Datasets:

---

renwei2024

/

nemo-github-issues

like 0

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[ "Can you also update latest TalkNet weights in pretrained link?", "LGTM\r\ni will update ckpts links in following days, as well as TTS notebook for inference", "@stasbel @blisc I found additional bug in HiFiGAN tuning. I will fix it and after that we will merge it", "@blisc let's merge it after this PR: https://github.com/NVIDIA/NeMo/pull/1994" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "This pull request **introduces 1 alert** when merging d5f866692e69b78a5565dd635e396521ebecbd67 into 91146cf360415e9656e62c635b48ad639b1b45af - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-a6442172376caa13bb8863b03862186b4e24cc86)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 00fed41bdbdd9eb4fc955af647ec4a1dd4f9e731 into 91146cf360415e9656e62c635b48ad639b1b45af - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-b5ddf66a90906faf3a185779453eb1f91747d55b)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 472a28dfaab024e4482513001c214de86f57769e into 91146cf360415e9656e62c635b48ad639b1b45af - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-6f080e2f767e616f56d728929f095d8b0bf0e951)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging d76389a45d60d479dcd2762af651bda1c07f87e4 into 91146cf360415e9656e62c635b48ad639b1b45af - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-9761851acbb4a9f8f36290a3dcaf07290b0c6e0f)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging f3f232e5bc91449ef5adc2f740c952352f87e9ff into 19af99bc487aeafa091bc5e8901919f0904f8cdb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-6818890be86776575324f89deefdc5fb0eccba71)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 63d783df6b838b80bd1f9b34f15fff7b6d8027bf into 19af99bc487aeafa091bc5e8901919f0904f8cdb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-b57b931e9ffed948bf2f98ee6c773c48d1034a94)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 51d302ca066cf57217f351213c5adef9a813616f into 19af99bc487aeafa091bc5e8901919f0904f8cdb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-936321f8ef7862cbc8b13c48442ba22ab52c82e0)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 5e2ee7b2268bc2367c46f156fe77e187d2ffb264 into 19af99bc487aeafa091bc5e8901919f0904f8cdb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-7ffe27a5669017e30f173fdba287ba4cc49b7138)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 62b6dd4201308a642bc4fa7da39c724f82ab596f into 19af99bc487aeafa091bc5e8901919f0904f8cdb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-cb184c85ad5630f8c510750591039cc0cd1ced5b)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 78245dd858b28c45189fcd9e6b325e07f4c96476 into 4bcb2321cd16835f63afe3dfe993e6d56bcf2c0c - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-886eb38056d6fb503030edeb79864b6fc8b558de)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 70852f1a84d70a4f28257f85fcb364470febc5e6 into 4bcb2321cd16835f63afe3dfe993e6d56bcf2c0c - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-348c3ebf285a19fc689c5bf856c93f9a34e3e0fa)\n\n**new alerts:**\n\n* 1 for Unused import", "This pull request **introduces 2 alerts** when merging b9724dc5b6b41c94d909be88b74f80c437b396ae into df0cbdb278e7e916c37d43126a6746870868f17e - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-59a3b39ed5ef209db4754d99b8a1d8f64e0771cd)\n\n**new alerts:**\n\n* 2 for Unused import", "This pull request **introduces 5 alerts** when merging e4e8f0ca4830608f346875d493a4fb45a8a8c26c into 360eb0422f53f1d3bb41fa941a44591bba4d7de9 - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-4508cbe5ae5363157b52fdd19b59385c58ba7573)\n\n**new alerts:**\n\n* 3 for Unused import\n* 1 for Except block handles 'BaseException'\n* 1 for Module is imported with 'import' and 'import from'", "Close in favor of #2015 " ]

[]

[]

[ "> get test is working, export is skipped (not working)\r\n\r\n\r\n- No, locally I get: tests/collections/nlp/test_megatron.py::TestMegatron::test_get_model FAILED \r\nE RuntimeError: The following operation failed in the TorchScript interpreter.\r\nE Traceback of TorchScript (most recent call last):\r\nE RuntimeError: The following operation failed in the TorchScript interpreter.\r\nE Traceback of TorchScript (most recent call last):\r\nE RuntimeError: nvrtc: error: invalid value for --gpu-architecture (-arch)" ]

