:
(LoRA versions - results reported in paper but not recommended for real use)
- [GleghornLab/DSM_150_ppi_lora](https://huggingface.co/GleghornLab/DSM_150_ppi_lora) - 150M parameter LoRA DSM-ppi model
- [GleghornLab/DSM_650_ppi_lora](https://huggingface.co/GleghornLab/DSM_650_ppi_lora) - 650M parameter LoRA DSM-ppi model
- [GleghornLab/DSM_150_ppi_control](https://huggingface.co/GleghornLab/DSM_150_ppi_control) - Control version of LoRA DSM-ppi
(Fully finetuned - recommended for real use)
- [Synthyra/DSM_ppi_full](https://huggingface.co/Synthyra/DSM_ppi_full) - 650M parameter DSM-ppi model
- **Datasets**:
- [Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50) - Open MetaGenomic dataset clustered at 50% identity (207M sequences)
- [GleghornLab/stringv12_modelorgs_9090](https://huggingface.co/datasets/GleghornLab/stringv12_modelorgs_9090) - STRING database model organisms (653k sequences)
- **Utility Models**:
- [GleghornLab/production_ss4_model](https://huggingface.co/GleghornLab/production_ss4_model) - Secondary structure prediction (4-class)
- [GleghornLab/production_ss9_model](https://huggingface.co/GleghornLab/production_ss9_model) - Secondary structure prediction (9-class)
## Usage
This section outlines how to use a trained `DSM` model for common generation tasks. The core generation logic is provided by the `GenerateMixin` class, used by `DSM` models.
First, ensure you have a trained model (either one you trained or a pre-trained one from Hugging Face Hub) and the necessary environment set up.
```python
import torch
from models.modeling_dsm import DSM # Or DSM_ppi for binder generation
# Load a pre-trained model
model_name_or_path = "GleghornLab/DSM_650" # Replace with your model of choice
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = DSM.from_pretrained(model_name_or_path).to(device).eval()
tokenizer = model.tokenizer
```
```console
You are using a model of type esm_diff to instantiate a model of type dsm. This is not supported for all configurations of models and can yield errors.
```
This warning is normal - all good!
### 1. Unconditional Sequence Generation
To generate a novel sequence of a specific length. DSM uses a progressive denoising approach.
```python
### Unconditional generation
length = 100
mask_token = tokenizer.mask_token
# optionally, enforce starting with methionine
input_tokens = tokenizer.encode('M' + ''.join([mask_token] * (length - 1)), add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
generated_sequences = model.decode_output(output)
print(f"Generated sequence: {generated_sequences[0]}")
```
```console
Generated sequence: MFRVDALQVAQQETLAIGRSTAYDKQESPSMAQRQVLTQLAAYGGENDLRQICIPAERRNFLSIANGASYQFVEEDNEANGGYWSPHKAGLPESACKRFI
```
### 2. Mask Filling (Inpainting)
To fill in masked regions of a template sequence:
```python
# Mask Filling / Inpainting
template_sequence = "MA<mask><mask><mask>KEG<mask><mask>STL"
input_tokens = tokenizer.encode(template_sequence, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
generated_sequences = model.decode_output(output)
print(f"Generated sequence: {generated_sequences[0]}")
```
```console
Generated sequence: MAVKFKEGGISTL
```
### 3. Conditional Generation (e.g., Binders - using DSM-ppi)
```python
# from models.modeling_dsm import DSM_ppi
# model_binder = DSM_ppi.from_pretrained("GleghornLab/DSM_650_ppi_lora").to(device).eval()
# The lora version from the paper leads to unreliable outputs
# Synthyra has generously trained a version through full fine tuning
model = DSM.from_pretrained("Synthyra/DSM_ppi_full").to(device).eval()
# BBF-14
target_seq = "MGTPLWALLGGPWRGTATYEDGTKVTLDYRYTRVSPDRLRADVTYTTPDGTTLEATVDLWKDANGVIRYHATYPDGTSADGTLTQLDADTLLATGTYDDGTKYTVTLTRVAPGSGWHHHHHH"
# For binder generation, the 'interactor' (SeqB) part is what gets generated/filled.
# Start with a fully masked interactor of desired length.
interactor_template_len = 256
interactor_template = ''.join([mask_token] * interactor_template_len)
combined_input_str = target_seq + '<eos>' + interactor_template
input_tokens = tokenizer.encode(combined_input_str, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=100, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
target, binder = model.decode_dual_input(output, seperator='<eos>')
# Parse out the generated interactor part based on EOS tokens.
# Example: generated_full_seq_str.split(model_binder.tokenizer.eos_token)[1]
print(f"Generated binder {binder[0]}")
```
```console
Generated binder HRHHHRRPTHARETEWLARMRLGIAEHQRIAVPRSDLEPDQMRERAADNQRLVKEYDQVIDHQTEGSTERLFEVLRVWEQVNTEQAHHEASAALEFGRVGYPDDEGGRAFYTQANAHKKDLVEYIGGIDEDAKWDPRIAWLMPEGGQPVKATVIGVSEERINGLKVLDDHWGRERRLWLINLFTALQAYDDPTRPTQVTLTPATDQLTNDVQYLLLSTRYTPPGVTTAVKIRKLDGRTLKVLTTEAPYVVRGATLS
```
Folded with Chai1:
`Synthyra/DSM_ppi_full` was actually trained to fill masks from any part of SeqA and SeqB. That means you can fully hallucinate plausibly interacting protein pairs.
```python
seq_a_length = 128
seq_b_length = 128
seq_a_template = ''.join([mask_token] * seq_a_length)
seq_b_template = ''.join([mask_token] * seq_b_length)
combined_input_str = seq_a_template + '<eos>' + seq_b_template
input_tokens = tokenizer.encode(combined_input_str, add_special_tokens=True, return_tensors='pt').to(device)
output = model.mask_diffusion_generate(
tokenizer=tokenizer,
input_tokens=input_tokens,
step_divisor=10, # lower is slower but better
temperature=1.0, # sampling temperature
remasking="random", # strategy for remasking tokens not kept
preview=False, # set this to True to watch the mask tokens get rilled in real time
slow=False, # adds a small delay to the real time filling (because it is usually very fast and watching carefully is hard!)
return_trajectory=False # set this to True to return the trajectory of the generation (what you watch in the preview)
) # Note: output will be a tuple if return_trajectory is True
seqa, seqb = model.decode_dual_input(output, seperator='<eos>')
# Parse out the generated interactor part based on EOS tokens.
# Example: generated_full_seq_str.split(model_binder.tokenizer.eos_token)[1]
print(f"SeqA: {seqa[0][5:]}") # remove cls token
print(f"SeqB: {seqb[0]}")
```
```console
SeqA: MVNLAKMRQRTEQNLREVSSFVKILFHTVLKFPMKINIGIHVHINMQAAQNAAADQNMQATNVIDLHNFKMGKDIGVDNKASATAHIYDEAHHTFLQLGAIKLLHAIPMIAGPVRCRLPIGFGHRFRG
SeqB: HYKNPMHSLLDSNVLHKDVVEVRLPIKIGMELDVMASAMREFLMPGTQQGDLRVIAEKRPVNKLHTYRRDLVKLLLAGAKLGTEAKSVELDLYRTELGGLVVYIININIATWDIIFAKVKICRGNDKP
```
Folded with Chai1:
## Demos
There are various demos with many more to come. For example, in `demo_dsm_ppi_full.py` (run by `python -m demos.demo_dsm_ppi_full`) we perform a test on DSM-ppi.
We take 1000 protein pairs from BIOGRID (real protein-protein interactions) and 1000 from Negatome (non interacting protein pairs) and mask the second sequence (SeqB) by 50%.
This acts as a sanity check, as we expect the accuracy on reconstructing real positive PPIs to be higher than the accuracy on non-interacting proteins.
Indeed, this is the case:
```console
==================================================
RESULTS COMPARISON
==================================================
Positive examples:
Mean accuracy: 0.495 ± 0.322
Processed: 1000 examples
Negative examples:
Mean accuracy: 0.227 ± 0.231
Processed: 1000 examples
Difference (Positive - Negative): 0.267
T-test: t=21.331, p=0.000
Difference is statistically significant (p < 0.05)
```
## Installation
1. **Clone the repository:**
```bash
git clone <repository-url>
cd <repository-name>
```
2. **Initialize the submodules:**
```bash
git submodule update --init --remote --recursive
```
3. **Set up the Python virtual environment:**
The `setup_bioenv.sh` script creates a virtual environment named `bioenv` in your home directory (`~/bioenv`), installs PyTorch with CUDA 12.6 support, and then installs all other dependencies from `requirements.txt`.
Make the script executable:
```bash
chmod +x setup_bioenv.sh
```
Run the script:
```bash
./setup_bioenv.sh
```
If you are not on a linux machine, you can install the requirements directly
```console
python -m pip install -r requirements.txt
```
4. **Activate the environment:**
Each time you want to work on this project, activate the virtual environment:
```bash
source ~/bioenv/bin/activate
```
5. **To deactivate the environment:**
```bash
deactivate
```
## Training
The primary script for training models is `training/train_dsm.py`. This script further pretrains an ESM2 checkpoint using the DSM objective (masked diffusion based on LLaDA) on a large protein sequence dataset like [OMG-prot50](https://huggingface.co/datasets/Synthyra/omg_prot50).
### Main Training Script: `train_dsm.py`
- **Base Model**: DSM models are extended from pre-trained ESM2 checkpoints (e.g., ESM2-150M, ESM2-650M).
- **Training Objective**: Masked diffusion loss, where the model predicts masked tokens. The loss is scaled by `1/(t + epsilon)` where `t` is the corruption level, penalizing errors more at low mask rates.
- **Language Modeling Head**: Uses a modified head with a soft-logit cap (`tau=30`) and tied output projection weights to the token embeddings.
- **Data Handling**:
- Training data can be streamed from datasets like [Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50) (a version of Open MetaGenomic dataset clustered at 50% identity).
- Uses `data.dataset_classes.SequenceDatasetFromList` for validation/test sets and `data.dataset_classes.IterableDatasetFromHF` for streaming training.
- `data.data_collators.SequenceCollator` is used for batching.
- **Training Process**:
- Utilizes Hugging Face `TrainingArguments`.
- A custom `IterableTrainer` (from `training.iterable_trainer.py`) handles iterable datasets.
- Uses AdamW optimizer and a cosine learning rate scheduler with linear warmup.
- Supports logging to Weights & Biases (wandb).
- The trained model can be pushed to Hugging Face Hub.
- Example checkpoints mentioned in the paper: [DSM-150](https://huggingface.co/GleghornLab/DSM_150) (from ESM2-150M, 100k steps, batch 32, seqlen 512, LR 1e-4) and [DSM-650](https://huggingface.co/GleghornLab/DSM_650) (from ESM2-650M, 100k steps, global batch 128, seqlen 2048, LR 1e-4).
**Usage Example:**
```bash
python -m training.train_dsm \
--model_path facebook/esm2_t33_650M_UR50D \
--save_path GleghornLab/DSM_650 \
--lr 1e-4 \
--batch_size 8 \
--grad_accum 16 \
--max_steps 100000 \
--save_every 1000 \
--fp16 \
--wandb_project "DSM_Training" \
--token <your_hf_token_if_needed_for_private_repo_or_saving>
```
**Key Command-Line Arguments for `train_dsm.py`:**
* `--token`: Hugging Face token.
* `--model_path`: Path to the base ESM2 model to start from.
* `--save_path`: Path to save the trained DSM model on Hugging Face Hub.
* `--lr`: Learning rate.
* `--batch_size`: Batch size per device.
* `--grad_accum`: Gradient accumulation steps.
* `--max_steps`: Maximum training steps.
* `--wandb_project`: Wandb project name (default: `DSM`).
* `--max_length`: Maximum sequence length.
* `--save_every`: Save model and evaluate every N steps.
* `--fp16`: Enable mixed-precision training.
* `--bugfix`: Use small batch size and max length for debugging.
### Other Training Scripts (e.g., for DSM-ppi)
The `training/` directory may also contain scripts like `train_dsm_bind.py`.
