Datasets:

---

renwei2024

/

nemo-github-issues

like 0

[]

[]

[]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "jenkins", "jenkins" ]

[ { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" }, { "id": 6627254865, "node_id": "LA_kwDOC_bI7s8AAAABiwPiUQ", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Multi%20Modal", "name": "Multi Modal", "color": "ededed", "default": false, "description": null } ]

[]

[ "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 6627254865, "node_id": "LA_kwDOC_bI7s8AAAABiwPiUQ", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Multi%20Modal", "name": "Multi Modal", "color": "ededed", "default": false, "description": null } ]

[]

[ "jenkins", "jenkins", "jenkins", "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[ "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "jenkins", "jenkins", "jenkins" ]

[]

[]

[ "Thanks ! " ]

[ { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null }, { "id": 4811884691, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR", "name": "ASR", "color": "F3B33E", "default": false, "description": "" } ]

[]

[ "This PR is stale because it has been open for 14 days with no activity. Remove stale label or comment or update or this will be closed in 7 days.", "This PR was closed because it has been inactive for 7 days since being marked as stale." ]

[ { "id": 1485815557, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU3", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" }, { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null } ]

[]

[ "This is a monolingual Hindi only model, it doesn't support English. ", "What can i use for hinglish(english + hindi)?\r\n\r\nOn Fri, 12 Apr, 2024, 5:33 pm Somshubra Majumdar, ***@***.***>\r\nwrote:\r\n\r\n> This is a monolingual Hindi only model, it doesn't support English.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/NVIDIA/NeMo/issues/8901#issuecomment-2051627631>, or\r\n> unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AJ6XWJZCH3YIE3ULGCFRUULY47EPVAVCNFSM6AAAAABGD7O7A6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANJRGYZDONRTGE>\r\n> .\r\n> You are receiving this because you authored the thread.Message ID:\r\n> ***@***.***>\r\n>\r\n", "How can i get this hindi model to work for hindi alone. I followed the steps but havent been successful", "This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.", "This issue was closed because it has been inactive for 7 days since being marked as stale." ]

[ { "id": 1485815557, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU3", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" }, { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null } ]

[ { "login": "ericharper", "id": 11999610, "node_id": "MDQ6VXNlcjExOTk5NjEw", "avatar_url": "https://avatars.githubusercontent.com/u/11999610?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ericharper", "html_url": "https://github.com/ericharper", "followers_url": "https://api.github.com/users/ericharper/followers", "following_url": "https://api.github.com/users/ericharper/following{/other_user}", "gists_url": "https://api.github.com/users/ericharper/gists{/gist_id}", "starred_url": "https://api.github.com/users/ericharper/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ericharper/subscriptions", "organizations_url": "https://api.github.com/users/ericharper/orgs", "repos_url": "https://api.github.com/users/ericharper/repos", "events_url": "https://api.github.com/users/ericharper/events{/privacy}", "received_events_url": "https://api.github.com/users/ericharper/received_events", "type": "User", "user_view_type": "public", "site_admin": false } ]

[ "This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.", "This issue was closed because it has been inactive for 7 days since being marked as stale." ]

[ { "id": 1485815557, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU3", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" }, { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null } ]

[]

[ "I'm on the commit:97d1abb2bca0b5daff6d434c4bb340d3bb702e86", "This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.", "This issue was closed because it has been inactive for 7 days since being marked as stale.", "Does this error solved? \r\nIt seems that _load_state_dict_from_disk expects model.ckpt file, but the result of untar model.nemo generates model_weights folder. \r\n", "same issue I manage to run the pretraining script by setting model.mm_cfg.llm.from_pretrained=null and it works but is he seems to pretraing the llm from scratch (?)" ]

[ { "id": 1485815557, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU3", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" }, { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null } ]

[]