[]

[]

[ ">>\"While this works for all restore_from calls, it is not the case for from_pretrained. All domains have their own way of populating the list of pretrained models - and therefore calling from_pretrained from ModelPT will fail since ModelPT has no information of the subclass pretrained models.\"\r\n\r\n- I don't understand this. \"from_pretrained\" just calls \"restore_from\" once .nemo file is downloaded.", "> > > \"While this works for all restore_from calls, it is not the case for from_pretrained. All domains have their own way of populating the list of pretrained models - and therefore calling from_pretrained from ModelPT will fail since ModelPT has no information of the subclass pretrained models.\"\r\n> \r\n> * I don't understand this. \"from_pretrained\" just calls \"restore_from\" once .nemo file is downloaded.", "> I don't understand this. \"from_pretrained\" just calls \"restore_from\" once .nemo file is downloaded.\r\nfrom_pretrained first loads up the list of pretrained models with download links from the `list_available_models` method.\r\n\r\nIf its called from `EncDecCTCModel` (say) - then it will load up the correct set of model `PretrainedInfo`. However, if its called from `ModelPT` - then `ModelPT.restore_from('stt_en_citrinet_256')` will fail - since ModelPT has an empty implementation of `list_available_models`. So it raises an error saying the name `stt_en_citrinet_256` is not valid.\r\n\r\nAn easy solution is to perform a DFS inside ModelPT for all possible subclass implementations of `list_available_models` - and populate a set of all pretraind model info. However, this is up for debate if core should do such a task. For an example of such a population, look to ASRModel's `list_available_models` implementation - it populates its list from all of its own subclasses. ModelPT would just do this, but at one level higher up the chain.", "> > I don't understand this. \"from_pretrained\" just calls \"restore_from\" once .nemo file is downloaded.\r\n> > from_pretrained first loads up the list of pretrained models with download links from the `list_available_models` method.\r\n> \r\n> If its called from `EncDecCTCModel` (say) - then it will load up the correct set of model `PretrainedInfo`. However, if its called from `ModelPT` - then `ModelPT.restore_from('stt_en_citrinet_256')` will fail - since ModelPT has an empty implementation of `list_available_models`. So it raises an error saying the name `stt_en_citrinet_256` is not valid.\r\n> \r\n> An easy solution is to perform a DFS inside ModelPT for all possible subclass implementations of `list_available_models` - and populate a set of all pretraind model info. However, this is up for debate if core should do such a task. For an example of such a population, look to ASRModel's `list_available_models` implementation - it populates its list from all of its own subclasses. ModelPT would just do this, but at one level higher up the chain.\r\n\r\nmakes sense", "how about, inside ```nemo.core.classes.common.FileIO.to_config_file``` we check if \"target\" is present for the model and if not, we add it there?", "> how about, inside nemo.core.classes.common.FileIO.to_config_file we check if \"target\" is present for the model and if not, we add it there?\r\n\r\nProbably not a good idea. ModelPT injects `target` which is safe. since all other config modules inside it are \"nearly\" independent and called explicitly by the constructor of the subclass of ModelPT.\r\n\r\nOn the other hand, `FileIO.to_config_file` is called under many different cases - simple modules like encoder or decoder or spec_augment all can call both `from_config_file` and `to_config_file`. Injecting parameters there would make the constructors of such classes have issues.", ">>\"ModelPT injects target which is safe.\"\r\n\r\n-- where? \r\n\r\nBasically, I'm saying that if we do ```.save_to``` with this version, then ```ModelPT.restore_from``` should work for the newly generated file", "Inside the constructor - https://github.com/NVIDIA/NeMo/blob/main/nemo/core/classes/modelPT.py#L99-L103", "I can add a save-restore-save test, but for some reason I think we have that already. The target injection has been there since b1 so it should be safe if the model tests regarding Citrinet and rest are passing. " ]

[]

[]

[ "Just saw @ekmb didn't review yet. Changes should be fine, but revert in a subsequent PR for NLP model test if needed. " ]

[]

[]

[]

[]