- DSM-ppi (e.g., [DSM-150-ppi](https://huggingface.co/GleghornLab/DSM_150_ppi_lora), [DSM-650-ppi](https://huggingface.co/GleghornLab/DSM_650_ppi_lora)) is fine-tuned on PPI datasets.
- Training involves conditioning on a target sequence (SeqA) to generate an interactor (SeqB) using the format `[CLS]--SeqA--[EOS]--[MASKED~SeqB]--[EOS]`.
- LoRA (Low-Rank Adaptation) can be applied to attention layers for efficient fine-tuning.
And `training/iterable_trainer.py` provides the `get_iterable_trainer` function used by `train_dsm.py` to enable training with iterable datasets.
## Evaluation
The repository includes a comprehensive suite for evaluating model performance, focusing on:
1. **Sequence Reconstruction (Mask Filling):**
* Evaluated by masking validation/test sets at various corruption rates (5% to 90%) and measuring cross-entropy loss, weighted F1 score, and Alignment Score (ASc) for the masked positions.
* The script `evaluation/mask_filling.py` is central to this.
2. **Unconditional Generation Quality:**
* Generate a corpus of sequences based on lengths from a reference set (e.g., validation data).
* Compare distributions (1-mers, 2-mers, 3-mers) of amino acids and predicted secondary structures between generated and natural sequences using χ² test and Jensen-Shannon (JS) divergence.
* Compare distributions of predicted functional annotations (e.g., using Annotation Vocabulary - AV terms).
* Scripts involved: `evaluation/unconditional_generation_tuning.py` (to find optimal generation parameters like temperature and step divisor `s`), `evaluation/unconditional_generation.py`, `evaluation/ss_pred.py` (using [production_ss4_model](https://huggingface.co/GleghornLab/production_ss4_model) or [production_ss9_model](https://huggingface.co/GleghornLab/production_ss9_model)), `evaluation/annotate_comparisons.py`, `evaluation/compare_distributions.py`, `evaluation/plot_distribution_comparisons.py`.
* The `run_eval_pipeline.py` script automates this workflow.
3. **Representation Quality (Model Probing):**
* Evaluate learned embeddings by training linear probes (or simple transformer blocks) on various downstream tasks (e.g., secondary structure prediction, localization prediction, etc.).
* Performance is compared against random vectors, randomized transformers, and other established pLMs.
* The assessment was done with [Protify](https://github.com/Synthyra/Protify), an open-source framework that can be used for pLM training and evaluation.
4. **Conditional Generation (Binder Design for DSM-ppi):**
* Evaluate DSM-ppi on benchmarks like BenchBB.
* Generate binders for target proteins using template-based masking strategies.
* Assess generated binders using *in-silico* tools like Synteract2 for predicted binding affinity (ppKd).
The `evaluation/` directory also contains a `readme.md` which provides further details on some evaluation workflows. Key metrics used include:
- **Alignment Score (ASc):** A normalized Needleman-Wunsch global alignment score (using BLOSUM62) to measure sequence similarity, robust to length variations. ASc(a, b) = l/(f(a, a) - f(a, b) + l).
- **Jensen-Shannon (JS) Divergence:** To compare distributions of k-mers and functional terms.
**Running the Full Unconditional Evaluation Pipeline:**
```bash
python run_eval_pipeline.py --token YOUR_HF_TOKEN --data_dir ./evaluation_results
```
Refer to `run_eval_pipeline.py --help` for more options, such as `--skip_tuning`.
### Mask Filling Evaluation
The script `evaluation/mask_filling.py` is used to evaluate models on their ability to predict masked tokens in a sequence across various masking rates.
- **Functionality:**
- Evaluates different models (DSM, DPLM, standard ESM models).
- Tests across multiple datasets ([Synthyra/omg_prot50](https://huggingface.co/datasets/Synthyra/omg_prot50), [GleghornLab/stringv12_modelorgs_9090](https://huggingface.co/datasets/GleghornLab/stringv12_modelorgs_9090)).
- Calculates metrics: loss, perplexity, precision, recall, F1, accuracy, MCC, and alignment score.
- Saves detailed results to CSV files.
- Can generate a summary plot comparing model performance across different mask rates using `evaluation/plot_mask_fill_results.py`.
- **Usage Example:**
```bash
python -m evaluation.mask_filling \
--token YOUR_HF_TOKEN \
--batch_size 4 \
--mask_rates 0.15 0.30 0.50 \
--data_splits valid test \
--results_dir ./results/mask_fill_custom
```
To generate a comparison plot from existing results:
```bash
python -m evaluation.mask_filling --generate_comparison_plot --results_dir ./results/mask_fill_custom --plot_output ./results/mask_fill_custom/comparison.png
```
### Other Evaluation Scripts
The `evaluation/` directory contains additional scripts for more specific analyses. These are typically run independently:
- `evaluation/all_targets_uncond.py` and `evaluation/all_targets_cond.py`: Likely for evaluating generation towards specific targets, unconditionally and conditionally.
- `evaluation/conditional_binder.py` and `evaluation/unconditional_binder.py`: Suggest evaluation focused on generating protein binders.
- `evaluation/unconditional_by_length.py`: May evaluate unconditional generation focusing on sequence length distributions.
- `evaluation/utils.py`: Utility functions for evaluation scripts.
Users should refer to individual scripts (e.g., using `python -m evaluation.<script_name> --help`) for their specific usage and arguments.
The `evaluation/` directory also contains a `readme.md` which provides further details on the unconditional generation evaluation workflow.
## Results
DSM demonstrates strong performance in both protein sequence generation and representation learning, establishing masked diffusion as a powerful paradigm.
- **Biomimetic Sequence Generation**: Unconditionally generated DSM sequences closely mimic natural protein distributions in terms of amino acid k-mers, predicted secondary structures (JS divergence < 0.01 for AA k-mers), and predicted functional annotations (AV terms, JS divergence ~0.1). This suggests DSM captures underlying biological principles.
- **Superior Sequence Reconstruction**: DSM models significantly outperform MLM-based ESM2 models in reconstructing sequences from highly corrupted inputs (up to 90% masking).
- At 90% masking, DSM achieves an Alignment Score (ASc) of ~0.27, considerably higher than random.
- DSM models show higher F1 scores in reconstruction tasks compared to DPLM models, especially at high mask rates.
- **High-Quality Embeddings**: DSM embeddings match or exceed the quality of those from comparably sized pLMs (ESM2, DPLM) and even larger autoregressive models (ProtCLM 1B) on various downstream tasks evaluated by linear probing. [DSM-650](https://huggingface.co/GleghornLab/DSM_650) generally provides the best representations among tested models of similar size.
- **Effective Binder Design (DSM-ppi):**
- DSM-ppi fine-tuned on protein-protein interaction data, demonstrates the ability to generate protein binders conditioned on target sequences.
- On the BenchBB benchmark, DSM-generated binders (both unconditional DSM and conditional DSM-ppi) show promising predicted binding affinities, in some cases superior to known binders. For example, designs for EGFR showed high predicted pKd and good structural metrics (ipTM, pTM with AlphaFold3).
- **Efficiency**: DSM can generate realistic protein sequences from a single forward pass during reconstruction tasks at high mask rates, offering potential efficiency advantages over iterative AR or some discrete diffusion models.
These results highlight DSM's capability to unify high-quality protein representation learning and biologically coherent generative modeling within a single framework.
## Cite
```
@misc{hallee2025diffusionsequencemodelsenhanced,
title={Diffusion Sequence Models for Enhanced Protein Representation and Generation},
author={Logan Hallee and Nikolaos Rafailidis and David B. Bichara and Jason P. Gleghorn},
year={2025},
eprint={2506.08293},
archivePrefix={arXiv},
primaryClass={q-bio.BM},
url={https://arxiv.org/abs/2506.08293},
}
```
|
blazarev/roberta-structure-hub
|
blazarev
| 2025-06-23T15:21:41Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"roberta",
"text-classification",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2025-06-23T15:21:15Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
AylaEmeryIris/Jaipur.hotel.couple.viral.video
|
AylaEmeryIris
| 2025-06-23T15:21:34Z | 0 | 0 | null |
[
"region:us"
] | null | 2025-06-23T15:18:50Z |
[](https://t-me-viral-now01.blogspot.com/2025/06/ghds.html)
|
Hachipo/OpenCoder-8B-Base-MIFT-en_newbase_v1-EnTrans_5000
|
Hachipo
| 2025-06-23T15:20:47Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"trl",
"sft",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-06-23T15:17:46Z |
---
library_name: transformers
tags:
- trl
- sft
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
tscstudios/iwal7zawwerd8k7vjzyubn9guup1_5d0a470f-5192-4946-888b-a4a18d6ca051
|
tscstudios
| 2025-06-23T15:19:07Z | 0 | 0 |
diffusers
|
[
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] |
text-to-image
| 2025-06-23T15:19:05Z |
---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: TOK
---
# Iwal7Zawwerd8K7Vjzyubn9Guup1_5D0A470F 5192 4946 888B A4A18D6Ca051
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `TOK` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "TOK",
"lora_weights": "https://huggingface.co/tscstudios/iwal7zawwerd8k7vjzyubn9guup1_5d0a470f-5192-4946-888b-a4a18d6ca051/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('tscstudios/iwal7zawwerd8k7vjzyubn9guup1_5d0a470f-5192-4946-888b-a4a18d6ca051', weight_name='lora.safetensors')
image = pipeline('TOK').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/tscstudios/iwal7zawwerd8k7vjzyubn9guup1_5d0a470f-5192-4946-888b-a4a18d6ca051/discussions) to add images that show off what you’ve made with this LoRA.
|
encku/glc-06-2025v2
|
encku
| 2025-06-23T15:17:39Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"vit",
"image-classification",
"autotrain",
"base_model:google/vit-large-patch32-384",
"base_model:finetune:google/vit-large-patch32-384",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2025-06-23T14:58:38Z |
---
tags:
- autotrain
- transformers
- image-classification
base_model: google/vit-large-patch32-384
widget:
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg
example_title: Tiger
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg
example_title: Teapot
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg
example_title: Palace
---
# Model Trained Using AutoTrain
- Problem type: Image Classification
## Validation Metrics
loss: 0.0068761385045945644
f1_macro: 0.998209495397491
f1_micro: 0.9982098102398854
f1_weighted: 0.9982094953974912
precision_macro: 0.9982146649917772
precision_micro: 0.9982098102398854
precision_weighted: 0.9982146649917772
recall_macro: 0.9982098102398854
recall_micro: 0.9982098102398854
recall_weighted: 0.9982098102398854
accuracy: 0.9982098102398854
|
ziadrone/qwen_test_3_vedant
|
ziadrone
| 2025-06-23T15:15:14Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-06-23T15:13:40Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
leobianco/npov_RM_google_S130104_LLM_false_STRUCT_false_epo3_lr1e-3_r8_2506231509
|
leobianco
| 2025-06-23T15:13:08Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T15:09:50Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
hasdal/c3e26798-d4f9-46ce-a0c1-aae4e18d08b1
|
hasdal
| 2025-06-23T15:13:07Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T11:56:49Z |
---
library_name: transformers
tags:
- unsloth
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
IntelliGrow/ppo-SnowballTarget
|
IntelliGrow
| 2025-06-23T15:11:52Z | 0 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"SnowballTarget",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SnowballTarget",
"region:us"
] |
reinforcement-learning
| 2025-06-23T15:11:48Z |
---
library_name: ml-agents
tags:
- SnowballTarget
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SnowballTarget
---
# **ppo** Agent playing **SnowballTarget**
This is a trained model of a **ppo** agent playing **SnowballTarget**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: IntelliGrow/ppo-SnowballTarget
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
Tyl3rDrden/trained-flux-dev-dreambooth-lora_LILY_3.1
|
Tyl3rDrden
| 2025-06-23T15:10:44Z | 0 | 0 |
diffusers
|
[
"diffusers",
"text-to-image",
"diffusers-training",
"lora",
"flux",
"flux-diffusers",
"template:sd-lora",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] |
text-to-image
| 2025-06-23T11:06:46Z |
---
base_model: black-forest-labs/FLUX.1-dev
library_name: diffusers
license: other
instance_prompt: a photo of <sksdog>
widget:
- text: <sksdog> fighting Iranian ballistic missiles
output:
url: image_0.png
- text: <sksdog> fighting Iranian ballistic missiles
output:
url: image_1.png
- text: <sksdog> fighting Iranian ballistic missiles
output:
url: image_2.png
- text: <sksdog> fighting Iranian ballistic missiles
output:
url: image_3.png
tags:
- text-to-image
- diffusers-training
- diffusers
- lora
- flux
- flux-diffusers
- template:sd-lora
---
<!-- This model card has been generated automatically according to the information the training script had access to. You
should probably proofread and complete it, then remove this comment. -->
# Flux DreamBooth LoRA - Tyl3rDrden/trained-flux-dev-dreambooth-lora_LILY_3.1
<Gallery />
## Model description
These are Tyl3rDrden/trained-flux-dev-dreambooth-lora_LILY_3.1 DreamBooth LoRA weights for black-forest-labs/FLUX.1-dev.
