Datasets:

---

DanielPFlorian

/

Transformers-Github-Issues

like 0

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5920?src=pr&el=h1) Report\n> Merging [#5920](https://codecov.io/gh/huggingface/transformers/pull/5920?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/32883b310ba30d72e67bb2ebb5847888f03a90a8&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5920?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5920 +/- ##\n=======================================\n Coverage 78.51% 78.51% \n=======================================\n Files 146 146 \n Lines 26214 26214 \n=======================================\n Hits 20583 20583 \n Misses 5631 5631 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5920?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5920/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.19% <0.00%> (-0.30%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5920/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.46% <0.00%> (ø)` | |\n| [src/transformers/generation\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5920/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `97.11% <0.00%> (+0.28%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5920?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5920?src=pr&el=footer). Last update [32883b3...2579206](https://codecov.io/gh/huggingface/transformers/pull/5920?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5919?src=pr&el=h1) Report\n> Merging [#5919](https://codecov.io/gh/huggingface/transformers/pull/5919?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/4781afd045b4722e7f28347f1c4f42a56a4550e8&el=desc) will **decrease** coverage by `0.17%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5919?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5919 +/- ##\n==========================================\n- Coverage 78.69% 78.51% -0.18% \n==========================================\n Files 146 146 \n Lines 26214 26214 \n==========================================\n- Hits 20628 20581 -47 \n- Misses 5586 5633 +47 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5919?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5919/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5919/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5919/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5919?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5919?src=pr&el=footer). Last update [4781afd...303f0ac](https://codecov.io/gh/huggingface/transformers/pull/5919?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Gunna merge this.\r\nThe packed dataset is definitely a win, label smoothing less clear.\r\nTODO: figure out loss function mystery." ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1843244711, "node_id": "MDU6TGFiZWwxODQzMjQ0NzEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/New%20model", "name": "New model", "color": "fbca04", "default": false, "description": "" } ]

[]

[ "Hi guys, let us know how we can help and also kindly add @apoorv2904 to the author list.\r\n\r\nAlthough the model weights are nothing particularly useful we do provide them for our colab so let us know if they are needed and how to provide them.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "Don´t let it die. In my tests this is the best performing model so far!", "@patrickvonplaten @sgugger \r\n\r\nI could try to include on huggingface/transformers if there is an interest from the core team. But I would have to depend on https://github.com/idiap/fast-transformers as they created optimized cuda/cpu c++ versions of the proposed attention. A MR with this dependency would be accepted by Huggingface? ", "would love if this comes in!", "Hey @bratao,\r\n\r\nYes, we would definitely be interested in this model and would also be fine with an optional dependency of `https://github.com/idiap/fast-transformers` Also pinging @joeddav @TevenLeScao here (in case you guys are interested in helping with the integration). \r\n\r\nI would also be happy to help you with the model integration otherwise @bratao :-) ", "Great, I´m on it @patrickvonplaten \r\n\r\nI will work on this on my free time, As soon as I have something, I put it here the fork.\r\n\r\nIf anyone else want to help or speed it up, just talk to me using the email in my profile!", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "What happened to this model? It was not finally integrated right? :( @bratao @patrickvonplaten " ]

[]

[ { "login": "sshleifer", "id": 6045025, "node_id": "MDQ6VXNlcjYwNDUwMjU=", "avatar_url": "https://avatars.githubusercontent.com/u/6045025?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sshleifer", "html_url": "https://github.com/sshleifer", "followers_url": "https://api.github.com/users/sshleifer/followers", "following_url": "https://api.github.com/users/sshleifer/following{/other_user}", "gists_url": "https://api.github.com/users/sshleifer/gists{/gist_id}", "starred_url": "https://api.github.com/users/sshleifer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sshleifer/subscriptions", "organizations_url": "https://api.github.com/users/sshleifer/orgs", "repos_url": "https://api.github.com/users/sshleifer/repos", "events_url": "https://api.github.com/users/sshleifer/events{/privacy}", "received_events_url": "https://api.github.com/users/sshleifer/received_events", "type": "User", "site_admin": false } ]

[ "maybe @sshleifer has an idea", "I can help! \r\nOn fairseq master (as of 5/28/20), that class seems to no longer have a `decoder` attribute.\r\n\r\nI think you want to change the `roberta.model.decoder` references to `self.model.encoder`, but hard to know without seeing the `state_dict`/handling the model interactively.\r\n\r\nThe best way to debug is to either instantiate the fairseq model in jupyter/ipython or set a breakpoint and see what the attributes are.\r\n\r\nIf you are stuck, feel free to upload your `model.pt` to some cloud storage and I can give it a shot!\r\n\r\n", "Hey @sshleifer!\r\n\r\nI did what you suggested and it worked, thanks a lot. You have to replace all the references to `roberta.model.decoder` with `roberta.model.encoder` as the attributes were just renamed.\r\n\r\nOn the other hand, I can't figure out what happened to `roberta.args.num_classes` that is used for the classification_head flag, which makes it useless for now.\r\n\r\nI would gladly commit the fix but I'm not a powergit user, so I'll leave it to the pros.\r\n\r\nThanks again!\r\n\r\n***\r\n\r\nEdit : the error that comes up with the flag. \r\n\r\n```python\r\nTraceback (most recent call last):\r\n File \"converter.py\", line 179, in <module>\r\n args.roberta_checkpoint_path, args.pytorch_dump_folder_path, args.classification_head\r\n File \"converter.py\", line 62, in convert_roberta_checkpoint_to_pytorch\r\n config.num_labels = roberta.args.num_classes\r\nAttributeError: 'Namespace' object has no attribute 'num_classes'\r\n```\r\n", "I think `num_classes` will be like something like\r\n\r\n`roberta.model.classification_heads[some_key].out_proj.weight.shape[0]`\r\nThere is likely only one possible key.", "I just checked with a model fine-tuned on MNLI and the key is classification_heads['mnli'], is this what you expected?", "sounds right! Don't lose that head!", "Hey @sshleifer,\r\nSorry if that's not the right place to ask but I couldn't find an answer to that question anywhere: is there a script like this one to convert a model.pt trained on gpu to a model.bin ? or should this script works both for cpu and gpu models ?\r\nThanks!", "Did our library save the `model.pt`?\r\n\r\nThe filenames don't really matter if the contents of the file are a `state_dict`.\r\nSo it may be as simple as, from the terminal,\r\n```bash\r\nmv model.pt pytorch_model.bin\r\n```\r\nThe library doesn't care if a `state_dict` was saved on gpu or cpu.\r\n\r\nIf that fails try to run `torch.load('model.pt', map_location='cpu')` in an interactive environment and see if it's a state dict.\r\n\r\n", "Thanks for the answer @sshleifer! HuggingFace was working well, it was my nvidia apex installation that was broken and returned errors in fast-bert that confused me. All works well now!" ]

[]

[]

[ "👍 " ]

[]

[]

[ "Let's consolidate the discussion to #5895 .\r\nDefinitely an issue!" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5914?src=pr&el=h1) Report\n> Merging [#5914](https://codecov.io/gh/huggingface/transformers/pull/5914?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/f19751117d54a4dd677c614f6e400a7ee49b3f24&el=desc) will **increase** coverage by `0.02%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5914?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5914 +/- ##\n==========================================\n+ Coverage 78.49% 78.51% +0.02% \n==========================================\n Files 146 146 \n Lines 26214 26214 \n==========================================\n+ Hits 20577 20583 +6 \n+ Misses 5637 5631 -6 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5914?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/generation\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5914/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `97.11% <0.00%> (+0.28%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5914/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5914/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+1.00%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5914?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5914?src=pr&el=footer). Last update [f197511...8901dff](https://codecov.io/gh/huggingface/transformers/pull/5914?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5913?src=pr&el=h1) Report\n> :exclamation: No coverage uploaded for pull request base (`master@f197511`). [Click here to learn what that means](https://docs.codecov.io/docs/error-reference#section-missing-base-commit).\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5913?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5913 +/- ##\n=========================================\n Coverage ? 78.69% \n=========================================\n Files ? 146 \n Lines ? 26214 \n Branches ? 0 \n=========================================\n Hits ? 20629 \n Misses ? 5585 \n Partials ? 0 \n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5913?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5913?src=pr&el=footer). Last update [f197511...547c2ad](https://codecov.io/gh/huggingface/transformers/pull/5913?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5912?src=pr&el=h1) Report\n> :exclamation: No coverage uploaded for pull request base (`master@f197511`). [Click here to learn what that means](https://docs.codecov.io/docs/error-reference#section-missing-base-commit).\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5912?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5912 +/- ##\n=========================================\n Coverage ? 78.46% \n=========================================\n Files ? 146 \n Lines ? 26214 \n Branches ? 0 \n=========================================\n Hits ? 20569 \n Misses ? 5645 \n Partials ? 0 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5912?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/configuration\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5912/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3hsbmV0LnB5) | `94.33% <ø> (ø)` | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5912?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5912?src=pr&el=footer). Last update [f197511...b3c3a0e](https://codecov.io/gh/huggingface/transformers/pull/5912?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1841528858, "node_id": "MDU6TGFiZWwxODQxNTI4ODU4", "url": "https://api.github.com/repos/huggingface/transformers/labels/Summarization", "name": "Summarization", "color": "b6f97f", "default": false, "description": "" } ]