[ { "login": "vsl9", "id": 4344862, "node_id": "MDQ6VXNlcjQzNDQ4NjI=", "avatar_url": "https://avatars.githubusercontent.com/u/4344862?v=4", "gravatar_id": "", "url": "https://api.github.com/users/vsl9", "html_url": "https://github.com/vsl9", "followers_url": "https://api.github.com/users/vsl9/followers", "following_url": "https://api.github.com/users/vsl9/following{/other_user}", "gists_url": "https://api.github.com/users/vsl9/gists{/gist_id}", "starred_url": "https://api.github.com/users/vsl9/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/vsl9/subscriptions", "organizations_url": "https://api.github.com/users/vsl9/orgs", "repos_url": "https://api.github.com/users/vsl9/repos", "events_url": "https://api.github.com/users/vsl9/events{/privacy}", "received_events_url": "https://api.github.com/users/vsl9/received_events", "type": "User", "user_view_type": "public", "site_admin": false } ]

[ "@vsl9 Could you please describe how the next parameters are calculated?\r\n\r\n```\r\n# 20ms is duration of a timestep at output of the model\r\ntime_stride = 0.02\r\n\r\n# calibration offset for timestamps: 180 ms\r\noffset = -0.18\r\n```\r\nand what should be the numbers for `stt_en_citrinet_1024` model?", "@Vishaal-MK, have you solved the issue? If not, can you please share an audio file to reproduce the issue?\r\nI've tested the notebook on long conference talks (~40 minutes long), it works fine.\r\n@marlon-br, `time_stride` is a duration of timestep for logits tensor (output of ASR model). QuartzNet (or Jasper) takes a mel spectrogram (computed using overlapping windows with `stride=10ms`) as input. The first convolutional layer has `stride=2` (so the ASR model downsamples input by 2). That's why timestep of logits tensor has duration of `20ms`.\r\n`offset` parameter is used to compensate slight delay of model's output. It was set manually by tuning the offset and listening to audio (but, of course, it can be calibrated automatically if ground truth timestamps at word level are available).\r\nCitrinet models have `3` convolutional layers with `stride=2`, so the resulting `downsampling factor is 8` and `time_stride=0.08`. It also means that resolution for word timestamps is 4 times worse in comparison with QuartzNet/Jasper. Another difference is that Citrinet is word-piece based model (not character based model like QuartzNet/Jasper), thus code for prediction of space character timestamps becomes slightly more complicated (depending on Citrinet's tokens).", "@vsl9 **thanks for the answer**, now it is more clear. I am experimenting with `stt_en_citrinet_1024` model now because it gives best results on the files I have in test.\r\n\r\n**Could you please confirm or correct my logic?** (I use [this wav file](https://drive.google.com/file/d/1iRzjd4Q-3i0olAUVg_rZUmoInD9vFoot/view?usp=sharing) as example)\r\n\r\nModel output is: \r\n`let me guess you're the kind of guy that ignores the rules because it makes you feel uncontroll am i right you're not wrong you think that's cute do you think it's cute six feet at all times you both know the rules`\r\n\r\nAnd by chars:\r\n```\r\nblank\r\nblank\r\nlet\r\nblank\r\nblank\r\nblank\r\nme\r\nblank\r\nguess\r\nblank\r\nblank\r\nblank\r\nblank\r\nyou\r\n##'\r\n##re\r\nblank\r\nthe\r\nblank\r\nkind\r\nblank\r\nblank\r\nof\r\nblank\r\ngu\r\n##y\r\nblank\r\nthat\r\nblank\r\ni\r\n##g\r\n##no\r\n##re\r\n##re\r\n##s\r\nblank\r\nthe\r\nru\r\nblank\r\n##les\r\n```\r\netc.\r\n\r\nIf we ignore `offset ` for now, does the next output for words look correct?\r\n```\r\nlet 0.16 0.24\r\nme 0.48 0.56\r\nguess 0.64 0.72\r\nyou're 1.04 1.28\r\nthe 1.36 1.44\r\nkind 1.52 1.6\r\nof 1.76 1.84\r\nguy 1.92 2.08\r\nthat 2.16 2.24\r\nignores 2.32 2.8000000000000003\r\nthe 2.88 2.96\r\nrules 2.96 3.2800000000000002\r\n```" ]

[]

[]