The weights were trained using [DreamBooth](https://dreambooth.github.io/) with the [Flux diffusers trainer](https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/README_flux.md).
Was LoRA for the text encoder enabled? False.
Pivotal tuning was enabled: True.
## Trigger words
To trigger image generation of trained concept(or concepts) replace each concept identifier in you prompt with the new inserted tokens:
to trigger concept `TOK` → use `<s0><s1>` in your prompt
## Download model
[Download the *.safetensors LoRA](Tyl3rDrden/trained-flux-dev-dreambooth-lora_LILY_3.1/tree/main) in the Files & versions tab.
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
from huggingface_hub import hf_hub_download
from safetensors.torch import load_file
pipeline = AutoPipelineForText2Image.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16).to('cuda')
pipeline.load_lora_weights('Tyl3rDrden/trained-flux-dev-dreambooth-lora_LILY_3.1', weight_name='pytorch_lora_weights.safetensors')
embedding_path = hf_hub_download(repo_id='Tyl3rDrden/trained-flux-dev-dreambooth-lora_LILY_3.1', filename='trained-flux-dev-dreambooth-lora_LILY_3.1_emb.safetensors', repo_type="model")
state_dict = load_file(embedding_path)
pipeline.load_textual_inversion(state_dict["clip_l"], token=[], text_encoder=pipeline.text_encoder, tokenizer=pipeline.tokenizer)
image = pipeline('<sksdog> fighting Iranian ballistic missiles').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## License
Please adhere to the licensing terms as described [here](https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md).
## Intended uses & limitations
#### How to use
```python
# TODO: add an example code snippet for running this diffusion pipeline
```
#### Limitations and bias
[TODO: provide examples of latent issues and potential remediations]
## Training details
[TODO: describe the data used to train the model]
|
zliang1233214/xiaoyi
|
zliang1233214
| 2025-06-23T15:10:22Z | 0 | 0 | null |
[
"safetensors",
"license:apache-2.0",
"region:us"
] | null | 2025-06-23T15:04:12Z |
---
license: apache-2.0
---
|
hoa12356/grpo_reasoning
|
hoa12356
| 2025-06-23T15:07:37Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"llama",
"trl",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T15:07:24Z |
---
base_model: unsloth/meta-llama-3.1-8b-instruct-unsloth-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- llama
- trl
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** hoa12356
- **License:** apache-2.0
- **Finetuned from model :** unsloth/meta-llama-3.1-8b-instruct-unsloth-bnb-4bit
This llama model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
bittu9988/Gemma_fine-trained_AGG
|
bittu9988
| 2025-06-23T15:06:16Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T15:06:00Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
SsharvienKumar/SASVi
|
SsharvienKumar
| 2025-06-23T15:05:05Z | 0 | 0 | null |
[
"arxiv:2502.09653",
"license:cc-by-4.0",
"region:us"
] | null | 2025-06-23T14:26:03Z |
---
license: cc-by-4.0
---
<div id="top" align="center">
# SASVi - Segment Any Surgical Video (IPCAI 2025)
[](https://arxiv.org/abs/2502.09653)
[](https://link.springer.com/article/10.1007/s11548-025-03408-y)
[](https://huggingface.co/SsharvienKumar/SASVi)
</div>
## Overview
SASVi leverages pre-trained frame-wise object detection and segmentation to re-prompt SAM2
for improved surgical video segmentation with scarcely annotated data.
## Example Results
* You can find the complete segmentations of the video datasets [here](https://huggingface.co/SsharvienKumar/SASVi/tree/main/dataset).
* Checkpoints of the all the overseers can be found [here](https://huggingface.co/SsharvienKumar/SASVi/tree/main/checkpoints).
## Setup
* Create a virtual environment of your choice and activate it: `conda create -n sasvi python=3.11 && conda activate sasvi`
* Install `torch>=2.3.1` and `torchvision>=0.18.1` following the instructions from [here](https://pytorch.org/get-started/locally/)
* Install the dependencies using `pip install -r requirements.txt`
* Install SDS_Playground from [here](https://github.com/MECLabTUDA/SDS_Playground)
* Install SAM2 using `cd src/sam2 && pip install -e .`
* Place SAM2 [checkpoints](https://github.com/facebookresearch/sam2/tree/main#model-description) at `src/sam2/checkpoints`
* Convert video files to frame folders using `bash helper_scripts/video_to_frames.sh`. The output should be in the format:
```
<video_root>
├── <video1>
│ ├── 0001.jpg
│ ├── 0002.jpg
│ └── ...
├── <video2>
│ ├── 0001.jpg
│ ├── 0002.jpg
│ └── ...
└── ...
```
## Overseer Model Training
We provide training scripts for three different overseer models (Mask R-CNN, DETR, Mask2Former)
on three different datasets (CaDIS, CholecSeg8k, Cataract1k).
You can run the training scripts as follows:
`python train_scripts/train_<OVERSEER>_<DATASET>.py`
## SASVi Inference
The frames in the video needs to be extracted beforehand and placed in the formatting above. More optional arguments can be found in the script directly.
```
python src/sam2/eval_sasvi.py \
--sam2_cfg configs/sam2.1_hiera_l.yaml \
--sam2_checkpoint ./checkpoints/<SAM2_CHECKPOINT>.pt \
--overseer_checkpoint <PATH_TO_OVERSEER_CHECKPOINT>.pth \
--overseer_type <NAME_OF_OVERSEER> \
--dataset_type <NAME_OF_DATASET> \
--base_video_dir <PATH_TO_VIDEO_ROOT> \
--output_mask_dir <OUTPUT_PATH_TO_SASVi_MASK> \
--overseer_mask_dir <OPTIONAL - OUTPUT_PATH_TO_OVERSEER_MASK>
```
## nnUNet Training & Inference
Fold 0: `nnUNetv2_train DATASET_ID 2d 0 -p nnUNetResEncUNetMPlans -tr nnUNetTrainer_400epochs --npz`
Fold 1: `nnUNetv2_train DATASET_ID 2d 1 -p nnUNetResEncUNetMPlans -tr nnUNetTrainer_400epochs --npz`
Fold 2: `nnUNetv2_train DATASET_ID 2d 2 -p nnUNetResEncUNetMPlans -tr nnUNetTrainer_400epochs --npz`
Fold 3: `nnUNetv2_train DATASET_ID 2d 3 -p nnUNetResEncUNetMPlans -tr nnUNetTrainer_400epochs --npz`
Fold 4: `nnUNetv2_train DATASET_ID 2d 4 -p nnUNetResEncUNetMPlans -tr nnUNetTrainer_400epochs --npz`
Then find the best configuration using
`nnUNetv2_find_best_configuration DATASET_ID -c 2d -p nnUNetResEncUNetMPlans -tr nnUNetTrainer_400epochs`
And run inference using
`nnUNetv2_predict -d DATASET_ID -i INPUT_FOLDER -o OUTPUT_FOLDER -f 0 1 2 3 4 -tr nnUNetTrainer_400epochs -c 2d -p nnUNetResEncUNetMPlans`
Once inference is completed, run postprocessing
`nnUNetv2_apply_postprocessing -i OUTPUT_FOLDER -o OUTPUT_FOLDER_PP -pp_pkl_file .../postprocessing.pkl -np 8 -plans_json .../plans.json`
## Evaluation
* For frame-wise segmentation evaluation:
* `python eval_scripts/eval_<OVERSEER>_frames.py`
* For frame-wise segmentation prediction on full videos:
* See `python eval_scripts/eval_MaskRCNN_videos.py` for an example.
* For video evaluation:
1. E.g. `python eval_scripts/eval_vid_T.py --segm_root <path_to_segmentation_root> --vid_pattern 'train' --mask_pattern '*.npz' --ignore 255 --device cuda`
2. E.g. `python eval_scripts/eval_vid_F.py --segm_root <path_to_segmentation_root> --frames_root <path_to_frames_root> --vid_pattern 'train' --frames_pattern '*.jpg' --mask_pattern '*.npz' --raft_iters 12 --device cuda`
## TODOs
* [ ] **The code will be refactored soon to be more modular and reusable!**
* [ ] Pre-process Cholec80 videos with out-of-body detection
* [ ] Improve SASVi by combining it with GT prompting (if available)
* [ ] Test SAM2 finetuning
## Citation
If you use SASVi in your research, please cite our paper:
```
@article{sivakumar2025sasvi,
title={SASVi: segment any surgical video},
author={Sivakumar, Ssharvien Kumar and Frisch, Yannik and Ranem, Amin and Mukhopadhyay, Anirban},
journal={International Journal of Computer Assisted Radiology and Surgery},
pages={1--11},
year={2025},
publisher={Springer}
}
```
|
bartowski/Triangulum-10B-GGUF
|
bartowski
| 2025-06-23T15:03:57Z | 1,591 | 0 | null |
[
"gguf",
"triangulum_10b",
"sft",
"chain_of_thought",
"ollama",
"text-generation-inference",
"llama_for_causal_lm",
"reasoning",
"CoT",
"text-generation",
"en",
"de",
"fr",
"it",
"pt",
"hi",
"es",
"th",
"base_model:prithivMLmods/Triangulum-10B",
"base_model:quantized:prithivMLmods/Triangulum-10B",
"license:llama3.1",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-01-04T05:31:27Z |
---
quantized_by: bartowski
pipeline_tag: text-generation
language:
- en
- de
- fr
- it
- pt
- hi
- es
- th
metrics:
- code_eval
- accuracy
- competition_math
- character
base_model: prithivMLmods/Triangulum-10B
license: llama3.1
tags:
- triangulum_10b
- sft
- chain_of_thought
- ollama
- text-generation-inference
- llama_for_causal_lm
- reasoning
- CoT
---
## Llamacpp imatrix Quantizations of Triangulum-10B
Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b4404">b4404</a> for quantization.