[ "Is it GPU or CPU memory that is exhausted ? And how many nodes are you using ? \r\n\r\nWhat version of NeMo are you using ? \r\nWithout sufficient details it's not possible to debug. \r\n\r\nWhat I can say is we train on nodes with 400 GB ram per node and A100 with 80GB gpu memory and train on 90-400K hours of speech without oom in either CPU or GPU memory. \r\n\r\nIf you can visibly see CPU ram constantly increase during training, a pseudo fix could be to use exp_manager.max_time_per_run and set it to a reasonable value like a day, then the job stops after a day and you can restart it and avoid memory leak. It's not a fix but a temporary solution ", "Is it GPU or CPU memory that is exhausted ? And how many nodes are you using ?\r\n1. CPU not GPU\r\n2. Just single node\r\nWe just added one row \r\nself.log('loss', loss_value, on_step=True, prog_bar=True, on_epoch=False,)\r\nin file: \r\nnemo/collections/asr/models/ctc_models.py\r\nPreviously, we used on_epoch=True, but now the problem still remains after changine to False.\r\n\r\nWhat version of NeMo are you using ?\r\n\r\n git log\r\ncommit 0d3d8fa1bc78135e72400e8deecd69cfb3d9aa3f (HEAD -> main)\r\nAuthor: anteju <[email protected]>\r\nDate: Wed Nov 15 16:56:29 2023 -0800\r\n\r\n [ASR] GSS-based mask estimator (#7849)\r\n \r\n * Added GSS-based mask estimator for multispeaker scenarios\r\n \r\n Signed-off-by: Ante Jukić <[email protected]>\r\n \r\n * Addressed PR comments\r\n \r\n Signed-off-by: Ante Jukić <[email protected]>\r\n \r\n ---------\r\n \r\n Signed-off-by: Ante Jukić <[email protected]>\r\n Co-authored-by: Taejin Park <[email protected]>\r\n\r\nActually, it's very easy to verify: you just submit a training task with, say librispeech data, you can observe you CPU memory keeps increasing within an epoch. \r\nBut such memory increase won't hurt since memory increase slow and after an epoch, memory usage somehwo is going down again. Here, if we decrease our training data down to 30k, for 1.2T cpu memory, we can finish an epoch normally.\r\n", "On Fri, Apr 12, 2024 at 2:21 PM Somshubra Majumdar ***@***.***>\r\nwrote:\r\n\r\n> Is it GPU or CPU memory that is exhausted ? And how many nodes are you\r\n> using ?\r\n>\r\n> What version of NeMo are you using ?\r\n> Without sufficient details it's not possible to debug.\r\n>\r\n> What I can say is we train on nodes with 400 GB ram per node and A100 with\r\n> 80GB gpu memory and train on 90-400K hours of speech without oom in either\r\n> CPU or GPU memory.\r\n>\r\nHow many nodes have you used ? if you use a lot of nodes, then you might\r\nnot trigger the bugs. Say, you have used 8 nodes, then there might be no\r\nissues ...\r\n\r\nRegards,\r\n\r\nHaihua\r\n\r\n> If you can visibly see CPU ram constantly increase during training, a\r\n> pseudo fix could be to use exp_manager.max_time_per_run and set it to a\r\n> reasonable value like a day, then the job stops after a day and you can\r\n> restart it and avoid memory leak. It's not a fix but a temporary solution\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/NVIDIA/NeMo/issues/8897#issuecomment-2051065239>, or\r\n> unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAZBHYOSYLVXC677XXF2LWLY454PZAVCNFSM6AAAAABGDMTIVKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANJRGA3DKMRTHE>\r\n> .\r\n> You are receiving this because you authored the thread.Message ID:\r\n> ***@***.***>\r\n>\r\n", "That Nemo version is 6 months old, can you use r1.23 and see if it persists ? We do not see constantly increasing CPU memory per epoch, but that may be because we use multiple nodes - min 4 nodes ", "Hi, is this issue resolved? I've been running into the same issue. (I can confirm that it happens on 1.23 as well)", "Hi there,\r\n\r\nJust checking here and wondering whether this is resolved?\r\nI am facing same issue.\r\n\r\nThank you.", "using multiple nodes to train can avoid the problem.\r\n\r\nOn Fri, May 24, 2024, 1:58 AM ROZBEH ***@***.***> wrote:\r\n\r\n> Hi there,\r\n>\r\n> Just checking here and wondering whether this is resolved?\r\n> I am facing same issue.\r\n>\r\n> Thank you.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/NVIDIA/NeMo/issues/8897#issuecomment-2127743659>, or\r\n> unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAZBHYJPQ36M6WWYAGGERQLZDYU2RAVCNFSM6AAAAABGDMTIVKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMRXG42DGNRVHE>\r\n> .\r\n> You are receiving this because you authored the thread.Message ID:\r\n> ***@***.***>\r\n>\r\n", "Thanks @haihua \nI'm indeed 5 nodes with 5 GPU each. Is that what you mean?", "Yes, that's it.\r\n\r\nOn Fri, May 24, 2024, 8:06 PM ROZBEH ***@***.***> wrote:\r\n\r\n> Thanks @haihua <https://github.com/haihua>\r\n> I'm indeed 5 nodes with 5 GPU each. Is that what you mean?\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/NVIDIA/NeMo/issues/8897#issuecomment-2129368588>, or\r\n> unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAZBHYLT2XC5A4QKW26PJHDZD4ULRAVCNFSM6AAAAABGDMTIVKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMRZGM3DQNJYHA>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "I see but the above issue is persistent with multi node and I'd like to get it working.", "This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.", "This issue was closed because it has been inactive for 7 days since being marked as stale.", "@ROZBEH did you ever solve the problem? @haihua did you find the origin of the problem?", "Unfortunately no" ]