[ "This pull request **introduces 1 alert** and **fixes 1** when merging 7bc984d05a1cb0684428995fdc4135ef3b9a8c36 into 8f7c18973bf92f30679c0f4416412d629bfc7bff - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-0ebacc9e5c231176025a17986eb8897d98fa1097)\n\n**new alerts:**\n\n* 1 for Unnecessary delete statement in function\n\n**fixed alerts:**\n\n* 1 for Unused import", "This pull request **fixes 1 alert** when merging 3fba4784f3f42e6202037bcbdb7c7a06dc8a4a99 into 8f7c18973bf92f30679c0f4416412d629bfc7bff - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-983adf1d405bed3dd4227058e8a8c9047c3fbe26)\n\n**fixed alerts:**\n\n* 1 for Unused import", "This pull request **fixes 1 alert** when merging 92d986b0082037ea9361fae554bbac85242cfb9c into 8f7c18973bf92f30679c0f4416412d629bfc7bff - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-551286c4266b331d1384c6590727e7dd136d54eb)\n\n**fixed alerts:**\n\n* 1 for Unused import", "This pull request **fixes 1 alert** when merging 1b7abb9f53e57c7996beea694942c80c55e6f282 into 8f7c18973bf92f30679c0f4416412d629bfc7bff - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-7e37fc8f694b12466dd09daf59e11e66e4c7fc95)\n\n**fixed alerts:**\n\n* 1 for Unused import", "Can we get this merged?", "I'm not sure what we're trying to accomplish with this PR. As @titu1994 has said, we can already restore weights and finetune in NeMo with .nemo files. Or to resume training, we use PTL.\r\n\r\nCan you please make it more clear in the README what feature is being added and how to use it?\r\n\r\nAlso, this PR is adding a new feature and should not be going into RC1. RC1 PRs are only for bug fixes and documentation.\r\n\r\nPlease rebase to main." ]

[]

[]

[ "Can I revert this? Our main branch is failing due to these DALI tests", "@blisc looking at it on main\r\n" ]

[]

[]

[]

[]

[]

[ "looks good to me. Please remove \"Draft\" status when you are done.\r\n" ]

[ { "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" } ]

[]

[ "What resource manager are you using? So far we tested multi-node in a SLURM environment only", "I am using google cloud vm without any resource manager.", "Doesn't GCP have at maximum just 8 GPUs? I'm not familiar with any one click solution to launching a 32 gpu (8 nodes, 4 GPUs per node) cluster on GCP - please point to such a config.\r\n\r\nIf you want to use just 8 GPUs, your trainer config would be \r\n\r\ntrainer = pl.Trainer(gpus=8, num_nodes=1, accelerator='ddp',max_epochs=200,amp_level='O1', precision=16, ...)\r\n\r\n", "in GCP you can spin up a vm with 8 nodes, and 4 GPUs \r\n\r\nI am using this command \r\n> `trainer = pl.Trainer(gpus=4, num_nodes=8, accelerator='ddp',max_epochs=200,amp_level='O1', precision=16,..)`\r\n\r\nI dont think this is a GCP issue but rather pytorch ligtning 1.7.1 issue where ddp hangs when initialising \r\n\r\n[https://github.com/PyTorchLightning/pytorch-lightning/issues/4612](url)\r\n\r\nIs there a way to update nemo to work with an older or newer version of pytorch?", "@sciai-ai just like workaround to fix your problem: \r\n```\r\npip uninstall -y torch pytorch-lightning torchtext torchaudio\r\npip install torchtext==0.8.0 torch==1.7.1 torchaudio==0.7.2 pytorch-lightning==1.2.5\r\n```\r\n\r\nI suppose problem related to https://github.com/PyTorchLightning/pytorch-lightning/issues/5865, https://github.com/PyTorchLightning/pytorch-lightning/issues/4612 and https://github.com/PyTorchLightning/pytorch-lightning/issues/6569", "@sciai-ai can you please try 1.0.0 to see if this issue still persists?", "@ericharper I noticed you re-opened this issue. We're still seeing this behavior in identical fashion, but also on CLI. Double checking, shall we re-open this one? ", "Feel free to use this issue. Please update with the code that was run and the log files.", "Great. Will do.", "@ericharper We have it resolved! By submitting the batch job in Slurm, we were able to get multi-GPU training working. We confirmed that the trainer was configured in the YAML file. " ]

[]

[]