Original model: https://huggingface.co/prithivMLmods/Triangulum-10B
All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)
Run them in [LM Studio](https://lmstudio.ai/)
## Prompt format
No prompt format found, check original model page
## Download a file (not the whole branch) from below:
| Filename | Quant type | File Size | Split | Description |
| -------- | ---------- | --------- | ----- | ----------- |
| [Triangulum-10B-f16.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-f16.gguf) | f16 | 20.62GB | false | Full F16 weights. |
| [Triangulum-10B-Q8_0.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q8_0.gguf) | Q8_0 | 10.96GB | false | Extremely high quality, generally unneeded but max available quant. |
| [Triangulum-10B-Q6_K_L.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q6_K_L.gguf) | Q6_K_L | 8.65GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. |
| [Triangulum-10B-Q6_K.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q6_K.gguf) | Q6_K | 8.46GB | false | Very high quality, near perfect, *recommended*. |
| [Triangulum-10B-Q5_K_L.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q5_K_L.gguf) | Q5_K_L | 7.59GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. |
| [Triangulum-10B-Q5_K_M.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q5_K_M.gguf) | Q5_K_M | 7.34GB | false | High quality, *recommended*. |
| [Triangulum-10B-Q5_K_S.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q5_K_S.gguf) | Q5_K_S | 7.14GB | false | High quality, *recommended*. |
| [Triangulum-10B-Q4_K_L.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q4_K_L.gguf) | Q4_K_L | 6.59GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. |
| [Triangulum-10B-Q4_1.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q4_1.gguf) | Q4_1 | 6.53GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. |
| [Triangulum-10B-Q4_K_M.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q4_K_M.gguf) | Q4_K_M | 6.29GB | false | Good quality, default size for most use cases, *recommended*. |
| [Triangulum-10B-Q4_K_S.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q4_K_S.gguf) | Q4_K_S | 5.95GB | false | Slightly lower quality with more space savings, *recommended*. |
| [Triangulum-10B-Q4_0.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q4_0.gguf) | Q4_0 | 5.93GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. |
| [Triangulum-10B-IQ4_NL.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-IQ4_NL.gguf) | IQ4_NL | 5.91GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. |
| [Triangulum-10B-Q3_K_XL.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q3_K_XL.gguf) | Q3_K_XL | 5.80GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
| [Triangulum-10B-IQ4_XS.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-IQ4_XS.gguf) | IQ4_XS | 5.60GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
| [Triangulum-10B-Q3_K_L.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q3_K_L.gguf) | Q3_K_L | 5.45GB | false | Lower quality but usable, good for low RAM availability. |
| [Triangulum-10B-Q3_K_M.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q3_K_M.gguf) | Q3_K_M | 5.05GB | false | Low quality. |
| [Triangulum-10B-IQ3_M.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-IQ3_M.gguf) | IQ3_M | 4.70GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
| [Triangulum-10B-Q3_K_S.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q3_K_S.gguf) | Q3_K_S | 4.59GB | false | Low quality, not recommended. |
| [Triangulum-10B-IQ3_XS.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-IQ3_XS.gguf) | IQ3_XS | 4.37GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
| [Triangulum-10B-Q2_K_L.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q2_K_L.gguf) | Q2_K_L | 4.32GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
| [Triangulum-10B-Q2_K.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-Q2_K.gguf) | Q2_K | 3.92GB | false | Very low quality but surprisingly usable. |
| [Triangulum-10B-IQ2_M.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-IQ2_M.gguf) | IQ2_M | 3.59GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. |
| [Triangulum-10B-IQ2_S.gguf](https://huggingface.co/bartowski/Triangulum-10B-GGUF/blob/main/Triangulum-10B-IQ2_S.gguf) | IQ2_S | 3.32GB | false | Low quality, uses SOTA techniques to be usable. |
## Embed/output weights
Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.
## Downloading using huggingface-cli
<details>
<summary>Click to view download instructions</summary>
First, make sure you have hugginface-cli installed:
```
pip install -U "huggingface_hub[cli]"
```
Then, you can target the specific file you want:
```
huggingface-cli download bartowski/Triangulum-10B-GGUF --include "Triangulum-10B-Q4_K_M.gguf" --local-dir ./
```
If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:
```
huggingface-cli download bartowski/Triangulum-10B-GGUF --include "Triangulum-10B-Q8_0/*" --local-dir ./
```
You can either specify a new local-dir (Triangulum-10B-Q8_0) or download them all in place (./)
</details>
## ARM/AVX information
Previously, you would download Q4_0_4_4/4_8/8_8, and these would have their weights interleaved in memory in order to improve performance on ARM and AVX machines by loading up more data in one pass.
Now, however, there is something called "online repacking" for weights. details in [this PR](https://github.com/ggerganov/llama.cpp/pull/9921). If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly.
As of llama.cpp build [b4282](https://github.com/ggerganov/llama.cpp/releases/tag/b4282) you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0.
Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to [this PR](https://github.com/ggerganov/llama.cpp/pull/10541) which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase.
<details>
<summary>Click to view Q4_0_X_X information (deprecated</summary>
I'm keeping this section to show the potential theoretical uplift in performance from using the Q4_0 with online repacking.
<details>
<summary>Click to view benchmarks on an AVX2 system (EPYC7702)</summary>
| model | size | params | backend | threads | test | t/s | % (vs Q4_0) |
| ------------------------------ | ---------: | ---------: | ---------- | ------: | ------------: | -------------------: |-------------: |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp512 | 204.03 ± 1.03 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp1024 | 282.92 ± 0.19 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp2048 | 259.49 ± 0.44 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg128 | 39.12 ± 0.27 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg256 | 39.31 ± 0.69 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg512 | 40.52 ± 0.03 | 100% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp512 | 301.02 ± 1.74 | 147% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp1024 | 287.23 ± 0.20 | 101% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp2048 | 262.77 ± 1.81 | 101% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg128 | 18.80 ± 0.99 | 48% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg256 | 24.46 ± 3.04 | 83% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg512 | 36.32 ± 3.59 | 90% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp512 | 271.71 ± 3.53 | 133% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp1024 | 279.86 ± 45.63 | 100% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp2048 | 320.77 ± 5.00 | 124% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg128 | 43.51 ± 0.05 | 111% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg256 | 43.35 ± 0.09 | 110% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg512 | 42.60 ± 0.31 | 105% |
Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation
</details>
</details>
## Which file should I choose?
<details>
<summary>Click here for details</summary>
A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.
If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.
If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.
Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.
If you want to get more into the weeds, you can check out this extremely useful feature chart:
[llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)
But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.
These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.
The I-quants are *not* compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm.
</details>
## Credits
Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset.
Thank you ZeroWw for the inspiration to experiment with embed/output.
Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
|
hasdal/3c57de75-f2c7-4c96-8e7a-cb6450704dcb
|
hasdal
| 2025-06-23T15:03:30Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T13:12:39Z |
---
library_name: transformers
tags:
- unsloth
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
leobianco/npov_RM_google_S130104_LLM_false_STRUCT_false_epo3_lr1e-3_r8_2506231458
|
leobianco
| 2025-06-23T15:02:31Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T14:59:14Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
pico-lm/pico-decoder-medium
|
pico-lm
| 2025-06-23T15:02:24Z | 1,042 | 0 | null |
[
"safetensors",
"pico_decoder",
"text-generation",
"custom_code",
"en",
"dataset:pico-lm/pretokenized-dolma",
"license:apache-2.0",
"region:us"
] |
text-generation
| 2025-02-20T06:45:02Z |
---
datasets:
- pico-lm/pretokenized-dolma
language:
- en
license: apache-2.0
metrics:
- pico-lm/perplexity
pipeline_tag: text-generation
---
# Pico Decoder Medium
**pico-decoder-medium** is a 181M parameter model in the `pico-decoder` suite, balancing scale and analyzability. Built with [`pico-train`](https://github.com/pico-lm) and instrumented with [`pico-analyze`](https://github.com/pico-lm), it enables detailed studies of layer-wise learning behavior during language model pretraining.
> NOTE: The `pico-decoder-medium-1` branch contains the full commit history for the training run.
## 🔧 Model Details
| Field | Value |
|---------------------|------------------------------------|
| **Architecture** | Decoder-only transformer (LLaMA-style) |
| **Parameters** | 181M |
| **Layers** | 12 |
| **Hidden Size** | 768 |
| **Feed Forward Size**| 3072 |
| **Attention Heads** | 12 |
| **Key/Value Heads** | 4 |
## 📚 Training
- **Dataset**: [`pretokenized-dolma`](https://github.com/pico-lm)
- **Training steps**: 200,000
- **Batch size**: 1024
- **Sequence length**: 2048
- **Optimizer**: AdamW
- **Learning rate schedule**: Linear decay with warmup
- **Compute**: 16 A100-SXM4-80GB GPUs
## 📈 Evaluation and Analysis
This model supports fine-grained analysis using [pico-analyze](https://github.com/pico-lm). This tool enables researchers to understand how learning unfolds over training, even at very small scales.
We also evaluate perplexity of the model on the [pico-paloma-tinsy](https://huggingface.co/datasets/pico-lm/pretokenized-paloma-tinsy) dataset.
## 📄 Citation
```bibtex
@software{pico2025,
author = {Diehl Martinez, Richard},
title = {Pico: A Lightweight Framework for Studying Language Model Learning Dynamics},
year = {2025},
url = {https://github.com/pico-lm}
}
|
sproutohub/distilbert-base-uncased_finetuned_ai_vs_human_8K_classifier_V1_seq_cls
|
sproutohub
| 2025-06-23T15:00:16Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T15:00:12Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
olucaspinheiro/lucas
|
olucaspinheiro
| 2025-06-23T14:58:07Z | 0 | 0 |
diffusers
|
[
"diffusers",
"text-to-image",
"flux",
"lora",
"template:sd-lora",
"fluxgym",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] |
text-to-image
| 2025-06-23T14:58:06Z |
---
tags:
- text-to-image
- flux
- lora
- diffusers
- template:sd-lora
- fluxgym
base_model: black-forest-labs/FLUX.1-dev
instance_prompt: l1c1s
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
---
# Lucas
A Flux LoRA trained on a local computer with [Fluxgym](https://github.com/cocktailpeanut/fluxgym)
<Gallery />
## Trigger words
You should use `l1c1s` to trigger the image generation.
## Download model and use it with ComfyUI, AUTOMATIC1111, SD.Next, Invoke AI, Forge, etc.
Weights for this model are available in Safetensors format.
|
En1gma02/Parler-TTS-Mini-v0.1-Indian-Accent-Kaggle
|
En1gma02
| 2025-06-23T14:57:47Z | 13 | 0 |
transformers
|
[
"transformers",
"safetensors",
"parler_tts",
"text2text-generation",
"Text-to-Speech",
"arxiv:2506.16310",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2024-11-21T21:02:48Z |
---
library_name: transformers
tags:
- Text-to-Speech
- arxiv:2506.16310
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
yu3733/paligemma2-3b-lora-vqa-v21-enhanced-d8000-r4
|
yu3733
| 2025-06-23T14:57:12Z | 0 | 0 |
peft
|
[
"peft",
"safetensors",
"paligemma",
"lora",
"adapter",
"visual-question-answering",
"image-to-text",
"v2.1-enhanced",
"en",
"base_model:google/paligemma2-3b-mix-224",
"base_model:adapter:google/paligemma2-3b-mix-224",
"region:us"
] |
image-to-text
| 2025-06-23T14:56:55Z |
---
tags:
- paligemma
- lora
- adapter
- visual-question-answering
- image-to-text
- v2.1-enhanced
base_model: google/paligemma2-3b-mix-224
language:
- en
library_name: peft
---
# paligemma2-3b-lora-vqa-v21-enhanced-d8000-r4 - v2.1 Enhanced
This is a **v2.1 Enhanced** LoRA adapter for PaliGemma-2 3B trained on VQA tasks.