[ { "login": "sshleifer", "id": 6045025, "node_id": "MDQ6VXNlcjYwNDUwMjU=", "avatar_url": "https://avatars.githubusercontent.com/u/6045025?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sshleifer", "html_url": "https://github.com/sshleifer", "followers_url": "https://api.github.com/users/sshleifer/followers", "following_url": "https://api.github.com/users/sshleifer/following{/other_user}", "gists_url": "https://api.github.com/users/sshleifer/gists{/gist_id}", "starred_url": "https://api.github.com/users/sshleifer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sshleifer/subscriptions", "organizations_url": "https://api.github.com/users/sshleifer/orgs", "repos_url": "https://api.github.com/users/sshleifer/repos", "events_url": "https://api.github.com/users/sshleifer/events{/privacy}", "received_events_url": "https://api.github.com/users/sshleifer/received_events", "type": "User", "site_admin": false } ]

[ "Hi @sshleifer, I will continue to push code to this PR and make it runnable asap.", "@sshleifer The code has been uploaded. (1) The test is runnable in TF2 and loads AESLC checkpoints successfully with correct outputs. (2) Code of models and layers (including decoding, beam search) are all in a single file which may look messy (sorry). (3) Most code is simply copied and pasted (then converted to TF2) from the original PEGASUS repo so you may refer to the original repo for clearer code if necessary. \r\n\r\nI think you can start from here. Please let me know if I can help further.", "@JingqingZ I'm gunna add torch first in #6340 (you will be a PR co-author). And then come back here to finish TF. No action needed from you, just an update.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null }, { "id": 1771187924, "node_id": "MDU6TGFiZWwxNzcxMTg3OTI0", "url": "https://api.github.com/repos/huggingface/transformers/labels/Core:%20Pipeline", "name": "Core: Pipeline", "color": "FF7066", "default": false, "description": "Internals of the library; Pipeline." } ]

[ { "login": "mfuntowicz", "id": 2241520, "node_id": "MDQ6VXNlcjIyNDE1MjA=", "avatar_url": "https://avatars.githubusercontent.com/u/2241520?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mfuntowicz", "html_url": "https://github.com/mfuntowicz", "followers_url": "https://api.github.com/users/mfuntowicz/followers", "following_url": "https://api.github.com/users/mfuntowicz/following{/other_user}", "gists_url": "https://api.github.com/users/mfuntowicz/gists{/gist_id}", "starred_url": "https://api.github.com/users/mfuntowicz/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/mfuntowicz/subscriptions", "organizations_url": "https://api.github.com/users/mfuntowicz/orgs", "repos_url": "https://api.github.com/users/mfuntowicz/repos", "events_url": "https://api.github.com/users/mfuntowicz/events{/privacy}", "received_events_url": "https://api.github.com/users/mfuntowicz/received_events", "type": "User", "site_admin": false } ]

[ "I also have this -- and related issues on v3.0.0, 3.0.2\r\n\r\non v3.0.0 I get Key Error for sequence terminal tokens -- I think this may be to do with removing special tokens but not masking their positions or due to a truncation error -- may be truncating one too few?\r\non v3.0.2 I get Key Error for tokens at the beginning of the sequence (Key Error : 0) -- may be an issue to do with the [CLS] token and/or tokens falling in the question span.\r\n\r\nI think the trick is to use the attention_mask * token_type_ids * start/end_scores. This will set the logits for all tokens outside the answer to 0 and can be done easily on batch tensors/GPU. I will see if I can put together a pull request.\r\n\r\nPlatform: Mac OS Catalina, GCP linux cuda 10.1\r\nPython version: 3.6.8\r\nPyTorch version (GPU?): 1.5.1, GPU\r\nUsing GPU in script?: Yes\r\nUsing distributed or parallel set-up in script?: No", "@brandenchan @jusjosgra thanks for reporting the issue and the steps to reproduce.\r\n\r\nWe did have an issue but it should have been fixed on master branch.\r\nWhen running the snippet you provided on master I get the following: \r\n\r\n`{'score': 0.05008925125002861, 'start': 679, 'end': 697, 'answer': 'Lord Eddard Stark,'}`\r\n\r\nIf you can checkout the master branch and give it a try to make sure it works on your side too.\r\n\r\n_If everything work as expected: I need to checkout with the team when we can do a maintenance release._", "I still have an error. It appears to be for an example where start and end are both predicted as 0 (a null answer).\r\nIn this case the valid results have been filtered to document characters only and so an index for 0 doesnt exist in feature.token_to_orig_map (the first index in my instance is 12, 0 doesnt exist).\r\n\r\nSo there needs to be a method to handle when the predicted span occurs outside the filtered feature dict. You could return a null object since these cases represent no answer or you could return the best answer found inside the valid span (i.e. mask the logits for non document tokens when getting the max value).", "Here is one solution:\r\n\r\noriginal code\r\n```python\r\n # Normalize logits and spans to retrieve the answer\r\n start_ = np.exp(start_ - np.log(np.sum(np.exp(start_), axis=-1, keepdims=True)))\r\n end_ = np.exp(end_ - np.log(np.sum(np.exp(end_), axis=-1, keepdims=True)))\r\n\r\n if kwargs[\"handle_impossible_answer\"]:\r\n min_null_score = min(min_null_score, (start_[0] * end_[0]).item())\r\n\r\n starts, ends, scores = self.decode(start_, end_, kwargs[\"topk\"], kwargs[\"max_answer_len\"])\r\n char_to_word = np.array(example.char_to_word_offset)\r\n\r\n # Convert the answer (tokens) back to the original text\r\n answers += [\r\n {\r\n \"score\": score.item(),\r\n \"start\": np.where(char_to_word == feature.token_to_orig_map[s])[0][0].item(),\r\n \"end\": np.where(char_to_word == feature.token_to_orig_map[e])[0][-1].item(),\r\n \"answer\": \" \".join(\r\n example.doc_tokens[feature.token_to_orig_map[s] : feature.token_to_orig_map[e] + 1] \r\n ),\r\n }\r\n for s, e, score in zip(starts, ends, scores)\r\n```\r\nmy update\r\n```python\r\n # Normalize logits and spans to retrieve the answer\r\n start_ = np.exp(start_ - np.log(np.sum(np.exp(start_), axis=-1, keepdims=True)))\r\n end_ = np.exp(end_ - np.log(np.sum(np.exp(end_), axis=-1, keepdims=True)))\r\n\r\n if kwargs[\"handle_impossible_answer\"]:\r\n min_null_score = min(min_null_score, (start_[0] * end_[0]).item())\r\n\r\n starts, ends, scores = self.decode(start_, end_, kwargs[\"topk\"], kwargs[\"max_answer_len\"])\r\n char_to_word = np.array(example.char_to_word_offset)\r\n\r\n # Convert the answer (tokens) back to the original text\r\n answers += [\r\n {\r\n \"score\": score.item(),\r\n \"start\": np.where(char_to_word == feature.token_to_orig_map[s])[0][0].item(),\r\n \"end\": np.where(char_to_word == feature.token_to_orig_map[e])[0][-1].item(),\r\n \"answer\": \" \".join(\r\n example.doc_tokens[feature.token_to_orig_map[s] : feature.token_to_orig_map[e] + 1]\r\n ),\r\n }\r\n if s in feature.token_to_orig_map and e in feature.token_to_orig_map # this condition handles the case when answer spans are outside the valid token range.\r\n else {\"score\": min_null_score, \"start\": 0, \"end\": 0, \"answer\": \"\"}\r\n for s, e, score in zip(starts, ends, scores)\r\n```\r\n\r\npersonally I would rather get the best valid span (max over the masked logits) rather than an error/null answer. This might be a more useful use of \"handle impossible answer\". Returning null answers might be the best default behaviour and \"best valid span\" might be a good alternative although this would involve a significant refactor of decode to mask the logits appropriately.", "I think there is another bug in the decode function (although I may be misunderstanding).\r\nYou compute negative log likelihoods as probabilities but in order to mask items you set them to 0. These items need to be set to a high negative number (e.g. -99) as valid values span zero.\r\nfor example:\r\n```python\r\ndef decode(...):\r\n\r\n...\r\n start_, end_ = (\r\n start_ - np.abs(-99 * np.array(feature.p_mask)),\r\n end_ - np.abs(-99 * np.array(feature.p_mask)),\r\n )\r\n\r\n # Mask CLS\r\n start_[0] = end_[0] = -99.\r\n```", "@mfuntowicz \r\nGot the same key error zero issue. Above code fixed it", "@jusjosgra Thanks for the provided solution here, do you want to submit a PR with your fix? Tag myself and @LysandreJik as reviewers and we'll merge it into master. \r\n\r\nOtherwise I'll do 😄.\r\n\r\nThanks!", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5909?src=pr&el=h1) Report\n> Merging [#5909](https://codecov.io/gh/huggingface/transformers/pull/5909?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/82dd96cae74797be0c1d330566df7f929214b278&el=desc) will **decrease** coverage by `0.13%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5909?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5909 +/- ##\n==========================================\n- Coverage 78.49% 78.35% -0.14% \n==========================================\n Files 146 146 \n Lines 26210 26211 +1 \n==========================================\n- Hits 20573 20538 -35 \n- Misses 5637 5673 +36 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5909?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/tokenization\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5909/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHMucHk=) | `90.00% <100.00%> (+0.04%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5909/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5909/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5909/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5909/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `95.53% <0.00%> (+69.51%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5909?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5909?src=pr&el=footer). Last update [82dd96c...c90f92b](https://codecov.io/gh/huggingface/transformers/pull/5909?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "Hi @krannnn , `DataCollatorForPermutationLanguageModeling` is added after 3.0, you will need to install from source if you want to run examples", "Hi @patil-suraj , out of curiosity, how do you install it? What do you mean by you will need to install from source?" ]