[ "Doesn't have to be this PR, but I think the contents of the README would be a great addition to the NeMo NLP Docs.\r\n" ]

[]

[]

[ "Thanks for this issue @Omarnabk! We moved the Label Studio ML backend to a new repo, so the import path needs to be updated, and you'll also need to install the ML backend in addition to Label Studio.\r\n\r\na PR to fix this in the Jupyter notebook is coming soon, but in the meantime you can do the following:\r\n1. After you install label studio, install the ML backend: `pip install git+git://github.com/heartexlabs/label-studio-ml-backend.git#egg=label-studio-ml` (soon there will be a proper label-studio-ml pip package but you can use this command in the meantime).\r\n2. update the import statement to import `from label_studio_ml` instead of `from label_studio.ml`\r\n\r\nlet us know if you have any more questions!\r\n", "https://github.com/NVIDIA/NeMo/pull/1989 should fix it." ]

[ { "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": 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" } ]

[]

[ "Please start using rc1 version, we are going to release it soon\r\n```\r\n python -m pip install git+https://github.com/NVIDIA/[email protected]#egg=nemo_toolkit[all]\r\n```" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "This pull request **introduces 4 alerts** when merging 146952ca71c98f2274c49f2472a080f786157e45 into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-f5bf2e88dcb9955a61864b2cba2d1fc2e6b13b3b)\n\n**new alerts:**\n\n* 2 for Variable defined multiple times\n* 1 for Wrong name for an argument in a call\n* 1 for Unused import", "This pull request **introduces 2 alerts** when merging 7b13704b3c80a645de62d7d8fd4deabd644ad0d8 into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-2d3f41de390ac4045d4332e16f5fe044c0ca63af)\n\n**new alerts:**\n\n* 1 for Wrong name for an argument in a call\n* 1 for Unused import", "This pull request **introduces 1 alert** when merging 9743c99d762b95fa230b8c047fda37c9d10b8d7a into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-0a3215727fef1e5e725e6276711ec435e66f42d1)\n\n**new alerts:**\n\n* 1 for Wrong name for an argument in a call", "This pull request **introduces 1 alert** when merging fca0fc66c678005af01de5970b44afd3a64c499f into 480bbc39b8c35f2f5ff0e2a2b61f4e2526de28eb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-dd7a2c4804b8bbe025bce68e9d2571d5833d00e6)\n\n**new alerts:**\n\n* 1 for Wrong name for an argument in a call", "This pull request **introduces 1 alert** when merging cdc88bc431d617643671602e4ec33dd2ce68421c into 480bbc39b8c35f2f5ff0e2a2b61f4e2526de28eb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-3e6422c6ea4187dd1b86d9b399253ff4109ee7e1)\n\n**new alerts:**\n\n* 1 for Wrong name for an argument in a call", "This pull request **introduces 1 alert** when merging 842012cb6f2bb7988a5035fad7e98ab5c8b54ec6 into 480bbc39b8c35f2f5ff0e2a2b61f4e2526de28eb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-38a4773fbc26e6c792d7ee14e9370cc2eae2f5fd)\n\n**new alerts:**\n\n* 1 for Wrong name for an argument in a call" ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "@okuchaiev , small enhancement to the quantization script.\r\n\r\nI'll file another PR to add README later" ]

[]

[]

[]

[]

[]