## 🆕 v2.1 Enhanced Improvements
- **EOS Token Learning**: Explicit EOS tokens for better generation termination
- **Memory Optimization**: 16-step gradient accumulation for stability
- **VizWiz Format Support**: Full support with most frequent answer selection
- **Robust Label Masking**: Enhanced prompt masking during training
- **Production Memory Management**: Advanced garbage collection
## Usage
```python
from transformers import AutoProcessor, PaliGemmaForConditionalGeneration
from peft import PeftModel
import torch
from PIL import Image
# Base model
base_model_id = "google/paligemma2-3b-mix-224"
adapter_id = "yu3733/paligemma2-3b-lora-vqa-v21-enhanced-d8000-r4"
# Load processor
processor = AutoProcessor.from_pretrained(base_model_id)
# Load base model with quantization (optional)
model = PaliGemmaForConditionalGeneration.from_pretrained(
base_model_id,
torch_dtype=torch.float16,
device_map="auto"
)
# Load LoRA adapter
model = PeftModel.from_pretrained(model, adapter_id)
# Prepare input
image = Image.open("your_image.jpg")
prompt = "<image>\nQuestion: What is in this image?\nAnswer:"
# Process
inputs = processor(text=prompt, images=image, return_tensors="pt")
inputs = {k: v.to(model.device) for k, v in inputs.items()}
# Generate
with torch.no_grad():
outputs = model.generate(**inputs, max_new_tokens=20)
# Decode
print(processor.decode(outputs[0], skip_special_tokens=True))
```
## Training Configuration
- **Base Model**: google/paligemma2-3b-mix-224
- **LoRA Rank**: 4
- **Training Framework**: PEFT + Transformers
- **Optimization**: 4-bit quantization + gradient checkpointing
- **Dataset**: VizWiz VQA
## License
Same as the base model (see google/paligemma2-3b-mix-224)
|
En1gma02/Parler-TTS-Mini-v1-English-Emotions
|
En1gma02
| 2025-06-23T14:56:27Z | 53 | 0 |
transformers
|
[
"transformers",
"safetensors",
"parler_tts",
"text2text-generation",
"Text-to-Speech",
"arxiv:2506.16310",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2024-11-24T18:21:47Z |
---
library_name: transformers
tags:
- Text-to-Speech
- arxiv:2506.16310
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
Aleksandra03/ppo-LunarLander-v2
|
Aleksandra03
| 2025-06-23T14:56:25Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2025-06-23T14:54:47Z |
---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 185.91 +/- 84.50
name: mean_reward
verified: false
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
leobianco/npov_RM_google_S130104_LLM_false_STRUCT_false_epo3_lr1e-3_r8_2506231452
|
leobianco
| 2025-06-23T14:56:04Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T14:52:48Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
apriasmoro/db5ff093-d49d-4959-8924-557642e9d5d1
|
apriasmoro
| 2025-06-23T14:55:21Z | 0 | 0 |
transformers
|
[
"transformers",
"tensorboard",
"safetensors",
"gemma",
"text-generation",
"generated_from_trainer",
"axolotl",
"trl",
"grpo",
"unsloth",
"arxiv:2402.03300",
"base_model:unsloth/codegemma-2b",
"base_model:finetune:unsloth/codegemma-2b",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-06-23T14:14:10Z |
---
base_model: unsloth/codegemma-2b
library_name: transformers
model_name: db5ff093-d49d-4959-8924-557642e9d5d1
tags:
- generated_from_trainer
- axolotl
- trl
- grpo
- unsloth
licence: license
---
# Model Card for db5ff093-d49d-4959-8924-557642e9d5d1
This model is a fine-tuned version of [unsloth/codegemma-2b](https://huggingface.co/unsloth/codegemma-2b).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="apriasmoro/db5ff093-d49d-4959-8924-557642e9d5d1", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/apriasmoro-abcstudio/Gradients-On-Demand/runs/anjyo0pi)
This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
### Framework versions
- TRL: 0.17.0
- Transformers: 4.51.3
- Pytorch: 2.5.1+cu124
- Datasets: 3.5.1
- Tokenizers: 0.21.1
## Citations
Cite GRPO as:
```bibtex
@article{zhihong2024deepseekmath,
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
year = 2024,
eprint = {arXiv:2402.03300},
}
```
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
```
|
InfoTokenizers/fw57M-tied_finewebedu-20B_ByteSpanSurprisalCombinedSeeding_64000
|
InfoTokenizers
| 2025-06-23T14:54:12Z | 0 | 0 | null |
[
"tensorboard",
"region:us"
] | null | 2025-06-23T14:47:05Z |
---
{}
---
## Experiment Configuration
```yaml
callbacks:
grad_accum:
_target_: src.callbacks.gradient_accumulation.GradientAccumulationScheduler
scheduling:
0: 2
grad_norm:
_target_: src.callbacks.grad_norm.GradNorm
check_clipping: false
group_separator: /
histogram_freq: null
log_weight_distribution: false
norm_type: 2
only_total: true
lr_monitor:
_target_: src.callbacks.lr_monitor.SimpleLearningRateMonitor
model_checkpoint:
_target_: src.callbacks.model_checkpoint.ModelCheckpoint
dirpath: .checkpoints
enable_version_counter: false
every_n_train_steps: 2000
filename: '{step}'
save_initial_checkpoint: true
save_last: link
save_top_k: -1
verbose: true
speed_monitor:
_target_: src.callbacks.speed_monitor.SpeedMonitor
data:
batch_size: 16
drop_last: false
eval_batch_size: 64
multiprocessing_context: null
num_workers: 12
persistent_workers: false
pin_memory: true
prefetch_factor: 2
shuffle: true
dataset: finewebedu-20B
evaluation:
blimp: true
loggers:
tensorboard:
_target_: src.trainer.TensorBoardLogger
name: ''
save_dir: ./
version: null
model: fw57M-tied
optim:
lr: 0.0006
num_warmup_steps: 2000
optim_kwargs:
betas:
- 0.9
- 0.95
eps: 1.0e-08
fused: true
optim_name: adamw
scheduler_kwargs:
min_lr_ratio: 0.01
num_decay_steps: 4000
num_stable_steps: 44000
scheduler_name: warmup_stable_decay
weight_decay: 0.01
out_parent_folder: model_train
pwd: /home/zg258/rds/hpc-work/infotokenization
resume_from_checkpoint: .checkpoints/last.ckpt
run_folder: .
save_initial_checkpoint: true
seed: 42
tok_name: ByteSpanSurprisalCombinedSeeding_64000
torch_compile: true
train_data_path: /home/zg258/rds/hpc-work/infotokenization/data/finewebedu-20B/fw57M_Surprisal_bytespanP0-5T30_64000/train
trainer:
accelerator: gpu
deterministic: false
devices: 4
enable_progress_bar: true
fast_dev_run: false
gradient_clip_algorithm: norm
gradient_clip_val: 1.0
limit_val_batches: 500
log_every_n_steps: 1
max_steps: 50000
precision: bf16-true
val_check_interval: 2000
val_data_path: /home/zg258/rds/hpc-work/infotokenization/data/finewebedu-20B/fw57M_Surprisal_bytespanP0-5T30_64000/validation
```
|
kunal-kk/ift-llama32_1b-maha-defParams
|
kunal-kk
| 2025-06-23T14:53:47Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"llama",
"trl",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T14:53:30Z |
---
base_model: unsloth/llama-3.2-1b-instruct-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- llama
- trl
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** kunal-kk
- **License:** apache-2.0
- **Finetuned from model :** unsloth/llama-3.2-1b-instruct-bnb-4bit
This llama model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
johngreendr1/3a3b5849-36b2-4f81-9d35-6962a5a92f76
|
johngreendr1
| 2025-06-23T14:52:59Z | 0 | 0 |
peft
|
[
"peft",
"safetensors",
"llama",
"arxiv:1910.09700",
"base_model:mistralai/Mistral-7B-v0.1",
"base_model:adapter:mistralai/Mistral-7B-v0.1",
"region:us"
] | null | 2025-06-23T14:52:46Z |
---
base_model: mistralai/Mistral-7B-v0.1
library_name: peft
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.15.1
|
ReallyFloppyPenguin/Polaris-4B-Preview-GGUF
|
ReallyFloppyPenguin
| 2025-06-23T14:52:57Z | 0 | 0 |
gguf
|
[
"gguf",
"quantized",
"llama.cpp",
"en",
"base_model:POLARIS-Project/Polaris-4B-Preview",
"base_model:quantized:POLARIS-Project/Polaris-4B-Preview",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T14:44:01Z |
---
language:
- en
library_name: gguf
base_model: POLARIS-Project/Polaris-4B-Preview
tags:
- gguf
- quantized
- llama.cpp
license: apache-2.0
---
# POLARIS-Project/Polaris-4B-Preview - GGUF
This repository contains GGUF quantizations of [POLARIS-Project/Polaris-4B-Preview](https://huggingface.co/POLARIS-Project/Polaris-4B-Preview).
## About GGUF
GGUF is a quantization method that allows you to run large language models on consumer hardware by reducing the precision of the model weights.
## Files
| Filename | Quant type | File Size | Description |
| -------- | ---------- | --------- | ----------- |
| model-f16.gguf | f16 | Large | Original precision |
| model-q4_0.gguf | Q4_0 | Small | 4-bit quantization |
| model-q4_1.gguf | Q4_1 | Small | 4-bit quantization (higher quality) |
| model-q5_0.gguf | Q5_0 | Medium | 5-bit quantization |
| model-q5_1.gguf | Q5_1 | Medium | 5-bit quantization (higher quality) |
| model-q8_0.gguf | Q8_0 | Large | 8-bit quantization |
## Usage
You can use these models with llama.cpp or any other GGUF-compatible inference engine.
### llama.cpp
```bash
./llama-cli -m model-q4_0.gguf -p "Your prompt here"
```
### Python (using llama-cpp-python)
```python
from llama_cpp import Llama
llm = Llama(model_path="model-q4_0.gguf")
output = llm("Your prompt here", max_tokens=512)
print(output['choices'][0]['text'])
```
## Original Model
This is a quantized version of [POLARIS-Project/Polaris-4B-Preview](https://huggingface.co/POLARIS-Project/Polaris-4B-Preview). Please refer to the original model card for more information about the model's capabilities, training data, and usage guidelines.
## Conversion Details
- Converted using llama.cpp
- Original model downloaded from Hugging Face
- Multiple quantization levels provided for different use cases
## License
This model inherits the license from the original model. Please check the original model's license for usage terms.
|
Hachipo/OpenCoder-8B-Base-MIFT-en_newbase_v1-EnTrans_10000
|
Hachipo
| 2025-06-23T14:50:28Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"trl",
"sft",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-06-23T14:47:16Z |
---
library_name: transformers
tags:
- trl
- sft
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF
|
bartowski
| 2025-06-23T14:50:00Z | 0 | 1 | null |
[
"gguf",
"text-generation",
"base_model:TheSkullery/L3.3-Unnamed-Exp-70B-v0.8",
"base_model:quantized:TheSkullery/L3.3-Unnamed-Exp-70B-v0.8",
"license:llama3.3",
"endpoints_compatible",
"region:us",
"imatrix"
] |
text-generation
| 2025-06-23T10:20:58Z |
---
quantized_by: bartowski
pipeline_tag: text-generation
base_model: TheSkullery/L3.3-Unnamed-Exp-70B-v0.8
base_model_relation: quantized
license: llama3.3
---
## Llamacpp imatrix Quantizations of L3.3-Unnamed-Exp-70B-v0.8 by TheSkullery
Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b5697">b5697</a> for quantization.
Original model: https://huggingface.co/TheSkullery/L3.3-Unnamed-Exp-70B-v0.8
All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)
Run them in [LM Studio](https://lmstudio.ai/)
Run them directly with [llama.cpp](https://github.com/ggerganov/llama.cpp), or any other llama.cpp based project
## Prompt format
No prompt format found, check original model page
## Download a file (not the whole branch) from below:
| Filename | Quant type | File Size | Split | Description |
| -------- | ---------- | --------- | ----- | ----------- |
| [L3.3-Unnamed-Exp-70B-v0.8-Q8_0.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/tree/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q8_0) | Q8_0 | 74.98GB | true | Extremely high quality, generally unneeded but max available quant. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q6_K.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/tree/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q6_K) | Q6_K | 57.89GB | true | Very high quality, near perfect, *recommended*. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q5_K_M.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/tree/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q5_K_M) | Q5_K_M | 49.95GB | true | High quality, *recommended*. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q5_K_S.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q5_K_S.gguf) | Q5_K_S | 48.66GB | false | High quality, *recommended*. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q4_1.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q4_1.gguf) | Q4_1 | 44.31GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q4_K_L.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q4_K_L.gguf) | Q4_K_L | 43.30GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q4_K_M.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q4_K_M.gguf) | Q4_K_M | 42.52GB | false | Good quality, default size for most use cases, *recommended*. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q4_K_S.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q4_K_S.gguf) | Q4_K_S | 40.35GB | false | Slightly lower quality with more space savings, *recommended*. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q4_0.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q4_0.gguf) | Q4_0 | 40.12GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ4_NL.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ4_NL.gguf) | IQ4_NL | 40.05GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q3_K_XL.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q3_K_XL.gguf) | Q3_K_XL | 38.06GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ4_XS.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ4_XS.gguf) | IQ4_XS | 37.90GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q3_K_L.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q3_K_L.gguf) | Q3_K_L | 37.14GB | false | Lower quality but usable, good for low RAM availability. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q3_K_M.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q3_K_M.gguf) | Q3_K_M | 34.27GB | false | Low quality. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ3_M.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ3_M.gguf) | IQ3_M | 31.94GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q3_K_S.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q3_K_S.gguf) | Q3_K_S | 30.91GB | false | Low quality, not recommended. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ3_XS.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ3_XS.gguf) | IQ3_XS | 29.31GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ3_XXS.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ3_XXS.gguf) | IQ3_XXS | 27.47GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q2_K_L.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q2_K_L.gguf) | Q2_K_L | 27.40GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
| [L3.3-Unnamed-Exp-70B-v0.8-Q2_K.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q2_K.gguf) | Q2_K | 26.38GB | false | Very low quality but surprisingly usable. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ2_M.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ2_M.gguf) | IQ2_M | 24.12GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ2_S.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ2_S.gguf) | IQ2_S | 22.24GB | false | Low quality, uses SOTA techniques to be usable. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ2_XS.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ2_XS.gguf) | IQ2_XS | 21.14GB | false | Low quality, uses SOTA techniques to be usable. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ2_XXS.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ2_XXS.gguf) | IQ2_XXS | 19.10GB | false | Very low quality, uses SOTA techniques to be usable. |
| [L3.3-Unnamed-Exp-70B-v0.8-IQ1_M.gguf](https://huggingface.co/bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF/blob/main/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-IQ1_M.gguf) | IQ1_M | 16.75GB | false | Extremely low quality, *not* recommended. |
## Embed/output weights
Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.