[ { "id": 4811884691, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR", "name": "ASR", "color": "F3B33E", "default": false, "description": "" }, { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 6627254865, "node_id": "LA_kwDOC_bI7s8AAAABiwPiUQ", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Multi%20Modal", "name": "Multi Modal", "color": "ededed", "default": false, "description": null } ]

[]

[]

[]

[]

[ "jenkins", "jenkins" ]

[ { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[]

[ { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null }, { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "jenkins", "jenkins", "This PR is stale because it has been open for 14 days with no activity. Remove stale label or comment or update or this will be closed in 7 days.", "This PR was closed because it has been inactive for 7 days since being marked as stale." ]

[]

[]

[]

[]

[ { "login": "okuchaiev", "id": 22035961, "node_id": "MDQ6VXNlcjIyMDM1OTYx", "avatar_url": "https://avatars.githubusercontent.com/u/22035961?v=4", "gravatar_id": "", "url": "https://api.github.com/users/okuchaiev", "html_url": "https://github.com/okuchaiev", "followers_url": "https://api.github.com/users/okuchaiev/followers", "following_url": "https://api.github.com/users/okuchaiev/following{/other_user}", "gists_url": "https://api.github.com/users/okuchaiev/gists{/gist_id}", "starred_url": "https://api.github.com/users/okuchaiev/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/okuchaiev/subscriptions", "organizations_url": "https://api.github.com/users/okuchaiev/orgs", "repos_url": "https://api.github.com/users/okuchaiev/repos", "events_url": "https://api.github.com/users/okuchaiev/events{/privacy}", "received_events_url": "https://api.github.com/users/okuchaiev/received_events", "type": "User", "user_view_type": "public", "site_admin": false } ]

[ "@artbataev this may be related to your observation of TDT being faster than CTC in some cases.", "If I recall correctly, max does work on GPU so I don't know why I forced it onto CPU. Maybe the slice? Either way, if it fixes the slowdown let's move the .CPU() to after max calculation.\r\n\r\nIt might have been that the logprobs can be super long for 12 hours so I wanted to preserve memory, in which case the cost of speed is not under consideration but I suppose we can make it a premature optimization and select whether to move the tensor to CPU before or after depending on length of the tensor. ", "It is pretty common to run long audio CTC - NFA does it constantly. But on average, it is more common to run short audio batches ", "> It might have been that the logprobs can be super long for 12 hours so I wanted to preserve memory\r\n\r\nThe tensor pointed to by \"prediction\" already existed on GPU. It is just being copied to CPU in this code. We should not expect any memory increase from my suggestion. In fact, we will see a CPU memory decrease.", "That's true. Can you see if moving it onto CPU at the end before return is sufficient to remove this bottleneck then let's do it in your other pr " ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[ "@cuichenx", "jenkins", "jenkins" ]