[]

[]

[ "Hi @vyaslkv , what is your transformers version ? I tried this from master branch and it worked.", "'3.0.2' Thanks for the quick reply, which version worked for you", "I also tried with the master branch by uninstalling the transformers and then using the repo\r\n", "worked!! Thanks @patil-suraj :)", "can you give me an example how to use this a short one", "<img width=\"1138\" alt=\"Screenshot 2020-07-20 at 7 07 12 PM\" src=\"https://user-images.githubusercontent.com/33617789/87943880-46e07900-cabc-11ea-85e1-c98e8462ac0f.png\">\r\n", "ValueError: You have to specify either decoder_input_ids or decoder_inputs_embeds\r\n", "You should use `.generate` method for generation. \r\n```python3\r\nmodel.generate(input_ids=inputs['input_ids'], attention_mask=inputs['attention_mask'])\r\n```\r\n\r\npinging @mrm8488 for exact `generate` arguments.", "<img width=\"1166\" alt=\"Screenshot 2020-07-20 at 8 55 41 PM\" src=\"https://user-images.githubusercontent.com/33617789/87955457-68952c80-cacb-11ea-889d-95de35eb4cc0.png\">\r\n", "is this correct? generating the text out of it the tokenizer decode part?", "I will write a in the model card the exact arguments to use it ASAP and post it here.", "Also @vyaslkv it would nice if you post code instead of screenshot so we can copy paste and try the code faster ;)", "```sh\r\ngit clone https://github.com/huggingface/transformers.git\r\npip install ./transformers\r\n```\r\n```python\r\nfrom transformers import AutoModelWithLMHead, AutoTokenizer\r\n\r\ntokenizer = AutoTokenizer.from_pretrained(\"mrm8488/t5-base-finetuned-wikiSQL-sql-to-en\")\r\nmodel = AutoModelWithLMHead.from_pretrained(\"mrm8488/t5-base-finetuned-wikiSQL-sql-to-en\")\r\n\r\ndef get_explanation(query):\r\n input_text = \"translante Sql to English: %s </s>\" % query\r\n features = tokenizer([input_text], return_tensors='pt')\r\n\r\n output = model.generate(input_ids=features['input_ids'], \r\n attention_mask=features['attention_mask'])\r\n \r\n return tokenizer.decode(output[0])\r\n\r\nquery = \"SELECT COUNT Params form model where location=HF-Hub\"\r\n\r\nget_explanation(query)\r\n```", "@mrm8488 can you also make something like nlp to sql", "@mrm8488 it doesn't work for longer queries or is there any particular format I should give", "> @mrm8488 can you also make something like nlp to sql\r\n\r\nI already did it. ", "> @mrm8488 it doesn't work for longer queries or is there any particular format I should give\r\n\r\nThe max number of f tokens is 128 but I am currently working on the 256 version", "@mrm8488 can you send me the link of nlp to sql", "https://huggingface.co/mrm8488/t5-base-finetuned-wikiSQL-sql-to-en", "Model card: https://github.com/huggingface/transformers/commit/61e8be9940096ce763872c8d1479965511d0b472", "@mrm8488 I think this is sql to English not English to SQL correct me If I am wrong", "English to SQL is t5-base-finetuned-wikiSQL or English to SQL is t5-small-finetuned-wikiSQL", "https://github.com/mrm8488/shared_colab_notebooks/blob/master/T5_finetuned_wikiSQL_demo.ipynb", "The main issue is solved, closing this for now. Feel free to re-open if the problem persists." ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "@kevin-yauris I had a similar problem with retraining fine-tuned model. Here is what I have done. \r\n\r\nDo not pass config parameter when creating your model with `from_pretrained()`. Just initialize it with something like this: \r\n```\r\nmodel = AutoModelForTokenClassification.from_pretrained(\r\n model_name,\r\n from_tf=bool(\".ckpt\" in model_name),\r\n cache_dir=cache_dir,\r\n )\r\n```\r\n\r\nThen, you will need to change the last layer in the model. I was using PyTorch to fine-tune a blank model initially, therefore these steps will work for PyTorch models.\r\n\r\nThe last layer in the `TokenClassification` model is called `classification`. It is simply a linear layer, so you can create new one with the correct shape and randomized weights, and assign it to the initialized model `classification` layer. Say before my layer was (768,5) with the initial 5 classes, and now I want 9 so make a final layer with shape (768,9).\r\n\r\n```\r\n#reinitiallize the final classification layer to match new number of labels\r\nmodel.classifier = torch.nn.Linear(in_features=model.classifier.in_features, out_features=config.num_labels, bias=True)\r\nmodel.config = config\r\nmodel.num_labels = config.num_labels\r\n```\r\nSince to initialize the model you will be loading config file from the fine-tuned model, you also want to change model config to your current one with the new classes, so the correct config gets exported after your model is trained. Also you will want to modify `num_labels` of the model, since that was initialized with the old number of classes in the old config.\r\n", "Hi @TarasPriadka thank you for answering\r\nI also to the same thing that you did but with Tensorflow https://discuss.huggingface.co/t/retrain-reuse-fine-tuned-models-on-different-set-of-labels/346/5?u=kevinyauris.\r\nI forget about the model.num_labels tho, thank you for the catch.\r\nI wonder if there is another way to do it since if we replace the last layer with randomized weights we can't use the learned weight for some labels that are the same with previous labels/classes.\r\nLet's say there are 3 classes in the initial model and now I want to add 1 more class but the other classes are the same. If we use this method all weights for the last layer are randomized and we need to fine-tune the model with all the data again instead of just give train data for the new class.", "@kevin-yauris I've seen your forum post since I've been looking for a solution. My idea is that you already have an `id2label` and `label2id` in the model, so you could find if the incoming labels are already trained in the fine-tuned model. You find those which are not and you add randomized layers for them. However I am not sure how you can take a layer, and then just add randomized rows to it.", "Hi @TarasPriadka ,\r\n\r\nThanks for sharing the solution. I followed the same steps which solved this error \r\n\r\nRuntimeError: Error(s) in loading state_dict for BertForTokenClassification:\r\nsize mismatch for classifier.weight: copying a param with shape torch.Size([17, 1024]) from checkpoint, the shape in current model is torch.Size([13, 1024]).\r\nsize mismatch for classifier.bias: copying a param with shape torch.Size([17]) from checkpoint, the shape in current model is torch.Size([13]).\r\n\r\n\r\nbut now, it throws another error -\r\n\r\n model_path=model_args.model_name_or_path if os.path.isdir(model_args.model_name_or_path) else None\r\n File \"/usr/local/lib/python3.7/site-packages/transformers/trainer.py\", line 514, in train\r\n optimizer.step()\r\n File \"/usr/local/lib/python3.7/site-packages/torch/optim/lr_scheduler.py\", line 67, in wrapper\r\n return wrapped(*args, **kwargs)\r\n File \"/usr/local/lib/python3.7/site-packages/transformers/optimization.py\", line 244, in step\r\n exp_avg.mul_(beta1).add_(grad, alpha=1.0 - beta1)\r\nRuntimeError: The size of tensor a (17) must match the size of tensor b (13) at non-singleton dimension 0\r\n\r\n\r\nI had first trained the model on a dataset having 17 classes and now I want to transfer this model to the 2nd dataset which has 13 labels. \r\n\r\nDo we have to change the num_labels for any other layer ?