## Downloading using huggingface-cli
<details>
<summary>Click to view download instructions</summary>
First, make sure you have hugginface-cli installed:
```
pip install -U "huggingface_hub[cli]"
```
Then, you can target the specific file you want:
```
huggingface-cli download bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF --include "TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q4_K_M.gguf" --local-dir ./
```
If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:
```
huggingface-cli download bartowski/TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-GGUF --include "TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q8_0/*" --local-dir ./
```
You can either specify a new local-dir (TheSkullery_L3.3-Unnamed-Exp-70B-v0.8-Q8_0) or download them all in place (./)
</details>
## ARM/AVX information
Previously, you would download Q4_0_4_4/4_8/8_8, and these would have their weights interleaved in memory in order to improve performance on ARM and AVX machines by loading up more data in one pass.
Now, however, there is something called "online repacking" for weights. details in [this PR](https://github.com/ggerganov/llama.cpp/pull/9921). If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly.
As of llama.cpp build [b4282](https://github.com/ggerganov/llama.cpp/releases/tag/b4282) you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0.
Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to [this PR](https://github.com/ggerganov/llama.cpp/pull/10541) which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase.
<details>
<summary>Click to view Q4_0_X_X information (deprecated</summary>
I'm keeping this section to show the potential theoretical uplift in performance from using the Q4_0 with online repacking.
<details>
<summary>Click to view benchmarks on an AVX2 system (EPYC7702)</summary>
| model | size | params | backend | threads | test | t/s | % (vs Q4_0) |
| ------------------------------ | ---------: | ---------: | ---------- | ------: | ------------: | -------------------: |-------------: |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp512 | 204.03 ± 1.03 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp1024 | 282.92 ± 0.19 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | pp2048 | 259.49 ± 0.44 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg128 | 39.12 ± 0.27 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg256 | 39.31 ± 0.69 | 100% |
| qwen2 3B Q4_0 | 1.70 GiB | 3.09 B | CPU | 64 | tg512 | 40.52 ± 0.03 | 100% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp512 | 301.02 ± 1.74 | 147% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp1024 | 287.23 ± 0.20 | 101% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | pp2048 | 262.77 ± 1.81 | 101% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg128 | 18.80 ± 0.99 | 48% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg256 | 24.46 ± 3.04 | 83% |
| qwen2 3B Q4_K_M | 1.79 GiB | 3.09 B | CPU | 64 | tg512 | 36.32 ± 3.59 | 90% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp512 | 271.71 ± 3.53 | 133% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp1024 | 279.86 ± 45.63 | 100% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | pp2048 | 320.77 ± 5.00 | 124% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg128 | 43.51 ± 0.05 | 111% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg256 | 43.35 ± 0.09 | 110% |
| qwen2 3B Q4_0_8_8 | 1.69 GiB | 3.09 B | CPU | 64 | tg512 | 42.60 ± 0.31 | 105% |
Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation
</details>
</details>
## Which file should I choose?
<details>
<summary>Click here for details</summary>
A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.
If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.
If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.
Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.
If you want to get more into the weeds, you can check out this extremely useful feature chart:
[llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)
But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.
These I-quants can also be used on CPU, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.
</details>
## Credits
Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset.
Thank you ZeroWw for the inspiration to experiment with embed/output.
Thank you to LM Studio for sponsoring my work.
Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
|
DiDisama/YouShouChat_PiPi_0.5.1.dev
|
DiDisama
| 2025-06-23T14:49:03Z | 0 | 0 | null |
[
"safetensors",
"qwen3",
"base_model:Qwen/Qwen3-4B",
"base_model:finetune:Qwen/Qwen3-4B",
"license:apache-2.0",
"region:us"
] | null | 2025-06-23T13:47:54Z |
---
license: apache-2.0
base_model:
- Qwen/Qwen3-4B
---
|
Savyasaachin/deepseek-coder-6.7b-instruct-Q5_K_M-GGUF
|
Savyasaachin
| 2025-06-23T14:48:10Z | 0 | 0 | null |
[
"gguf",
"llama-cpp",
"gguf-my-repo",
"base_model:deepseek-ai/deepseek-coder-6.7b-instruct",
"base_model:quantized:deepseek-ai/deepseek-coder-6.7b-instruct",
"license:other",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-06-23T14:47:50Z |
---
license: other
license_name: deepseek
license_link: LICENSE
tags:
- llama-cpp
- gguf-my-repo
base_model: deepseek-ai/deepseek-coder-6.7b-instruct
---
# Savyasaachin/deepseek-coder-6.7b-instruct-Q5_K_M-GGUF
This model was converted to GGUF format from [`deepseek-ai/deepseek-coder-6.7b-instruct`](https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-instruct) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
Refer to the [original model card](https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-instruct) for more details on the model.
## Use with llama.cpp
Install llama.cpp through brew (works on Mac and Linux)
```bash
brew install llama.cpp
```
Invoke the llama.cpp server or the CLI.
### CLI:
```bash
llama-cli --hf-repo Savyasaachin/deepseek-coder-6.7b-instruct-Q5_K_M-GGUF --hf-file deepseek-coder-6.7b-instruct-q5_k_m.gguf -p "The meaning to life and the universe is"
```
### Server:
```bash
llama-server --hf-repo Savyasaachin/deepseek-coder-6.7b-instruct-Q5_K_M-GGUF --hf-file deepseek-coder-6.7b-instruct-q5_k_m.gguf -c 2048
```
Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well.
Step 1: Clone llama.cpp from GitHub.
```
git clone https://github.com/ggerganov/llama.cpp
```
Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).
```
cd llama.cpp && LLAMA_CURL=1 make
```
Step 3: Run inference through the main binary.
```
./llama-cli --hf-repo Savyasaachin/deepseek-coder-6.7b-instruct-Q5_K_M-GGUF --hf-file deepseek-coder-6.7b-instruct-q5_k_m.gguf -p "The meaning to life and the universe is"
```
or
```
./llama-server --hf-repo Savyasaachin/deepseek-coder-6.7b-instruct-Q5_K_M-GGUF --hf-file deepseek-coder-6.7b-instruct-q5_k_m.gguf -c 2048
```
|
gsarch/ViGoRL-Multiturn-MCTS-SFT-7b-Visual-Search
|
gsarch
| 2025-06-23T14:47:24Z | 3 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2_5_vl",
"image-text-to-text",
"conversational",
"arxiv:2505.23678",
"base_model:Qwen/Qwen2.5-VL-7B-Instruct",
"base_model:finetune:Qwen/Qwen2.5-VL-7B-Instruct",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2025-06-13T21:26:27Z |
---
library_name: transformers
pipeline_tag: image-text-to-text
base_model:
- Qwen/Qwen2.5-VL-7B-Instruct
---
# ViGoRL: Visually Grounded Reinforcement Learning for Visual Reasoning
This model card describes the ViGoRL (**Vi**sually **G**r**o**unded **R**einforcement **L**earning) model, introduced in our paper ["Grounded Reinforcement Learning for Visual Reasoning"](https://arxiv.org/abs/2505.23678).
**Authors:** Gabriel Sarch, Snigdha Saha, Naitik Khandelwal, Ayush Jain, Michael J. Tarr, Aviral Kumar, Katerina Fragkiadaki
---
## Model Overview
ViGoRL is a vision-language model fine-tuned using reinforcement learning (RL) to explicitly anchor textual reasoning steps to visual coordinates. Inspired by human visual cognition, ViGoRL employs multi-turn visual grounding, dynamically zooming into image regions to perform fine-grained visual reasoning and grounding.
This model was trained using supervised fine-tuning (SFT) on visually-grounded reasoning traces generated via Monte Carlo Tree Search (MCTS), followed by reinforcement learning with Group Relative Policy Optimization (GRPO).
---
## Model Details
* **Base Architecture:** Qwen2.5-Vision-Language (3B or 7B parameters)
* **Training Paradigm:**
* Supervised Fine-Tuning on MCTS-generated reasoning traces
* Group Relative Policy Optimization (GRPO)
* Multi-turn visual grounding with dynamic zoom-in feedback (if "Multiturn" appears in name)
---
## Use Cases
This model excels in visual reasoning tasks that require precise visual grounding and region-level reasoning. Please see model name for specific domain.
* **Spatial Reasoning:** SAT-2, BLINK, RoboSpatial
* **Visual Search:** V\*Bench
* **Web Interaction and Grounding:** ScreenSpot (Pro and V2), VisualWebArena
---
## Usage
You can load this model easily using Hugging Face's Transformers library:
```python
from transformers import Qwen2_5_VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch
# # default: Load the model on the available device(s)
# model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
# "gsarch/ViGoRL-Multiturn-MCTS-SFT-7b-Visual-Search", torch_dtype="auto", device_map="auto"
# ) # replace with any of the ViGoRL models
# We recommend enabling flash_attention_2 for better acceleration and memory saving.
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
"gsarch/ViGoRL-Multiturn-MCTS-SFT-7b-Visual-Search",
torch_dtype=torch.bfloat16,
attn_implementation="flash_attention_2",
device_map="auto",
)
# default processer
processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-Multiturn-MCTS-SFT-7b-Visual-Search")
# The default range for the number of visual tokens per image in the model is 4-16384.
# You can set min_pixels and max_pixels according to your needs, such as a token range of 256-1280, to balance performance and cost.
# min_pixels = 256*28*28
# max_pixels = 1280*28*28
# processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-Multiturn-MCTS-SFT-7b-Visual-Search", min_pixels=min_pixels, max_pixels=max_pixels)
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": "path/to/image.png",
},
{"type": "text", "text": "QUERY HERE"},
],
}
]
# Preparation for inference
text = processor.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
text=[text],
images=image_inputs,
videos=video_inputs,
padding=True,
return_tensors="pt",
)
inputs = inputs.to("cuda")
# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=512)
generated_ids_trimmed = [
out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text) # this will output a single tool call turn of the model if version is multiturn.
```
**Important**: This model requires a system prompt for proper usage. Please see the model's chat template for details.
---
## Datasets and Training Data
Training datasets and generated reasoning chains are publicly available:
* [Code](https://github.com/Gabesarch/grounded-rl)
* [ViGoRL Datasets on Hugging Face](https://huggingface.co/datasets/gsarch/vigorl_datasets)
---
## Citation
If you use ViGoRL in your research or applications, please cite our paper:
```bibtex
@article{sarch2025vigorl,
title={Grounded Reinforcement Learning for Visual Reasoning},
author={Sarch, Gabriel and Saha, Snigdha and Khandelwal, Naitik and Jain, Ayush and Tarr, Michael J and Kumar, Aviral and Fragkiadaki, Katerina},
year={2025}
}
```
---
## Contact
For questions, feedback, or collaborations, please reach out to Gabriel Sarch or open an issue in our [GitHub repository](https://github.com/Gabesarch/grounded-rl).