[ { "id": 4418253970, "node_id": "LA_kwDOC_bI7s8AAAABB1k0kg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS", "name": "TTS", "color": "49AF28", "default": false, "description": "" }, { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[ "jenkins", "jenkins" ]

[]

[]

[]

[ { "id": 1485815557, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU3", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" } ]

[]

[]

[ { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null }, { "id": 4811884691, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR", "name": "ASR", "color": "F3B33E", "default": false, "description": "" }, { "id": 4847373924, "node_id": "LA_kwDOC_bI7s8AAAABIO0OZA", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/common", "name": "common", "color": "c5def5", "default": false, "description": "" } ]

[]

[ "This PR is stale because it has been open for 14 days with no activity. Remove stale label or comment or update or this will be closed in 7 days.", "This PR was closed because it has been inactive for 7 days since being marked as stale." ]

[ { "id": 1485815557, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU3", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" } ]

[]

[ "NeMo main has an updated version of transcribe() that supports more functionality and deprecated old arguments with new ways of doing things. \r\n\r\nYou can use and install Nemo 1.23.0 branch to get it to work. Or you can use the following to fix transcribe on main. \r\n\r\nChange \r\nlogits = quartznet.transcribe(wave_file, logprobs=True)[0]\r\n\r\nTo \r\n\r\n```python\r\n\r\nLogits_hyps = asr_model.transcribe(\r\n self.audio_file_list, batch_size=self.asr_batch_size, return_hypotheses=True\r\n ) # type: List[nemo_asr.parts.Hypothesis]\r\n \r\nLogits = [hyp.alignments for hyp in Logits_hyps]\r\n```\r\n", "Thank you very much!" ]

[ { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" }, { "id": 6627254865, "node_id": "LA_kwDOC_bI7s8AAAABiwPiUQ", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Multi%20Modal", "name": "Multi Modal", "color": "ededed", "default": false, "description": null } ]

[]

[ "jenkins", "jenkins", "jenkins", "Set unet_precision when user not set it.\r\n", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "LGTM, We did convergence runs and convergence looks as expected on GBS=8, 32. We see measurable speedup for this change.\r\nOne minor comment:\r\n- We should try to JIT the python code, but maybe that can happen as a separate PR.\r\n\r\n@ericharper @cuichenx can you pls review/comment?", "jenkins" ]

[ { "id": 1923495043, "node_id": "MDU6TGFiZWwxOTIzNDk1MDQz", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/core", "name": "core", "color": "f25e3a", "default": false, "description": "Changes to NeMo Core" }, { "id": 4811884691, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR", "name": "ASR", "color": "F3B33E", "default": false, "description": "" }, { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" }, { "id": 6627254865, "node_id": "LA_kwDOC_bI7s8AAAABiwPiUQ", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Multi%20Modal", "name": "Multi Modal", "color": "ededed", "default": false, "description": null } ]

[]

[ "jenkins", "Will reopen with another branch soon." ]

[ { "id": 1923495043, "node_id": "MDU6TGFiZWwxOTIzNDk1MDQz", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/core", "name": "core", "color": "f25e3a", "default": false, "description": "Changes to NeMo Core" } ]

[]

[ "> Overall looks good, although we are still hashing out the design in [NVIDIA/apex#1794](https://github.com/NVIDIA/apex/pull/1794). As discussed in [NVIDIA/apex#1794 (comment)](https://github.com/NVIDIA/apex/pull/1794#issuecomment-2050894033), I think we should set `MegatronDistributedFusedAdam._step_support_amp_scaling=False` to signal that the NeMo grad scaler can accommodate the distributed optimizer (unlike the plain PyTorch grad scaler). As a bonus, this approach fixes the grad scaling issue even without needing to update Apex.\r\n\r\nYou are probably right, to make sure it won't hurt perf I need to confirm with our usecases. Anyway, I think that relates to the changes in APEX, the changes here have no relation to gradient clipping.", "jenkins", "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "Tested that the guard works correctly now, and will also work in 1.6 release where version would be 1.6.0.", "jenkins" ]

[ { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[]

[ { "id": 6627254865, "node_id": "LA_kwDOC_bI7s8AAAABiwPiUQ", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Multi%20Modal", "name": "Multi Modal", "color": "ededed", "default": false, "description": null } ]

[]

[ "jenkins", "jenkins" ]

[ { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" }, { "id": 6627254865, "node_id": "LA_kwDOC_bI7s8AAAABiwPiUQ", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Multi%20Modal", "name": "Multi Modal", "color": "ededed", "default": false, "description": null } ]

[]

[ "jenkins", "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "@mikolajblaz can you review?", "jenkins", "jenkins", "jenkins" ]

[ { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[]

[]

[]

[ "@JimmyZhang12 thanks for your review. Addressed your comments. Could you please approve?", "jenkins", "jenkins" ]

[ { "id": 1485815557, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU3", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" }, { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null } ]

[]

[ "{\r\n \"train_loss\": 8.534393310546875,\r\n \"learning_rate\": 0.00012372159108053893,\r\n \"global_step\": 5610,\r\n \"train_backward_timing in s\": 0.15601778030395508,\r\n \"train_step_timing in s\": 0.34882450103759766,\r\n \"training_batch_wer\": 0.18061673641204834,\r\n \"epoch\": 30,\r\n \"trainer/global_step\": 5610,\r\n \"_timestamp\": 1712744688.4421802,\r\n \"_runtime\": 11167.270045042038,\r\n \"_step\": 293353,\r\n \"validation_step_timing in s\": 0.02845907211303711,\r\n \"val_loss\": 6.631827354431152,\r\n \"val_wer\": 0.12062931060791016,\r\n \"test_step_timing in s\": 0.043920040130615234,\r\n \"test_loss\": 10.514272689819336,\r\n \"test_wer\": 0.17779342830181122,\r\n \"_wandb\": {\r\n \"runtime\": 11166\r\n }", "i used common voice13 hindi as dataset\r\n", "my issue is that wer for the model i used i.e conformer_large_ctc is very low .what could be the reason for such a low wer.\r\n@titu1994 sir", "I'm confused, the metric is \"word error rate\" - it's good if it's low. Ideal wer is 0", "actually i was worried about overfitting of data btw thanks sir 🙌", "This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.", "This issue was closed because it has been inactive for 7 days since being marked as stale." ]

[ { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null } ]

[ { "login": "okuchaiev", "id": 22035961, "node_id": "MDQ6VXNlcjIyMDM1OTYx", "avatar_url": "https://avatars.githubusercontent.com/u/22035961?v=4", "gravatar_id": "", "url": "https://api.github.com/users/okuchaiev", "html_url": "https://github.com/okuchaiev", "followers_url": "https://api.github.com/users/okuchaiev/followers", "following_url": "https://api.github.com/users/okuchaiev/following{/other_user}", "gists_url": "https://api.github.com/users/okuchaiev/gists{/gist_id}", "starred_url": "https://api.github.com/users/okuchaiev/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/okuchaiev/subscriptions", "organizations_url": "https://api.github.com/users/okuchaiev/orgs", "repos_url": "https://api.github.com/users/okuchaiev/repos", "events_url": "https://api.github.com/users/okuchaiev/events{/privacy}", "received_events_url": "https://api.github.com/users/okuchaiev/received_events", "type": "User", "user_view_type": "public", "site_admin": false } ]

[ "We don't have pretrained onnx models released in NeMo. Nvidia Riva might have .Riva files that contain a compiled model but that isn't supported in NeMo ", "This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.", "This issue was closed because it has been inactive for 7 days since being marked as stale." ]

[ { "id": 1923495043, "node_id": "MDU6TGFiZWwxOTIzNDk1MDQz", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/core", "name": "core", "color": "f25e3a", "default": false, "description": "Changes to NeMo Core" }, { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" }, { "id": 6627254865, "node_id": "LA_kwDOC_bI7s8AAAABiwPiUQ", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Multi%20Modal", "name": "Multi Modal", "color": "ededed", "default": false, "description": null } ]