[ "Can you try installing from main branch the most recent version?", "Thank you for your prompt response! \r\nI've tried to installed it like this: \r\napt-get update && apt-get install -y libsndfile1 ffmpeg\r\npip install Cython\r\npip install nemo_toolkit[all]==1.0.0b3\r\n\r\nand I got another error related to writing the onnx file. \r\n\r\n\r\n\r\nI'm not sure if this is what you mean by installing from main but I've tried,also, installing from the main git branch and I got an error in the nemo import related to pyramid installation: \r\n from pyramid.session import UnencryptedCookieSessionFactoryConfig ImportError: cannot import name 'UnencryptedCookieSessionFactoryConfig' from 'pyramid.session' (unknown location) \r\n \r\n Thank you again. \r\n", "Hello again,\r\nI think I found the issue, it's related to the path. \r\nI did as u said and It worked fine! \r\nThank you for your time. " ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "This pull request **introduces 2 alerts** and **fixes 2** when merging b07c408dc655861e9fbfee59e04050c384b155ae into cd194c88d7581e684a5ad6a0d2031a6b68300969 - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-693d8acbfaccfaceca6d6d47d3c4d90daddf0595)\n\n**new alerts:**\n\n* 2 for Unused import\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging d8c01c8e9d1f658331625ece3546ed378f437f65 into c51685e03f52d3428d19b7edccc1bbd0da5d8edb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-444fa37ca1051604057b75abe005d9d56912b647)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging 3a743b27a22859df175db4cd5fe4081cbf7991b8 into c51685e03f52d3428d19b7edccc1bbd0da5d8edb - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-c31f6a43a14ad9264be6e8499ceba3e4bcb3ff8e)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "@titu1994 Can you take a look at the asr datalayer and feature preprocessor changes?", "This pull request **fixes 2 alerts** when merging ba98c826cef0ba3474ac754616eb5994d5ebfb90 into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-f388eb5a7bd2d246c9eedd416da4c8e4468952de)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **introduces 1 alert** and **fixes 2** when merging 15f23537a2f9d69dea4c0bd060dc0b6fb6d37d67 into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-84f03f38d93d8787c91791ba0164c3ec7ca234a2)\n\n**new alerts:**\n\n* 1 for Unnecessary pass\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging 9cfc863397b2bf76b1ad052ff63897050746cabe into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-68b75ca060b278581e25fcaf74cd8cb507986d26)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging c1a21e52e25b204692e2e8cd7ce0f8563105c9ac into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-8ffa4f91de1621c29e7b0d0a3097cff7e5d6b28e)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging 808f91b3a43a4a9efee94b9e43f0d0c5241d41ed into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-2131f1b6061885f4d7304bba58cd30706fde73e1)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging 47da2edc561674eebe9674c41c72ee363b33b981 into d79d5d3ed06d86d24b82214363755174313d6d9d - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-b86e4ec6b3afdcb630b644aff86991255154209d)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging 38d7546e4462b458fbee19bb2056a99dc22c2ca1 into 16e0f1a4f486510ca4e50464d5bb5ee94cb766f2 - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-3e0d4148e38fb84aec471195e240c7f5657bd959)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging c750b741b5fe64e99f0538991b74cae08e2b595b into 02dc1af77c3c1edca766c95c30985deea5a7ca58 - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-775cb9cb1a1cc43796be4450d2deccc20441c8f9)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "This pull request **fixes 2 alerts** when merging 91596604d52f2fed3e8dba293709c07eeddf93ee into 02dc1af77c3c1edca766c95c30985deea5a7ca58 - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-2c08ecaaee888d774a24778d54e5b865214b2051)\n\n**fixed alerts:**\n\n* 2 for Mismatch in multiple assignment", "@blisc I use misc as ability to match sample from dataloader and file name from manifest.\r\n\r\nalso TalkNet uses load_audio=False flag. \r\n\r\nI suggest to discuss before removing it.", "> @blisc I use misc as ability to match sample from dataloader and file name from manifest.\r\n> \r\n> also TalkNet uses load_audio=False flag.\r\n> \r\n> I suggest to discuss before removing it.\r\n\r\nSeeing as how only TalkNet seems to use it and TalkNet already has it's own datalayer at https://github.com/NVIDIA/NeMo/blob/main/nemo/collections/asr/data/audio_to_text.py#L291. I will continue to remove these from the base class. Feel free to add such functionality to the derived class instead.", "@blisc \r\n\r\nIt is not a problem to implement it in derived class, but why do you necessary want to remove it from base class? Maybe will be another case for using load_audio=False." ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "@titu1994 Thanks! Updated. Will merge after the checks. " ]

[]

[]

[]

[]

[]

[ "@skyw any thoughts?", "It should be fixed in 21.03.\r\nWhat failed is the checker when calling `torch.onnx.export`. It needs opset13 support which comes with Pytorch 1.8 release. 21.02 was based on a dev version of 1.8.", "thanks! Close this question." ]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[ "Hi, unfortunately, at this moment we only support Linux (Ubuntu). " ]

[]