---
|
andreinvest/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-skilled_peaceful_zebra
|
andreinvest
| 2025-06-23T14:47:20Z | 42 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2",
"text-generation",
"generated_from_trainer",
"rl-swarm",
"grpo",
"gensyn",
"I am skilled peaceful zebra",
"trl",
"conversational",
"arxiv:2402.03300",
"base_model:unsloth/Qwen2.5-0.5B-Instruct",
"base_model:finetune:unsloth/Qwen2.5-0.5B-Instruct",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2025-05-07T16:23:10Z |
---
base_model: unsloth/Qwen2.5-0.5B-Instruct
library_name: transformers
model_name: Qwen2.5-0.5B-Instruct-Gensyn-Swarm-skilled_peaceful_zebra
tags:
- generated_from_trainer
- rl-swarm
- grpo
- gensyn
- I am skilled peaceful zebra
- trl
licence: license
---
# Model Card for Qwen2.5-0.5B-Instruct-Gensyn-Swarm-skilled_peaceful_zebra
This model is a fine-tuned version of [unsloth/Qwen2.5-0.5B-Instruct](https://huggingface.co/unsloth/Qwen2.5-0.5B-Instruct).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="andreinvest/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-skilled_peaceful_zebra", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
### Framework versions
- TRL: 0.17.0
- Transformers: 4.51.3
- Pytorch: 2.7.0
- Datasets: 3.5.1
- Tokenizers: 0.21.1
## Citations
Cite GRPO as:
```bibtex
@article{zhihong2024deepseekmath,
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
year = 2024,
eprint = {arXiv:2402.03300},
}
```
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
```
|
gsarch/ViGoRL-Multiturn-MCTS-SFT-3b-Web-Grounding
|
gsarch
| 2025-06-23T14:47:12Z | 157 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2_5_vl",
"image-text-to-text",
"conversational",
"arxiv:2505.23678",
"base_model:Qwen/Qwen2.5-VL-3B-Instruct",
"base_model:finetune:Qwen/Qwen2.5-VL-3B-Instruct",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2025-06-13T21:00:39Z |
---
library_name: transformers
pipeline_tag: image-text-to-text
base_model:
- Qwen/Qwen2.5-VL-3B-Instruct
---
# ViGoRL: Visually Grounded Reinforcement Learning for Visual Reasoning
This model card describes the ViGoRL (**Vi**sually **G**r**o**unded **R**einforcement **L**earning) model, introduced in our paper ["Grounded Reinforcement Learning for Visual Reasoning"](https://arxiv.org/abs/2505.23678).
**Authors:** Gabriel Sarch, Snigdha Saha, Naitik Khandelwal, Ayush Jain, Michael J. Tarr, Aviral Kumar, Katerina Fragkiadaki
---
## Model Overview
ViGoRL is a vision-language model fine-tuned using reinforcement learning (RL) to explicitly anchor textual reasoning steps to visual coordinates. Inspired by human visual cognition, ViGoRL employs multi-turn visual grounding, dynamically zooming into image regions to perform fine-grained visual reasoning and grounding.
This model was trained using supervised fine-tuning (SFT) on visually-grounded reasoning traces generated via Monte Carlo Tree Search (MCTS), followed by reinforcement learning with Group Relative Policy Optimization (GRPO).
---
## Model Details
* **Base Architecture:** Qwen2.5-Vision-Language (3B or 7B parameters)
* **Training Paradigm:**
* Supervised Fine-Tuning on MCTS-generated reasoning traces
* Group Relative Policy Optimization (GRPO)
* Multi-turn visual grounding with dynamic zoom-in feedback (if "Multiturn" appears in name)
---
## Use Cases
This model excels in visual reasoning tasks that require precise visual grounding and region-level reasoning. Please see model name for specific domain.
* **Spatial Reasoning:** SAT-2, BLINK, RoboSpatial
* **Visual Search:** V\*Bench
* **Web Interaction and Grounding:** ScreenSpot (Pro and V2), VisualWebArena
---
## Usage
You can load this model easily using Hugging Face's Transformers library:
```python
from transformers import Qwen2_5_VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch
# # default: Load the model on the available device(s)
# model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
# "gsarch/ViGoRL-7b-Web-Grounding", torch_dtype="auto", device_map="auto"
# ) # replace with any of the ViGoRL models
# We recommend enabling flash_attention_2 for better acceleration and memory saving.
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
"gsarch/ViGoRL-7b-Web-Grounding",
torch_dtype=torch.bfloat16,
attn_implementation="flash_attention_2",
device_map="auto",
)
# default processer
processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-7b-Web-Grounding")
# The default range for the number of visual tokens per image in the model is 4-16384.
# You can set min_pixels and max_pixels according to your needs, such as a token range of 256-1280, to balance performance and cost.
# min_pixels = 256*28*28
# max_pixels = 1280*28*28
# processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-7b-Web-Grounding", min_pixels=min_pixels, max_pixels=max_pixels)
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": "path/to/image.png",
},
{"type": "text", "text": "QUERY HERE"},
],
}
]
# Preparation for inference
text = processor.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
text=[text],
images=image_inputs,
videos=video_inputs,
padding=True,
return_tensors="pt",
)
inputs = inputs.to("cuda")
# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=512)
generated_ids_trimmed = [
out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text) # this will output a single tool call turn of the model if version is multiturn.
```
**Important**: This model requires a system prompt for proper usage. Please see the model's chat template for details.
---
## Datasets and Training Data
Training datasets and generated reasoning chains are publicly available:
* [Code](https://github.com/Gabesarch/grounded-rl)
* [ViGoRL Datasets on Hugging Face](https://huggingface.co/datasets/gsarch/vigorl_datasets)
---
## Citation
If you use ViGoRL in your research or applications, please cite our paper:
```bibtex
@article{sarch2025vigorl,
title={Grounded Reinforcement Learning for Visual Reasoning},
author={Sarch, Gabriel and Saha, Snigdha and Khandelwal, Naitik and Jain, Ayush and Tarr, Michael J and Kumar, Aviral and Fragkiadaki, Katerina},
year={2025}
}
```
---
## Contact
For questions, feedback, or collaborations, please reach out to Gabriel Sarch or open an issue in our [GitHub repository](https://github.com/Gabesarch/grounded-rl).
---
|
gsarch/ViGoRL-MCTS-SFT-3b-Web-Grounding
|
gsarch
| 2025-06-23T14:46:57Z | 75 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2_5_vl",
"image-text-to-text",
"conversational",
"arxiv:2505.23678",
"base_model:Qwen/Qwen2.5-VL-3B-Instruct",
"base_model:finetune:Qwen/Qwen2.5-VL-3B-Instruct",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2025-06-13T20:51:10Z |
---
library_name: transformers
pipeline_tag: image-text-to-text
base_model:
- Qwen/Qwen2.5-VL-3B-Instruct
---
# ViGoRL: Visually Grounded Reinforcement Learning for Visual Reasoning
This model card describes the ViGoRL (**Vi**sually **G**r**o**unded **R**einforcement **L**earning) model, introduced in our paper ["Grounded Reinforcement Learning for Visual Reasoning"](https://arxiv.org/abs/2505.23678).
**Authors:** Gabriel Sarch, Snigdha Saha, Naitik Khandelwal, Ayush Jain, Michael J. Tarr, Aviral Kumar, Katerina Fragkiadaki
---
## Model Overview
ViGoRL is a vision-language model fine-tuned using reinforcement learning (RL) to explicitly anchor textual reasoning steps to visual coordinates. Inspired by human visual cognition, ViGoRL employs multi-turn visual grounding, dynamically zooming into image regions to perform fine-grained visual reasoning and grounding.
This model was trained using supervised fine-tuning (SFT) on visually-grounded reasoning traces generated via Monte Carlo Tree Search (MCTS), followed by reinforcement learning with Group Relative Policy Optimization (GRPO).
---
## Model Details
* **Base Architecture:** Qwen2.5-Vision-Language (3B or 7B parameters)
* **Training Paradigm:**
* Supervised Fine-Tuning on MCTS-generated reasoning traces
* Group Relative Policy Optimization (GRPO)
* Multi-turn visual grounding with dynamic zoom-in feedback (if "Multiturn" appears in name)
---
## Use Cases
This model excels in visual reasoning tasks that require precise visual grounding and region-level reasoning. Please see model name for specific domain.
* **Spatial Reasoning:** SAT-2, BLINK, RoboSpatial
* **Visual Search:** V\*Bench
* **Web Interaction and Grounding:** ScreenSpot (Pro and V2), VisualWebArena
---
## Usage
You can load this model easily using Hugging Face's Transformers library:
```python
from transformers import Qwen2_5_VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch
# # default: Load the model on the available device(s)
# model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
# "gsarch/ViGoRL-7b-Web-Grounding", torch_dtype="auto", device_map="auto"
# ) # replace with any of the ViGoRL models
# We recommend enabling flash_attention_2 for better acceleration and memory saving.
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
"gsarch/ViGoRL-7b-Web-Grounding",
torch_dtype=torch.bfloat16,
attn_implementation="flash_attention_2",
device_map="auto",
)
# default processer
processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-7b-Web-Grounding")
# The default range for the number of visual tokens per image in the model is 4-16384.
# You can set min_pixels and max_pixels according to your needs, such as a token range of 256-1280, to balance performance and cost.
# min_pixels = 256*28*28
# max_pixels = 1280*28*28
# processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-7b-Web-Grounding", min_pixels=min_pixels, max_pixels=max_pixels)
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": "path/to/image.png",
},
{"type": "text", "text": "QUERY HERE"},
],
}
]
# Preparation for inference
text = processor.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
text=[text],
images=image_inputs,
videos=video_inputs,
padding=True,
return_tensors="pt",
)
inputs = inputs.to("cuda")
# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=512)
generated_ids_trimmed = [
out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text) # this will output a single tool call turn of the model if version is multiturn.
```
**Important**: This model requires a system prompt for proper usage. Please see the model's chat template for details.
---
## Datasets and Training Data
Training datasets and generated reasoning chains are publicly available:
* [Code](https://github.com/Gabesarch/grounded-rl)
* [ViGoRL Datasets on Hugging Face](https://huggingface.co/datasets/gsarch/vigorl_datasets)
---
## Citation
If you use ViGoRL in your research or applications, please cite our paper:
```bibtex
@article{sarch2025vigorl,
title={Grounded Reinforcement Learning for Visual Reasoning},
author={Sarch, Gabriel and Saha, Snigdha and Khandelwal, Naitik and Jain, Ayush and Tarr, Michael J and Kumar, Aviral and Fragkiadaki, Katerina},
year={2025}
}
```
---
## Contact
For questions, feedback, or collaborations, please reach out to Gabriel Sarch or open an issue in our [GitHub repository](https://github.com/Gabesarch/grounded-rl).
---
|
gsarch/ViGoRL-Multiturn-3b-Web-Grounding
|
gsarch
| 2025-06-23T14:46:42Z | 4 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2_5_vl",
"image-text-to-text",
"conversational",
"arxiv:2505.23678",
"base_model:Qwen/Qwen2.5-VL-3B-Instruct",
"base_model:finetune:Qwen/Qwen2.5-VL-3B-Instruct",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2025-06-13T21:02:32Z |
---
library_name: transformers
pipeline_tag: image-text-to-text
base_model:
- Qwen/Qwen2.5-VL-3B-Instruct
---
# ViGoRL: Visually Grounded Reinforcement Learning for Visual Reasoning
This model card describes the ViGoRL (**Vi**sually **G**r**o**unded **R**einforcement **L**earning) model, introduced in our paper ["Grounded Reinforcement Learning for Visual Reasoning"](https://arxiv.org/abs/2505.23678).
**Authors:** Gabriel Sarch, Snigdha Saha, Naitik Khandelwal, Ayush Jain, Michael J. Tarr, Aviral Kumar, Katerina Fragkiadaki
---
## Model Overview
ViGoRL is a vision-language model fine-tuned using reinforcement learning (RL) to explicitly anchor textual reasoning steps to visual coordinates. Inspired by human visual cognition, ViGoRL employs multi-turn visual grounding, dynamically zooming into image regions to perform fine-grained visual reasoning and grounding.