[]

[]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "jenkins", "Hi @vysarge , I do not totally understand what the first input dim and the second dim mean, could you please describe it? And how does it work if `sequence_parallel=True` and `context_parallel_size > 1`? Thanks.", "> Hi @vysarge , I do not totally understand what the first input dim and the second dim mean, could you please describe it? And how does it work if `sequence_parallel=True` and `context_parallel_size > 1`? Thanks.\r\n\r\nHm, I thought sequence parallel and context parallel were mutually exclusive. If they're used together, do CP and TP counts split on the same dimension? Making the change you suggested should fix this case too if so, let me test it.", "> > Hi @vysarge , I do not totally understand what the first input dim and the second dim mean, could you please describe it? And how does it work if `sequence_parallel=True` and `context_parallel_size > 1`? Thanks.\r\n> \r\n> Hm, I thought sequence parallel and context parallel were mutually exclusive. If they're used together, do CP and TP counts split on the same dimension? Making the change you suggested should fix this case too if so, let me test it.\r\n\r\nSequence parallel and context parallel can work together. Sequence parallel only split sequence of LayerNorm. In addition, context parallel split sequence of the whole transformer layer. Yeah, you can test my suggested fix, I think it should work for all cases. Thanks.", "> > > Hi @vysarge , I do not totally understand what the first input dim and the second dim mean, could you please describe it? And how does it work if `sequence_parallel=True` and `context_parallel_size > 1`? Thanks.\r\n> > \r\n> > \r\n> > Hm, I thought sequence parallel and context parallel were mutually exclusive. If they're used together, do CP and TP counts split on the same dimension? Making the change you suggested should fix this case too if so, let me test it.\r\n> \r\n> Sequence parallel and context parallel can work together. Sequence parallel only split sequence of LayerNorm. In addition, context parallel split sequence of the whole transformer layer. Yeah, you can test my suggested fix, I think it should work for all cases. Thanks.\r\n\r\nTested and this fix is working well; PR has been updated, thanks.", "LGTM, thanks.", "jenkins", "This PR is stale because it has been open for 14 days with no activity. Remove stale label or comment or update or this will be closed in 7 days." ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "Re #8803: `nemo/collections/nlp/models/information_retrieval/megatron_sbert_model.py` was removed with [this commit](https://github.com/NVIDIA/NeMo/commit/b273fa6caf41bde383e93adfaa10349a3f361807).", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins" ]

[ { "id": 1485815557, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU3", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" }, { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" } ]

[]

[ "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins" ]

[]

[]

[ "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins" ]

[ { "id": 6876332615, "node_id": "LA_kwDOC_bI7s8AAAABmdyCRw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/Run%20CICD", "name": "Run CICD", "color": "fef2c0", "default": false, "description": "" } ]

[]

[ "Can this be merged @oyilmaz-nvidia ?", "jenkins" ]

[]

[]