[ { "login": "fayejf", "id": 36722593, "node_id": "MDQ6VXNlcjM2NzIyNTkz", "avatar_url": "https://avatars.githubusercontent.com/u/36722593?v=4", "gravatar_id": "", "url": "https://api.github.com/users/fayejf", "html_url": "https://github.com/fayejf", "followers_url": "https://api.github.com/users/fayejf/followers", "following_url": "https://api.github.com/users/fayejf/following{/other_user}", "gists_url": "https://api.github.com/users/fayejf/gists{/gist_id}", "starred_url": "https://api.github.com/users/fayejf/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/fayejf/subscriptions", "organizations_url": "https://api.github.com/users/fayejf/orgs", "repos_url": "https://api.github.com/users/fayejf/repos", "events_url": "https://api.github.com/users/fayejf/events{/privacy}", "received_events_url": "https://api.github.com/users/fayejf/received_events", "type": "User", "user_view_type": "public", "site_admin": false } ]

[ "Hi thanks for your question! The offset in manifest (for all audio related manifest not limited to VAD) means the **starting point (second) of the audio snippet**. \r\nFor example` \"audio_a.wav, offset:0, duarion:10\"` means \"audio snippet of audio_a from 0 to 10 seconds\"; SImilarly \r\n`\"audio_a.wav, offset:5, duarion:5\" `means \"audio snippet of audio_a from 5 to 10 seconds\";\r\n", "Thank you so much! @fayejf \n\nI am closing the issue.\n\nRegards." ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "This pull request **introduces 2 alerts** when merging 82fa2be0844042c845af3cf43e7e5e2cf989eab0 into da48c997177f89a9e305e508ab27b15954044814 - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-87eabc38d29520a4b3fc1bcfa066c69c4774062f)\n\n**new alerts:**\n\n* 1 for Unnecessary pass\n* 1 for Unused import", "> \r\n> \r\n> This pull request **introduces 2 alerts** when merging [82fa2be](https://github.com/NVIDIA/NeMo/commit/82fa2be0844042c845af3cf43e7e5e2cf989eab0) into [da48c99](https://github.com/NVIDIA/NeMo/commit/da48c997177f89a9e305e508ab27b15954044814) - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-87eabc38d29520a4b3fc1bcfa066c69c4774062f)\r\n> \r\n> **new alerts:**\r\n> \r\n> * 1 for Unnecessary pass\r\n> \r\n> * 1 for Unused import\r\n\r\nSome LGTM warnings to address.", "This pull request **introduces 2 alerts** when merging 70625bf33217750fe1931d091ca44f4c8d305487 into 532c71b8c632bc9c0fc1096d7630a02da3bee00e - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-5af65e7126afc589c92131f898452b6562e14acf)\n\n**new alerts:**\n\n* 1 for Unnecessary pass\n* 1 for Unused import" ]

[ { "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" } ]

[]

[ "To save checkpoints at every epoch pass `+exp_manager.checkpoint_callback_params.save_top_k=num_epochs` while calling script or add through yaml file ", "nemo saves the last checkpoint, not the ones at the end of each epoch. It can also saves the top K best ones. You need to set K along with what metric to monitor and if best value is min or max. You may add it to your config file:\r\n\r\nexp_manager:\r\n checkpoint_callback_params:\r\n # in case of multiple validation sets, first one is used\r\n monitor: \"val_wer\"\r\n mode: \"min\"\r\n save_top_k: 3\r\n" ]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[]

[ "wrong branch. I applied this fix to rc", "This pull request **introduces 10 alerts** and **fixes 3** when merging 398409b424148af7b95a84fce097d7f053b6bca7 into 6594dd6dcd268cc7eeb1caf08a7b9ed18cd40556 - [view on LGTM.com](https://lgtm.com/projects/g/NVIDIA/NeMo/rev/pr-13346d8c5f9ee999c6b43b23568223cccbc61a20)\n\n**new alerts:**\n\n* 6 for Unused import\n* 2 for Modification of dictionary returned by locals\\(\\)\n* 1 for Signature mismatch in overriding method\n* 1 for Mismatch in multiple assignment\n\n**fixed alerts:**\n\n* 2 for Unused import\n* 1 for Explicit export is not defined" ]

[]

[]

[ "At the moment we do not test/support Arm architectures. Do you know which module need MKL? Are you able to install Pytorch on Arm without MKL?" ]