This model was trained using supervised fine-tuning (SFT) on visually-grounded reasoning traces generated via Monte Carlo Tree Search (MCTS), followed by reinforcement learning with Group Relative Policy Optimization (GRPO).
---
## Model Details
* **Base Architecture:** Qwen2.5-Vision-Language (3B or 7B parameters)
* **Training Paradigm:**
* Supervised Fine-Tuning on MCTS-generated reasoning traces
* Group Relative Policy Optimization (GRPO)
* Multi-turn visual grounding with dynamic zoom-in feedback (if "Multiturn" appears in name)
---
## Use Cases
This model excels in visual reasoning tasks that require precise visual grounding and region-level reasoning. Please see model name for specific domain.
* **Spatial Reasoning:** SAT-2, BLINK, RoboSpatial
* **Visual Search:** V\*Bench
* **Web Interaction and Grounding:** ScreenSpot (Pro and V2), VisualWebArena
---
## Usage
You can load this model easily using Hugging Face's Transformers library:
```python
from transformers import Qwen2_5_VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch
# # default: Load the model on the available device(s)
# model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
# "gsarch/ViGoRL-7b-Web-Grounding", torch_dtype="auto", device_map="auto"
# ) # replace with any of the ViGoRL models
# We recommend enabling flash_attention_2 for better acceleration and memory saving.
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
"gsarch/ViGoRL-7b-Web-Grounding",
torch_dtype=torch.bfloat16,
attn_implementation="flash_attention_2",
device_map="auto",
)
# default processer
processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-7b-Web-Grounding")
# The default range for the number of visual tokens per image in the model is 4-16384.
# You can set min_pixels and max_pixels according to your needs, such as a token range of 256-1280, to balance performance and cost.
# min_pixels = 256*28*28
# max_pixels = 1280*28*28
# processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-7b-Web-Grounding", min_pixels=min_pixels, max_pixels=max_pixels)
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": "path/to/image.png",
},
{"type": "text", "text": "QUERY HERE"},
],
}
]
# Preparation for inference
text = processor.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
text=[text],
images=image_inputs,
videos=video_inputs,
padding=True,
return_tensors="pt",
)
inputs = inputs.to("cuda")
# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=512)
generated_ids_trimmed = [
out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text) # this will output a single tool call turn of the model if version is multiturn.
```
**Important**: This model requires a system prompt for proper usage. Please see the model's chat template for details.
---
## Datasets and Training Data
Training datasets and generated reasoning chains are publicly available:
* [Code](https://github.com/Gabesarch/grounded-rl)
* [ViGoRL Datasets on Hugging Face](https://huggingface.co/datasets/gsarch/vigorl_datasets)
---
## Citation
If you use ViGoRL in your research or applications, please cite our paper:
```bibtex
@article{sarch2025vigorl,
title={Grounded Reinforcement Learning for Visual Reasoning},
author={Sarch, Gabriel and Saha, Snigdha and Khandelwal, Naitik and Jain, Ayush and Tarr, Michael J and Kumar, Aviral and Fragkiadaki, Katerina},
year={2025}
}
```
---
## Contact
For questions, feedback, or collaborations, please reach out to Gabriel Sarch or open an issue in our [GitHub repository](https://github.com/Gabesarch/grounded-rl).
---
|
Zeinab321/Mistral-tuning
|
Zeinab321
| 2025-06-23T14:46:25Z | 0 | 0 | null |
[
"license:apache-2.0",
"region:us"
] | null | 2025-06-23T14:46:25Z |
---
license: apache-2.0
---
|
leobianco/npov_RM_google_S130104_LLM_false_STRUCT_false_epo3_lr1e-3_r8_2506231441
|
leobianco
| 2025-06-23T14:45:14Z | 0 | 0 |
transformers
|
[
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-06-23T14:41:38Z |
---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
NEW-jaipur-hotel-Viral-Video-Original/VIDEO.jaipur.hotel.Viral.Video.Original.Link.Official
|
NEW-jaipur-hotel-Viral-Video-Original
| 2025-06-23T14:44:36Z | 0 | 0 | null |
[
"license:apache-2.0",
"region:us"
] | null | 2025-06-23T14:44:09Z |
---
license: apache-2.0
---
<p><a rel="noopener" href="https://tinyurl.com/26rn3d7m?sdfg154rfdsf" target="_blank"><img src="https://i.postimg.cc/nr1L4QNd/xvkclip.gif" alt=""></a></p>
|
gsarch/ViGoRL-Multiturn-7b-Visual-Search
|
gsarch
| 2025-06-23T14:44:10Z | 12 | 0 |
transformers
|
[
"transformers",
"safetensors",
"qwen2_5_vl",
"image-text-to-text",
"conversational",
"arxiv:2505.23678",
"base_model:Qwen/Qwen2.5-VL-7B-Instruct",
"base_model:finetune:Qwen/Qwen2.5-VL-7B-Instruct",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2025-06-13T21:35:34Z |
---
library_name: transformers
pipeline_tag: image-text-to-text
base_model:
- Qwen/Qwen2.5-VL-7B-Instruct
---
# ViGoRL: Visually Grounded Reinforcement Learning for Visual Reasoning
This model card describes the ViGoRL (**Vi**sually **G**r**o**unded **R**einforcement **L**earning) model, introduced in our paper ["Grounded Reinforcement Learning for Visual Reasoning"](https://arxiv.org/abs/2505.23678).
**Authors:** Gabriel Sarch, Snigdha Saha, Naitik Khandelwal, Ayush Jain, Michael J. Tarr, Aviral Kumar, Katerina Fragkiadaki
---
## Model Overview
ViGoRL is a vision-language model fine-tuned using reinforcement learning (RL) to explicitly anchor textual reasoning steps to visual coordinates. Inspired by human visual cognition, ViGoRL employs multi-turn visual grounding, dynamically zooming into image regions to perform fine-grained visual reasoning and grounding.
This model was trained using supervised fine-tuning (SFT) on visually-grounded reasoning traces generated via Monte Carlo Tree Search (MCTS), followed by reinforcement learning with Group Relative Policy Optimization (GRPO).
---
## Model Details
* **Base Architecture:** Qwen2.5-Vision-Language (3B or 7B parameters)
* **Training Paradigm:**
* Supervised Fine-Tuning on MCTS-generated reasoning traces
* Group Relative Policy Optimization (GRPO)
* Multi-turn visual grounding with dynamic zoom-in feedback (if "Multiturn" appears in name)
---
## Use Cases
This model excels in visual reasoning tasks that require precise visual grounding and region-level reasoning. Please see model name for specific domain.
* **Spatial Reasoning:** SAT-2, BLINK, RoboSpatial
* **Visual Search:** V\*Bench
* **Web Interaction and Grounding:** ScreenSpot (Pro and V2), VisualWebArena
---
## Usage
You can load this model easily using Hugging Face's Transformers library:
```python
from transformers import Qwen2_5_VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch
# # default: Load the model on the available device(s)
# model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
# "gsarch/ViGoRL-Multiturn-3b-Visual-Search", torch_dtype="auto", device_map="auto"
# ) # replace with any of the ViGoRL models
# We recommend enabling flash_attention_2 for better acceleration and memory saving.
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
"gsarch/ViGoRL-Multiturn-3b-Visual-Search",
torch_dtype=torch.bfloat16,
attn_implementation="flash_attention_2",
device_map="auto",
)
# default processer
processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-Multiturn-3b-Visual-Search")
# The default range for the number of visual tokens per image in the model is 4-16384.
# You can set min_pixels and max_pixels according to your needs, such as a token range of 256-1280, to balance performance and cost.
# min_pixels = 256*28*28
# max_pixels = 1280*28*28
# processor = AutoProcessor.from_pretrained("gsarch/ViGoRL-Multiturn-3b-Visual-Search", min_pixels=min_pixels, max_pixels=max_pixels)
# messages = [
# {
# "role": "user",
# "content": [
# {
# "type": "image",
# "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg",
# },
# {"type": "text", "text": "What color is the leash."},
# ],
# }
# ]
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": "path/to/image.png",
},
{"type": "text", "text": "QUERY HERE"},
],
}
]
# Preparation for inference
text = processor.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
text=[text],
images=image_inputs,
videos=video_inputs,
padding=True,
return_tensors="pt",
)
inputs = inputs.to("cuda")
# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=512)
generated_ids_trimmed = [
out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text) # this will output a single tool call turn of the model if version is multiturn.
# Example output of gsarch/ViGoRL-Multiturn-3b-Visual-Search: ['<think> The leash appears to be red, as seen near the dog\'s paw and the person\'s hand. (1028, 1093). </think>\n<tool_call>\n{"name": "search_coordinate", "arguments": {"coordinate": [1028, 1093]}}\n</tool_call>']
```
**Important**: This model requires a system prompt for proper usage. Please see the model's chat template for details.
---
## Datasets and Training Data
Training datasets and generated reasoning chains are publicly available:
* [Code](https://github.com/Gabesarch/grounded-rl)
* [ViGoRL Datasets on Hugging Face](https://huggingface.co/datasets/gsarch/vigorl_datasets)
---
## Citation
If you use ViGoRL in your research or applications, please cite our paper:
```bibtex
@article{sarch2025vigorl,
title={Grounded Reinforcement Learning for Visual Reasoning},
author={Sarch, Gabriel and Saha, Snigdha and Khandelwal, Naitik and Jain, Ayush and Tarr, Michael J and Kumar, Aviral and Fragkiadaki, Katerina},
year={2025}
}
```
---
## Contact
For questions, feedback, or collaborations, please reach out to Gabriel Sarch or open an issue in our [GitHub repository](https://github.com/Gabesarch/grounded-rl).
---
|
Subsets and Splits
Filtered Qwen2.5 Distill Models
Identifies specific configurations of models by filtering cards that contain 'distill', 'qwen2.5', '7b' while excluding certain base models and incorrect model ID patterns, uncovering unique model variants.
Filtered Model Cards Count
Finds the count of entries with specific card details that include 'distill', 'qwen2.5', '7b' but exclude certain base models, revealing valuable insights about the dataset's content distribution.
Filtered Distill Qwen 7B Models
Filters for specific card entries containing 'distill', 'qwen', and '7b', excluding certain strings and patterns, to identify relevant model configurations.
Filtered Qwen-7b Model Cards
The query performs a detailed filtering based on specific keywords and excludes certain entries, which could be useful for identifying a specific subset of cards but does not provide deeper insights or trends.
Filtered Qwen 7B Model Cards
The query filters for specific terms related to "distilled" or "distill", "qwen", and "7b" in the 'card' column but excludes certain base models, providing a limited set of entries for further inspection.
Qwen 7B Distilled Models
The query provides a basic filtering of records to find specific card names that include keywords related to distilled Qwen 7b models, excluding a particular base model, which gives limited insight but helps in focusing on relevant entries.
Qwen 7B Distilled Model Cards
The query filters data based on specific keywords in the modelId and card fields, providing limited insight primarily useful for locating specific entries rather than revealing broad patterns or trends.
Qwen 7B Distilled Models
Finds all entries containing the terms 'distilled', 'qwen', and '7b' in a case-insensitive manner, providing a filtered set of records but without deeper analysis.
Distilled Qwen 7B Models
The query filters for specific model IDs containing 'distilled', 'qwen', and '7b', providing a basic retrieval of relevant entries but without deeper analysis or insight.
Filtered Model Cards with Distill Qwen2.
Filters and retrieves records containing specific keywords in the card description while excluding certain phrases, providing a basic count of relevant entries.
Filtered Model Cards with Distill Qwen 7
The query filters specific variations of card descriptions containing 'distill', 'qwen', and '7b' while excluding a particular base model, providing limited but specific data retrieval.
Distill Qwen 7B Model Cards
The query filters and retrieves rows where the 'card' column contains specific keywords ('distill', 'qwen', and '7b'), providing a basic filter result that can help in identifying specific entries.