\r\n\r\nThanks,", "@vikas95 I am not sure, but just changing the model's `num_labels` seemed to be working for me. However, I was scaling up labels, not reducing them. I would assume that it should have the same solution. Maybe you can share your model's layers before and after applying my fix with `print(model)`, and we can take a look into a possible solution.", "Hi @TarasPriadka ,\r\n\r\nThanks for the suggestion, I printed the model after loading the checkpoint and after updating the classification layer. \r\n\r\nThe classification layer output dimension is changing from your mentioned solution i.e.,\r\n\r\ninitially after loading the checkpoint, model = AutoModelForTokenClassification.from_pretrained(\r\n model_args.model_name_or_path,\r\n from_tf=bool(\".ckpt\" in model_args.model_name_or_path),\r\n cache_dir=model_args.cache_dir,\r\n )\r\nThe classification layer size is - (classifier): Linear(in_features=1024, out_features=17, bias=True)\r\n\r\nand after updating the classification layer, the size is - (classifier): Linear(in_features=1024, out_features=13, bias=True)\r\n\r\nThe rest of the layers look similar but I am still not sure why its throwing the previously mentioned error. \r\n\r\n-Vikas", "@vikas95, so the shape of the model is fine. The issue is in `exp_avg.mul_(beta1).add_(grad, alpha=1.0 - beta1)`. The exception highlights that `exp_avg` is of size 17 and its trying to add `grad` which is 13. So the problem is in `exp_avg`, since it wasn't updated along with everything else. Can you share the whole chunk of code where you initialize the model, trainer, etc?", "Hi @TarasPriadka ,\r\n\r\nHere is the part where I initialize the model (which is from run_ner.py (https://github.com/huggingface/transformers/blob/6e8a38568eb874f31eb49c42285c3a634fca12e7/examples/token-classification/run_ner.py#L158)) -\r\n\r\n labels = get_labels(data_args.labels)\r\n label_map = {i: label for i, label in enumerate(labels)}\r\n num_labels = len(labels)\r\n \r\n config = AutoConfig.from_pretrained(\r\n model_args.config_name if model_args.config_name else model_args.model_name_or_path,\r\n num_labels=num_labels,\r\n id2label=label_map,\r\n label2id={label: i for i, label in enumerate(labels)},\r\n cache_dir=model_args.cache_dir,\r\n )\r\n\r\n tokenizer = AutoTokenizer.from_pretrained(\r\n model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path,\r\n cache_dir=model_args.cache_dir,\r\n use_fast=model_args.use_fast,\r\n )\r\n model = AutoModelForTokenClassification.from_pretrained(\r\n model_args.model_name_or_path,\r\n from_tf=bool(\".ckpt\" in model_args.model_name_or_path),\r\n cache_dir=model_args.cache_dir,\r\n )\r\n\r\n model.classifier = torch.nn.Linear(in_features=model.classifier.in_features, out_features=config.num_labels, bias=True)\r\n model.config = config \r\n model.num_labels = config.num_labels\r\n", "@vikas95 Can you also share the trainer code", "@TarasPriadka - Its the same as in run_ner.py \r\nI haven't changed any other part of the code. \r\n", "@vikas95 can you check if in your model folder you have this file `optimizer.pt` and `scheduler.pt`\r\n", "@TarasPriadka ,\r\n\r\nThanks for the help, I was giving a specific checkpoint directory as the model path i.e., \"datasetA_model/checkpoint-6000/\" which had both optimizer.pt and scheduler.pt\r\n\r\nbut then I changed the model path to just \"datasetA_model/\" and it works fine with no errors. \r\nI am guessing that if I just give the \"datasetA_model/\" as model path then it would select the highest checkpoint ? \r\n\r\nAnyway, thanks a lot for looking at the problem and for all the quick responses and help 😬\r\n", "@vikas95 This was a great deal of fun. When you are running \r\n```\r\ntrainer.train(\r\n model_path=model_name if os.path.isdir(model_name) else None\r\n )\r\n```\r\ntrainer loads in those files, and initializes Adam optimizer with them. Optimizer breaks since you are changing the shape of the output layer, but optimizer was initialized with the other shape. What you can do, is either delete that file, or just run `trainer.train()` without parameters.", "Cool, this makes sense. \r\nThanks again for the explanation. I was actually trying with just trainer.train() for last 30 minutes and it works fine. \r\n\r\nThanks again for all the help and explanations. ", "Does anyone know what is the alternative method in Pytorch?", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "Thank you all for this solution, it worked for me but I encountered another problem while training getting this error:\r\n`ValueError: Expected input batch_size (3200) to match target batch_size (32).`\r\nMy batch size is indeed 32. If I change it to other value e.g. 16 the error will be:\r\n`ValueError: Expected input batch_size (1600) to match target batch_size (16).` it always multiplies by 100 which is a weird behavior because when trying to run the exact same code but on an original pre-trained model (in my case is `xlm-roberta-base`), to fine-tune it on classification task, it works just fine.\r\n\r\n\r\nHere is my code:\r\n```\r\nconfig = XLMRobertaConfig.from_pretrained(\"../xlm-roberta_domains_classifier/model\", output_hidden_states=True, \r\n num_labels=len(train_df.label.unique()),\r\n id2label=id2label, label2id=label2id)\r\nmodel = XLMRobertaForSequenceClassification.from_pretrained('../xlm-roberta_domains_classifier/model')\r\nmodel.cuda()\r\nmodel.classifier = torch.nn.Linear(in_features=model.classifier.out_proj.in_features, out_features=config.num_labels, bias=True)\r\nmodel.config = config\r\nmodel.num_labels = config.num_labels\r\ntokenizer = XLMRobertaTokenizer.from_pretrained('../xlm-roberta_domains_classifier/model')\r\nmodel.cuda()\r\n```\r\n\r\nModel summary:\r\n```\r\nXLMRobertaForSequenceClassification(\r\n (roberta): RobertaModel(\r\n (embeddings): RobertaEmbeddings(\r\n (word_embeddings): Embedding(250002, 768, padding_idx=1)\r\n (position_embeddings): Embedding(514, 768, padding_idx=1)\r\n (token_type_embeddings): Embedding(1, 768)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (encoder): RobertaEncoder(\r\n (layer): ModuleList(\r\n (0): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (1): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (2): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (3): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (4): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (5): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (6): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (7): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (8): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (9): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (10): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (11): RobertaLayer(\r\n (attention): RobertaAttention(\r\n (self): RobertaSelfAttention(\r\n (query): Linear(in_features=768, out_features=768, bias=True)\r\n (key): Linear(in_features=768, out_features=768, bias=True)\r\n (value): Linear(in_features=768, out_features=768, bias=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n (output): RobertaSelfOutput(\r\n (dense): Linear(in_features=768, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n (intermediate): RobertaIntermediate(\r\n (dense): Linear(in_features=768, out_features=3072, bias=True)\r\n )\r\n (output): RobertaOutput(\r\n (dense): Linear(in_features=3072, out_features=768, bias=True)\r\n (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\r\n (dropout): Dropout(p=0.1, inplace=False)\r\n )\r\n )\r\n )\r\n )\r\n )\r\n (classifier): Linear(in_features=768, out_features=12, bias=False)\r\n)\r\n```\r\n\r\n\r\nData preparation:\r\n```\r\ntrain_encoding = tokenizer(train_df.text.to_list(), return_tensors='pt', padding=True, truncation=True).to(device)\r\ntrain_input_ids = train_encoding['input_ids'].to(device)\r\ntrain_attention_mask = train_encoding['attention_mask'].to(device)\r\ntrain_labels = torch.tensor(train_df.label.to_list()).unsqueeze(0).to(device)[0]\r\n\r\nval_encoding = tokenizer(val_df.text.to_list(), return_tensors='pt', padding=True, truncation=True).to(device)\r\nval_input_ids = val_encoding['input_ids'].to(device)\r\nval_attention_mask = val_encoding['attention_mask'].to(device)\r\nval_labels = torch.tensor(val_df.label.to_list()).unsqueeze(0).to(device)[0]\r\n\r\ntest_encoding = tokenizer(test_df.text.to_list(), return_tensors='pt', padding=True, truncation=True).to(device)\r\ntest_input_ids = test_encoding['input_ids'].to(device)\r\ntest_attention_mask = test_encoding['attention_mask'].to(device)\r\ntest_labels = torch.tensor(test_df.label.to_list()).unsqueeze(0).to(device)[0]\r\n\r\n\r\nbatch_size = 32\r\ntrain_data = TensorDataset(train_input_ids, train_attention_mask, train_labels)\r\ntrain_sampler = RandomSampler(train_data)\r\ntrain_dataloader = DataLoader(train_data, sampler=train_sampler, batch_size=batch_size)\r\n\r\nvalidation_data = TensorDataset(val_input_ids, val_attention_mask, val_labels)\r\nvalidation_sampler = SequentialSampler(validation_data)\r\nvalidation_dataloader = DataLoader(validation_data, sampler=validation_sampler, batch_size=batch_size)\r\n\r\ntest_data = TensorDataset(test_input_ids, test_attention_mask, test_labels)\r\ntest_sampler = SequentialSampler(test_data)\r\ntest_dataloader = DataLoader(test_data, sampler=test_sampler, batch_size=batch_size)\r\n```\r\n\r\nTraining logic:\r\n```\r\noptimizer = AdamW(model.parameters(),\r\n lr = 4e-5, \r\n eps = 1e-8 # args.adam_epsilon - default is 1e-8.