[ "use ctc models for aligning very long audio duration files. ", "You can use hybrid models in CTC mode automatically. @erastorgueva-nv could you advise ", "It should work if you also add the parameter `use_local_attention=true` (reference: https://docs.nvidia.com/nemo-framework/user-guide/latest/nemotoolkit/tools/nemo_forced_aligner.html#optional-parameters)", "I tried using the model `stt_en_conformer_ctc_small` and got the same result, then I tried playing with the Viterbi device and the transcriber device values, passing both of those values as CPU killed the process, no error log or anything, just this:\r\n`Transcribing: 0% 0/1 [00:00<?, ?it/s]^C`\r\n\r\nWhen passing `transcribe_device='cuda'` I got the CUDA OOM error again.\r\nWhen passing `viterbi_device='cpu'` I got the following output again: `Transcribing: 0% 0/1 [00:00<?, ?it/s]^C`\r\n\r\nThen I tried running the command using the following model:\r\n\r\n`stt_en_conformer_ctc_small`\r\n\r\nhere I got a different error shared below:\r\n\r\n```\r\n!HYDRA_FULL_ERROR=1 python $NEMO_DIR_PATH/tools/nemo_forced_aligner/align.py \\\r\n pretrained_name=\"stt_en_conformer_ctc_small\" \\\r\n manifest_filepath=$manifest_filepath \\\r\n output_dir=$WORK_DIR/nfa_output/ \\\r\n use_local_attention=true \\\r\n transcribe_device='cuda' \\\r\n viterbi_device='cuda'\r\n```\r\n\r\nLogs:\r\n\r\n```\r\n[NeMo I 2024-04-11 23:21:25 features:289] PADDING: 0\r\n[NeMo I 2024-04-11 23:21:26 save_restore_connector:263] Model EncDecCTCModelBPE was successfully restored from /root/.cache/torch/NeMo/NeMo_1.23.0rc0/stt_en_conformer_ctc_small/5d2d8e5b2b5adb8f5091363c6ba19c55/stt_en_conformer_ctc_small.nemo.\r\n[NeMo I 2024-04-11 23:21:26 align:250] Flag use_local_attention is set to True => will try to use local attention for model if it allows it\r\nTranscribing: 100% 1/1 [00:06<00:00, 6.46s/it]\r\nError executing job with overrides: ['pretrained_name=stt_en_conformer_ctc_small', 'manifest_filepath=WORK_DIR/manifest.json', 'output_dir=WORK_DIR/nfa_output/', 'use_local_attention=true', 'transcribe_device=cuda', 'viterbi_device=cuda']\r\nTraceback (most recent call last):\r\n File \"/content/NeMo/tools/nemo_forced_aligner/align.py\", line 352, in <module>\r\n main()\r\n File \"/usr/local/lib/python3.10/dist-packages/nemo/core/config/hydra_runner.py\", line 129, in wrapper\r\n _run_hydra(\r\n File \"/usr/local/lib/python3.10/dist-packages/hydra/_internal/utils.py\", line 394, in _run_hydra\r\n _run_app(\r\n File \"/usr/local/lib/python3.10/dist-packages/hydra/_internal/utils.py\", line 457, in _run_app\r\n run_and_report(\r\n File \"/usr/local/lib/python3.10/dist-packages/hydra/_internal/utils.py\", line 223, in run_and_report\r\n raise ex\r\n File \"/usr/local/lib/python3.10/dist-packages/hydra/_internal/utils.py\", line 220, in run_and_report\r\n return func()\r\n File \"/usr/local/lib/python3.10/dist-packages/hydra/_internal/utils.py\", line 458, in <lambda>\r\n lambda: hydra.run(\r\n File \"/usr/local/lib/python3.10/dist-packages/hydra/_internal/hydra.py\", line 132, in run\r\n _ = ret.return_value\r\n File \"/usr/local/lib/python3.10/dist-packages/hydra/core/utils.py\", line 260, in return_value\r\n raise self._return_value\r\n File \"/usr/local/lib/python3.10/dist-packages/hydra/core/utils.py\", line 186, in run_job\r\n ret.return_value = task_function(task_cfg)\r\n File \"/content/NeMo/tools/nemo_forced_aligner/align.py\", line 318, in main\r\n (log_probs_batch, y_batch, T_batch, U_batch, utt_obj_batch, output_timestep_duration,) = get_batch_variables(\r\n File \"/content/NeMo/tools/nemo_forced_aligner/utils/data_prep.py\", line 748, in get_batch_variables\r\n utt_obj = get_utt_obj(\r\n File \"/content/NeMo/tools/nemo_forced_aligner/utils/data_prep.py\", line 320, in get_utt_obj\r\n word_tokens_cased = restore_token_case(word, word_tokens)\r\n File \"/content/NeMo/tools/nemo_forced_aligner/utils/data_prep.py\", line 164, in restore_token_case\r\n if token_char == word[word_char_pointer]:\r\nIndexError: string index out of range\r\n```\r\n\r\n\r\nIs there a way to know how many resources we need to run any model?", "After some trial and error, I figured out my text had some characters that didn't belong to the transcription.\r\n\r\nI was able to align the 1h 30m audio I needed.\r\n\r\nThanks for the help!", "Glad you were able to get it to work! A pointer about the command you posted:\r\n```\r\n!HYDRA_FULL_ERROR=1 PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True python $NEMO_DIR_PATH/tools/nemo_forced_aligner/align.py \\\r\n pretrained_name=\"stt_en_fastconformer_hybrid_large_pc\" \\\r\n manifest_filepath=$manifest_filepath \\\r\n output_dir=$WORK_DIR/nfa_output/ \\\r\n total_buffer_in_secs=2 \\\r\n chunk_batch_size=16 \\\r\n chunk_len_in_secs=0.8\r\n```\r\n^ in the above command, buffered chunk streaming is not being done (and the params `total_buffer_in_secs`, `chunk_batch_size`, `chunk_len_in_secs` are being ignored). If you want to use it, you need to add `use_buffered_chunked_streaming=true`. But seems like you do not need it.\r\n" ]

[]

[]

[]

[ { "id": 4811884691, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ikw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/ASR", "name": "ASR", "color": "F3B33E", "default": false, "description": "" } ]

[]

[ "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[ "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins", "jenkins" ]

[ { "id": 1485815558, "node_id": "MDU6TGFiZWwxNDg1ODE1NTU4", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/documentation", "name": "documentation", "color": "0075ca", "default": true, "description": "Improvements or additions to documentation" }, { "id": 4418253970, "node_id": "LA_kwDOC_bI7s8AAAABB1k0kg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/TTS", "name": "TTS", "color": "49AF28", "default": false, "description": "" } ]

[]

[ "jenkins", "jenkins" ]

[ { "id": 4811886210, "node_id": "LA_kwDOC_bI7s8AAAABHs-Ogg", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/NLP", "name": "NLP", "color": "CBF09A", "default": false, "description": "" }, { "id": 4840216855, "node_id": "LA_kwDOC_bI7s8AAAABIH_ZFw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/CI", "name": "CI", "color": "1D76DB", "default": false, "description": "" } ]

[]

[ "jenkins" ]

[ { "id": 4585431311, "node_id": "LA_kwDOC_bI7s8AAAABEVAhDw", "url": "https://api.github.com/repos/NVIDIA/NeMo/labels/stale", "name": "stale", "color": "ededed", "default": false, "description": null } ]

[ { "login": "okuchaiev", "id": 22035961, "node_id": "MDQ6VXNlcjIyMDM1OTYx", "avatar_url": "https://avatars.githubusercontent.com/u/22035961?v=4", "gravatar_id": "", "url": "https://api.github.com/users/okuchaiev", "html_url": "https://github.com/okuchaiev", "followers_url": "https://api.github.com/users/okuchaiev/followers", "following_url": "https://api.github.com/users/okuchaiev/following{/other_user}", "gists_url": "https://api.github.com/users/okuchaiev/gists{/gist_id}", "starred_url": "https://api.github.com/users/okuchaiev/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/okuchaiev/subscriptions", "organizations_url": "https://api.github.com/users/okuchaiev/orgs", "repos_url": "https://api.github.com/users/okuchaiev/repos", "events_url": "https://api.github.com/users/okuchaiev/events{/privacy}", "received_events_url": "https://api.github.com/users/okuchaiev/received_events", "type": "User", "user_view_type": "public", "site_admin": false } ]

[ "Here is the tutorial for VAD: https://github.com/NVIDIA/NeMo/blob/main/tutorials/asr/Voice_Activity_Detection.ipynb \r\n\r\nCurrently we don;t support training VAD and SD model together", "> Here is the tutorial for VAD: https://github.com/NVIDIA/NeMo/blob/main/tutorials/asr/Voice_Activity_Detection.ipynb\r\n> \r\n> Currently we don;t support training VAD and SD model together\r\n\r\nThank you for your answer! Additionally, I have a question about training dataset. Can one example of dataset contain more than 2 speech chunks?", "I didn;t understand your question well: `Can one utterance contain more than 2 speech chunks?` \r\nDo you mean, if VAD is frame or segment based? \r\n\r\n", "This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.", "This issue was closed because it has been inactive for 7 days since being marked as stale." ]