\r\n )\r\nfrom transformers import get_linear_schedule_with_warmup\r\nepochs = 3\r\ntotal_steps = len(train_dataloader) * epochs\r\nscheduler = get_linear_schedule_with_warmup(optimizer, \r\n num_warmup_steps = 0, # Default value in run_glue.py\r\n num_training_steps = total_steps)\r\n\r\n\r\n\r\nseed_val = 42\r\nrandom.seed(seed_val)\r\nnp.random.seed(seed_val)\r\ntorch.manual_seed(seed_val)\r\ntorch.cuda.manual_seed_all(seed_val)\r\n# Store the average loss after each epoch so we can plot them.\r\nloss_values = []\r\n# For each epoch...\r\nfor epoch_i in range(0, epochs):\r\n \r\n # ========================================\r\n # Training\r\n # ========================================\r\n \r\n print(\"\")\r\n print('======== Epoch {:} / {:} ========'.format(epoch_i + 1, epochs))\r\n print('Training...')\r\n t0 = time.time()\r\n total_loss = 0\r\n model.train()\r\n for step, batch in enumerate(train_dataloader):\r\n if step % 50 == 0 and not step == 0:\r\n elapsed = format_time(time.time() - t0)\r\n \r\n print(' Batch {:>5,} of {:>5,}. Elapsed: {:}.'.format(step, len(train_dataloader), elapsed))\r\n b_input_ids = batch[0].to(device)\r\n b_input_mask = batch[1].to(device)\r\n b_labels = batch[2].to(device) \r\n model.zero_grad() \r\n \r\n outputs = model(b_input_ids, \r\n token_type_ids=None, \r\n attention_mask=b_input_mask, \r\n labels=b_labels)\r\n \r\n loss = outputs[0]\r\n total_loss += loss.item()\r\n loss.backward()\r\n torch.nn.utils.clip_grad_norm_(model.parameters(), 1.0)\r\n optimizer.step()\r\n scheduler.step()\r\n avg_train_loss = total_loss / len(train_dataloader) \r\n \r\n loss_values.append(avg_train_loss)\r\n print(\"\")\r\n print(\" Average training loss: {0:.2f}\".format(avg_train_loss))\r\n print(\" Training epcoh took: {:}\".format(format_time(time.time() - t0)))\r\n \r\n # ========================================\r\n # Validation\r\n # ========================================\r\n print(\"\")\r\n print(\"Running Validation...\")\r\n t0 = time.time()\r\n model.eval()\r\n eval_loss, eval_accuracy = 0, 0\r\n nb_eval_steps, nb_eval_examples = 0, 0\r\n for batch in validation_dataloader:\r\n \r\n batch = tuple(t.to(device) for t in batch)\r\n \r\n b_input_ids, b_input_mask, b_labels = batch\r\n \r\n with torch.no_grad(): \r\n outputs = model(b_input_ids, \r\n token_type_ids=None, \r\n attention_mask=b_input_mask)\r\n \r\n logits = outputs[0]\r\n logits = logits.detach().cpu().numpy()\r\n label_ids = b_labels.to('cpu').numpy()\r\n \r\n tmp_eval_accuracy = flat_accuracy(logits, label_ids)\r\n \r\n eval_accuracy += tmp_eval_accuracy\r\n nb_eval_steps += 1\r\n print(\" Accuracy: {0:.2f}\".format(eval_accuracy/nb_eval_steps))\r\n print(\" Validation took: {:}\".format(format_time(time.time() - t0)))\r\nprint(\"\")\r\nprint(\"Training complete!\")\r\n\r\n```\r\n\r\nThe error occurs when reaching this line:\r\n```\r\noutputs = model(b_input_ids, \r\n token_type_ids=None, \r\n attention_mask=b_input_mask, \r\n labels=b_labels)\r\n```" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "@n1t0 ", "This seems to work for me, I guess it crashes because you don't have enough memory. Unfortunately `TextDataset` has not been optimized for fast tokenizers yet, so it does a lot more work than needed when using them. It's probably better to use python tokenizers for now with `TextDataset`.\r\n\r\nAlso, maybe the [huggingface/nlp](https://github.com/huggingface/nlp) library might be better suited here. cc @lhoestq ", "You could try\r\n```python\r\nfrom transformers import AutoTokenizer\r\nfrom nlp import load_dataset\r\n\r\ntokenizer = AutoTokenizer.from_pretrained(\"roberta-base\", use_fast=True)\r\ndataset = load_dataset(\"text\", data_files=\"path/to/wiki.train.raw\", split=\"train\")\r\ntokenized_dataset = dataset.map(lambda ex: tokenizer(ex[\"text\"]), batched=True)\r\nprint(tokenized_dataset[0][\"input_ids\"])\r\n```\r\nWe're still working on making it as fast as we can, but at least you won't have any memory issues.", "Re @n1t0 comment: \"I guess it crashes because you don't have enough memory\" this is correct. (I was hoping I could get away with 61.0 GiB, the standard for an AWS `p3.2xlarge`.)\r\n\r\nRe @lhoestq your code ran without errors for me. Thanks! \r\n\r\nI did get a lot of the [`Token indices sequence length is longer than the specified maximum sequence length for this model (522 > 512). Running this sequence through the model will result in indexing errors`](https://github.com/huggingface/transformers/issues/1791) warnings which I wasn't getting before.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[ { "id": 2796628563, "node_id": "MDU6TGFiZWwyNzk2NjI4NTYz", "url": "https://api.github.com/repos/huggingface/transformers/labels/WIP", "name": "WIP", "color": "234C99", "default": false, "description": "Label your PR/Issue with WIP for some long outstanding Issues/PRs that are work in progress" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5903?src=pr&el=h1) Report\n> Merging [#5903](https://codecov.io/gh/huggingface/transformers/pull/5903?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/492bb6aa486856f8243dfeb533ed1b23e996e403?el=desc) will **decrease** coverage by `2.80%`.\n> The diff coverage is `83.75%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5903?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5903 +/- ##\n==========================================\n- Coverage 80.12% 77.31% -2.81% \n==========================================\n Files 169 152 -17 \n Lines 32317 26290 -6027 \n==========================================\n- Hits 25893 20326 -5567 \n+ Misses 6424 5964 -460 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5903?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/theseus/theseus\\_list.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90aGVzZXVzL3RoZXNldXNfbGlzdC5weQ==) | `67.64% <67.64%> (ø)` | |\n| [src/transformers/theseus/theseus\\_module.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90aGVzZXVzL3RoZXNldXNfbW9kdWxlLnB5) | `88.88% <88.88%> (ø)` | |\n| [src/transformers/theseus/\\_\\_init\\_\\_.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90aGVzZXVzL19faW5pdF9fLnB5) | `100.00% <100.00%> (ø)` | |\n| [src/transformers/theseus/layerdrop\\_list.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90aGVzZXVzL2xheWVyZHJvcF9saXN0LnB5) | `100.00% <100.00%> (ø)` | |\n| [src/transformers/theseus/mixout\\_list.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90aGVzZXVzL21peG91dF9saXN0LnB5) | `100.00% <100.00%> (ø)` | |\n| [src/transformers/theseus/theseus\\_errors.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90aGVzZXVzL3RoZXNldXNfZXJyb3JzLnB5) | `100.00% <100.00%> (ø)` | |\n| [src/transformers/commands/env.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb21tYW5kcy9lbnYucHk=) | `0.00% <0.00%> (-100.00%)` | :arrow_down: |\n| [src/transformers/commands/\\_\\_init\\_\\_.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb21tYW5kcy9fX2luaXRfXy5weQ==) | `0.00% <0.00%> (-100.00%)` | :arrow_down: |\n| [src/transformers/commands/transformers\\_cli.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb21tYW5kcy90cmFuc2Zvcm1lcnNfY2xpLnB5) | `0.00% <0.00%> (-87.50%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.68%)` | :arrow_down: |\n| ... and [166 more](https://codecov.io/gh/huggingface/transformers/pull/5903/diff?src=pr&el=tree-more) | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5903?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5903?src=pr&el=footer). Last update [492bb6a...20dd34f](https://codecov.io/gh/huggingface/transformers/pull/5903?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "which examples have you tested this with?", "> which examples have you tested this with?\n\nrun_glue and run_ner but more to come!", "Want an early review for this? Would be wonderful! And it can save me some time before I do the docs. @sshleifer ", "> run_glue and run_ner but more to come! \r\n\r\n:heart_eyes: can't wait to re-fine-tune my NER models :hugs: ", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "Well I've been really busy recently but don't close it for me stalebot!", "Thanks for reopening it @LysandreJik! I was stuck with some details and I'll probably get it done soon.", "Sounds great, looking forward to it!", "Will move to an independent package. Closing this." ]

[ { "id": 1108649053, "node_id": "MDU6TGFiZWwxMTA4NjQ5MDUz", "url": "https://api.github.com/repos/huggingface/transformers/labels/Help%20wanted", "name": "Help wanted", "color": "008672", "default": false, "description": "Extra attention is needed, help appreciated" }, { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Very strange indeed. Nothing comes to mind, but if you can show the fairseq discrepancies I can take a look.", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "@astariul Hi I also encounter this problem. The finetuned BART just gives the same output for `<s>` under eval mode, have you fixed it? ", "@turing-yfqiu I'm sorry it's long time ago, I don't remember how I did...." ]

[ { "id": 2155169140, "node_id": "MDU6TGFiZWwyMTU1MTY5MTQw", "url": "https://api.github.com/repos/huggingface/transformers/labels/trainer", "name": "trainer", "color": "2ef289", "default": false, "description": "" } ]

[ { "login": "julien-c", "id": 326577, "node_id": "MDQ6VXNlcjMyNjU3Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/326577?v=4", "gravatar_id": "", "url": "https://api.github.com/users/julien-c", "html_url": "https://github.com/julien-c", "followers_url": "https://api.github.com/users/julien-c/followers", "following_url": "https://api.github.com/users/julien-c/following{/other_user}", "gists_url": "https://api.github.com/users/julien-c/gists{/gist_id}", "starred_url": "https://api.github.com/users/julien-c/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/julien-c/subscriptions", "organizations_url": "https://api.github.com/users/julien-c/orgs", "repos_url": "https://api.github.com/users/julien-c/repos", "events_url": "https://api.github.com/users/julien-c/events{/privacy}", "received_events_url": "https://api.github.com/users/julien-c/received_events", "type": "User", "site_admin": false } ]

[ "training loss should be printed every 500 iteration, but there is no log in pretraing.\r\n(parser.add_argument(\"--logsteps\", help=\"logging steps\", type=int, default=500)", "Hi! This wasn't intentional, so we've fixed it with #6097. If you rerun the script, you should see the losses now." ]

[]

[]

[ "You need to pass `output_hidden_states=True` when calling your model, or set this option when instantiating it. For instance, if you're using a pretrained model:\r\n```\r\nfrom transformers import DistilBertModel\r\nmodel = DistilBertModel.from_pretrained('distilbert-base-uncased', output_hidden_states=True)\r\n```" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[ { "login": "sshleifer", "id": 6045025, "node_id": "MDQ6VXNlcjYwNDUwMjU=", "avatar_url": "https://avatars.githubusercontent.com/u/6045025?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sshleifer", "html_url": "https://github.com/sshleifer", "followers_url": "https://api.github.com/users/sshleifer/followers", "following_url": "https://api.github.com/users/sshleifer/following{/other_user}", "gists_url": "https://api.github.com/users/sshleifer/gists{/gist_id}", "starred_url": "https://api.github.com/users/sshleifer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sshleifer/subscriptions", "organizations_url": "https://api.github.com/users/sshleifer/orgs", "repos_url": "https://api.github.com/users/sshleifer/repos", "events_url": "https://api.github.com/users/sshleifer/events{/privacy}", "received_events_url": "https://api.github.com/users/sshleifer/received_events", "type": "User", "site_admin": false } ]

[]

[]

[ { "login": "sshleifer", "id": 6045025, "node_id": "MDQ6VXNlcjYwNDUwMjU=", "avatar_url": "https://avatars.githubusercontent.com/u/6045025?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sshleifer", "html_url": "https://github.com/sshleifer", "followers_url": "https://api.github.com/users/sshleifer/followers", "following_url": "https://api.github.com/users/sshleifer/following{/other_user}", "gists_url": "https://api.github.com/users/sshleifer/gists{/gist_id}", "starred_url": "https://api.github.com/users/sshleifer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sshleifer/subscriptions", "organizations_url": "https://api.github.com/users/sshleifer/orgs", "repos_url": "https://api.github.com/users/sshleifer/repos", "events_url": "https://api.github.com/users/sshleifer/events{/privacy}", "received_events_url": "https://api.github.com/users/sshleifer/received_events", "type": "User", "site_admin": false } ]

[]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[ { "id": 1108649053, "node_id": "MDU6TGFiZWwxMTA4NjQ5MDUz", "url": "https://api.github.com/repos/huggingface/transformers/labels/Help%20wanted", "name": "Help wanted", "color": "008672", "default": false, "description": "Extra attention is needed, help appreciated" }, { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Do you want to wait for a stable release for the torch-xla? Also shouldn't the end user pass in \"(num_tpu_cores=8)\" when they are creating the lightning's trainer. It should automatically handle the rest for us i think. Also we have \"bloat16\" for TPUs as well.", "It totally could work out of the box. In which case this issue could be as simple as running a shell command on a tpu machine, seeing that it works well, and then checking in the working command or commenting here :).\r\n", "Alright cool! Let's see if I can make an attempt to do it! Very new to this examples section of hugging face. ", "Hey! So i tried running it on the [colab](https://colab.research.google.com/drive/16q2GWrnZ0Tjg1OxJQUcaWKCWwn3Jh5z0?usp=sharing) first, but it seems there's some error wrt bart model. The traceback it threw, (another issue open as well https://github.com/huggingface/transformers/issues/5915).\r\n\r\nTo reproduce the same, I executed the finetune.sh with\r\n```bash\r\n!sh finetune.sh \\\r\n--data_dir /content/xsum \\\r\n--model_name_or_path facebook/bart-base \\\r\n--output_dir=xsum_results \\\r\n--train_batch_size=2 \\\r\n--eval_batch_size=2 \\\r\n--num_train_epochs 1 \\\r\n--n_tpu_cores 8 \\\r\n--tpu_cores 8\r\n```\r\nAnd modified the args in the shell snip when we invoke the python finetune.py (removed fp16 and gpus to 0)\r\n\r\n\r\n```\r\nAttempted to call `variable.set_data(tensor)`, but `variable` and `tensor` have incompatible tensor type.\r\nTraceback (most recent call last):\r\n File \"/usr/local/lib/python3.6/dist-packages/torch_xla/distributed/xla_multiprocessing.py\", line 119, in _start_fn\r\n fn(gindex, *args)\r\n File \"/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/distrib_parts.py\", line 222, in tpu_train\r\n self.run_pretrain_routine(model)\r\n File \"/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/trainer.py\", line 1196, in run_pretrain_routine\r\n False)\r\n File \"/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/evaluation_loop.py\", line 293, in _evaluate\r\n output = self.evaluation_forward(model, batch, batch_idx, dataloader_idx, test_mode)\r\n File \"/usr/local/lib/python3.6/dist-packages/pytorch_lightning/trainer/evaluation_loop.py\", line 470, in evaluation_forward\r\n output = model.validation_step(*args)\r\n File \"finetune.py\", line 145, in validation_step\r\n return self._generative_step(batch)\r\n File \"finetune.py\", line 176, in _generative_step\r\n decoder_start_token_id=self.decoder_start_token_id,\r\n File \"/usr/local/lib/python3.6/dist-packages/torch/autograd/grad_mode.py\", line 15, in decorate_context\r\n return func(*args, **kwargs)\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/generation_utils.py\", line 248, in generate\r\n if self.get_output_embeddings() is None:\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/modeling_bart.py\", line 1113, in get_output_embeddings\r\n return _make_linear_from_emb(self.model.shared) # make it on the fly\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/modeling_bart.py\", line 190, in _make_linear_from_emb\r\n lin_layer.weight.data = emb.weight.data\r\n```", "Makes sense. You could try to instantiate\r\n```python\r\nself.lm_head = _make_linear_from_emb(self.model.shared) \r\n```\r\nin BartForConditionalGeneration.__init__ \r\nand then have `get_output_embeddings` return `self.lm_head`.\r\n", "Alright! Thanks; Will keep you posted 🎉;\r\n\r\nEdit -1:\r\n\r\nSo things seems to work after i make the changes you said. The trouble is that training seems to be frozen on TPUs.", "`emb.weight.data` is...\r\n```\r\ntensor([[-0.0370, 0.1117, 0.1829, ..., 0.2054, 0.0578, -0.0750],\r\n [ 0.0055, -0.0049, -0.0069, ..., -0.0030, 0.0038, 0.0087],\r\n [-0.0448, 0.4604, -0.0604, ..., 0.1073, 0.0310, 0.0477],\r\n ...,\r\n [-0.0138, 0.0278, -0.0467, ..., 0.0455, -0.0265, 0.0125],\r\n [-0.0043, 0.0153, -0.0567, ..., 0.0496, 0.0108, -0.0099],\r\n [ 0.0053, 0.0324, -0.0179, ..., -0.0085, 0.0223, -0.0020]],\r\n device='xla:1')\r\n```\r\n...and `lin_layer.weight.data` is...\r\n```\r\ntensor([[-1.0449e-03, 4.0973e-03, -9.7727e-04, ..., 8.2363e-04,\r\n -3.2153e-03, 3.5317e-03],\r\n [ 2.3644e-03, 3.5527e-03, -1.2428e-03, ..., -1.0983e-04,\r\n -2.1916e-03, 5.3099e-05],\r\n [-4.2492e-03, 3.8183e-04, 3.2527e-03, ..., -4.4359e-03,\r\n 7.6555e-04, -4.1728e-03],\r\n ...,\r\n [-4.3412e-03, 2.8537e-03, 7.9720e-04, ..., 2.9499e-03,\r\n 2.6357e-03, -3.5283e-03],\r\n [ 3.7042e-03, -3.0546e-03, 3.9206e-03, ..., -2.3771e-03,\r\n 4.3551e-03, 1.1703e-04],\r\n [ 3.5616e-03, -3.1224e-03, 1.3898e-03, ..., -2.1096e-05,\r\n 5.4077e-04, 1.6183e-03]])\r\n```\r\n... and you try to do...\r\n```\r\nlin_layer.weight.data = emb.weight.data\r\n```\r\n\r\nIsn't the problem that they are on different devices?", "We can remove the thing which does it on fly(refer to sshleifer comment above). That won't work in case of TPUs.", "Good to know. Once you get everything working it would be great to have all the required changes consolidated into one PR.", "I am not quite sure that lightning does this optimization when we use multiple TPU cores (only available at the nightly-xla's). Refer [here](https://github.com/pytorch/xla/issues/1870#issuecomment-623603323).", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=h1) Report\n> Merging [#5894](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/09a2f40684f77e62d0fd8485fe9d2d610390453f&el=desc) will **decrease** coverage by `1.19%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5894 +/- ##\n==========================================\n- Coverage 78.49% 77.30% -1.20% \n==========================================\n Files 146 146 \n Lines 26210 26210 \n==========================================\n- Hits 20573 20261 -312 \n- Misses 5637 5949 +312 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/training\\_args\\_tf.py](https://codecov.io/gh/huggingface/transformers/pull/5894/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzX3RmLnB5) | `47.45% <ø> (ø)` | |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5894/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `23.38% <0.00%> (-73.39%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5894/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.71% <0.00%> (+0.50%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5894/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=footer). Last update [09a2f40...b9754e4](https://codecov.io/gh/huggingface/transformers/pull/5894?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=h1) Report\n> Merging [#5893](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/09a2f40684f77e62d0fd8485fe9d2d610390453f&el=desc) will **increase** coverage by `0.17%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5893 +/- ##\n==========================================\n+ Coverage 78.49% 78.66% +0.17% \n==========================================\n Files 146 146 \n Lines 26210 26210 \n==========================================\n+ Hits 20573 20619 +46 \n+ Misses 5637 5591 -46 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/training\\_args.py](https://codecov.io/gh/huggingface/transformers/pull/5893/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90cmFpbmluZ19hcmdzLnB5) | `77.55% <ø> (ø)` | |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5893/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5893/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `95.53% <0.00%> (+69.51%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=footer). Last update [09a2f40...762ff74](https://codecov.io/gh/huggingface/transformers/pull/5893?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "Working command after MbartConfig change:\r\n```bash\r\nexport d=mbart_benchmark_data\r\npython examples/benchmarking/run_benchmark.py \\\r\n --models facebook/mbart-large-en-ro \\\r\n --log_filename $d/log.txt \\\r\n --inference_memory_csv \\ \r\n $d/inference_memory.csv \\ \r\n --train_memory_csv $d/train_memory.csv \\\r\n --train_time_csv $d/train_time.csv \\\r\n --inference_time_csv $d/inference_time.csv \\\r\n --fp16 --log_print --training --save_to_csv \\\r\n --batch_sizes 4 8 12 16\r\n```\r\n\r\n\r\n", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "fixed indirectly by #6441 " ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "You can't load directly a TF checkpoint if it's not been generated by the TensorFlow side of the library (which can't be the case here since you're using a TF1 checkpoint and the lib requires TF2.0). You probably have to use the [conversion script](https://huggingface.co/transformers/converting_tensorflow_models.html).", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "How did you install the package ?\r\n\r\nAccording to the [README](https://github.com/huggingface/transformers#from-source) you should do :\r\n\r\n```console\r\ngit clone https://github.com/huggingface/transformers\r\ncd transformers\r\npip install -e .\r\n```\r\n", "Hi ,\r\n\r\nApologies for raising this issue, the error was something specific to singularity images that I have been using to run transformers on HPC. \r\nPlease pardon this issue. \r\n\r\nAnyway thanks @Colanim. I knew the steps, but the error was from somewhere else. \r\n" ]

[]

[ { "login": "stas00", "id": 10676103, "node_id": "MDQ6VXNlcjEwNjc2MTAz", "avatar_url": "https://avatars.githubusercontent.com/u/10676103?v=4", "gravatar_id": "", "url": "https://api.github.com/users/stas00", "html_url": "https://github.com/stas00", "followers_url": "https://api.github.com/users/stas00/followers", "following_url": "https://api.github.com/users/stas00/following{/other_user}", "gists_url": "https://api.github.com/users/stas00/gists{/gist_id}", "starred_url": "https://api.github.com/users/stas00/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/stas00/subscriptions", "organizations_url": "https://api.github.com/users/stas00/orgs", "repos_url": "https://api.github.com/users/stas00/repos", "events_url": "https://api.github.com/users/stas00/events{/privacy}", "received_events_url": "https://api.github.com/users/stas00/received_events", "type": "User", "site_admin": false } ]

[ "Spent an hour testing the new version on real data:\r\n\r\n(1) `./train_mbart_enro_multigpu` works!\r\n(2) `./train_mbart_enro_multigpu --logger_name wandb` ~doesn't work~ works.\r\n(3) trainer.test seems to try to call `main` again, and then runs into an \"output directory\" already exists error.\r\n\r\nIt hangs for me with versions:\r\n```bash\r\npytorch-lightning==0.8.5\r\ntorch==1.5.1+cu101\r\n```\r\n\r\ncc @nateraw @williamFalcon ", "@stas00 do you think multi-gpu test coverage (at least on machines with multiple GPU, not CI yet) is possible?", "> @stas00 do you think multi-gpu test coverage (at least on machines with multiple GPU, not CI yet) is possible?\r\n\r\nAre you referring to `pytest --cov`? Yes, of course, it should work.\r\n\r\nThe main issue with multiprocessing and coverage is a potential unregistered coverage if the sub-process hasn't finished its work right away and is hanging around. But we already have this issue in the benchmarking tests - I need to go back and pound at it some more.", "Sorry, I'm referring to getting that test to pass.\r\nAt the moment it fails (twice!).", "Oh, I see what you meant. Let me investigate.", "Yeah, multigpu is a problem. Here is a totally different manifestation of it - hanging and other subtests failing:\r\n\r\n```\r\npytest --disable-warnings examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller\r\n――――――――――――――――――――――――――――――――――――――――― TestSummarizationDistiller.test_distill_checkpointing_with_teacher ―――――――――――――――――――――――――――――――――――――――――\r\n\r\nself = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_distill_checkpointing_with_teacher>\r\n\r\n def test_distill_checkpointing_with_teacher(self):\r\n updates = dict(\r\n student_encoder_layers=2,\r\n student_decoder_layers=1,\r\n max_epochs=4,\r\n val_check_interval=0.25,\r\n alpha_hid=2.0,\r\n model_name_or_path=\"IGNORE_THIS_IT_DOESNT_GET_USED\",\r\n )\r\n model = self._test_distiller_cli(updates, check_contents=False)\r\n \r\n ckpts = list(Path(model.output_dir).glob(\"*.ckpt\"))\r\n> self.assertEqual(1, len(ckpts))\r\nE AssertionError: 1 != 0\r\n\r\nexamples/seq2seq/test_seq2seq_examples.py:173: AssertionError\r\n\r\n examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_distill_checkpointing_with_teacher ⨯ 14% █▌ using module BartTranslationDistiller\r\nEpoch 2: 100%|████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:00<00:00, 11.53it/s, loss=2.970, v_num=19]\r\n \r\n\r\n――――――――――――――――――――――――――――――――――――――――――――――――――― TestSummarizationDistiller.test_distill_mbart ――――――――――――――――――――――――――――――――――――――――――――――――――――\r\n\r\nself = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_distill_mbart>\r\n\r\n def test_distill_mbart(self):\r\n updates = dict(\r\n student_encoder_layers=2,\r\n student_decoder_layers=1,\r\n num_train_epochs=4,\r\n val_check_interval=0.25,\r\n alpha_hid=2.0,\r\n task=\"translation\",\r\n model_name_or_path=\"IGNORE_THIS_IT_DOESNT_GET_USED\",\r\n tokenizer_name=MBART_TINY,\r\n teacher=MBART_TINY,\r\n src_lang=\"en_XX\",\r\n tgt_lang=\"ro_RO\",\r\n )\r\n model = self._test_distiller_cli(updates, check_contents=False)\r\n assert model.model.config.model_type == \"mbart\"\r\n \r\n ckpts = list(Path(model.output_dir).glob(\"*.ckpt\"))\r\n> self.assertEqual(1, len(ckpts))\r\nE AssertionError: 1 != 0\r\n\r\nexamples/seq2seq/test_seq2seq_examples.py:224: AssertionError\r\n\r\n examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_distill_mbart ⨯ 29% ██▉ using module SummarizationModule\r\nEpoch 2: 100%|███████████████████████████████████████████████████████████████████████████████████| 4/4 [00:00<00:00, 4.23it/s, loss=10.785, v_num=19]\r\n \r\n\r\n――――――――――――――――――――――――――――――――――――――――――――――――― TestSummarizationDistiller.test_distill_no_teacher ―――――――――――――――――――――――――――――――――――――――――――――――――\r\n\r\nself = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_distill_no_teacher>\r\n\r\n def test_distill_no_teacher(self):\r\n updates = dict(student_encoder_layers=2, student_decoder_layers=1, no_teacher=True)\r\n> self._test_distiller_cli(updates)\r\n\r\nexamples/seq2seq/test_seq2seq_examples.py:159: \r\n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \r\n\r\nself = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_distill_no_teacher>\r\nupdates = {'no_teacher': True, 'student_decoder_layers': 1, 'student_encoder_layers': 2}, check_contents = True\r\n\r\n def _test_distiller_cli(self, updates, check_contents=True):\r\n default_updates = dict(\r\n label_smoothing=0.0,\r\n early_stopping_patience=-1,\r\n train_batch_size=1,\r\n eval_batch_size=2,\r\n max_epochs=2,\r\n alpha_mlm=0.2,\r\n alpha_ce=0.8,\r\n do_predict=True,\r\n model_name_or_path=\"sshleifer/tinier_bart\",\r\n teacher=CHEAP_ARGS[\"model_name_or_path\"],\r\n val_check_interval=0.5,\r\n alpha_encoder_loss=0.4,\r\n )\r\n default_updates.update(updates)\r\n args_d: dict = CHEAP_ARGS.copy()\r\n tmp_dir = make_test_data_dir()\r\n output_dir = tempfile.mkdtemp(prefix=\"output_\")\r\n\r\n args_d.update(data_dir=tmp_dir, output_dir=output_dir, **default_updates)\r\n model = distill_main(argparse.Namespace(**args_d))\r\n if not check_contents:\r\n return model\r\n contents = os.listdir(output_dir)\r\n contents = {os.path.basename(p) for p in contents}\r\n ckpt_files = [p for p in contents if p.endswith(\"ckpt\")]\r\n> assert len(ckpt_files) > 0\r\nE AssertionError: assert 0 > 0\r\nE + where 0 = len([])\r\n\r\nexamples/seq2seq/test_seq2seq_examples.py:271: AssertionError\r\n\r\n examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_distill_no_teacher ⨯ 43% ████▍\r\n examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_distill_t5 s 57% █████▊\r\n ../../../../../home/stas/anaconda3/envs/main-38/lib/python3.8/unittest/case.py::TestSummarizationDistiller.test_hub_configs s 71% ███████▎\r\n examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_loss_fn ✓ 86% ████████▋\r\n examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller.test_multigpu s 100% ██████████\r\n============================================================== short test summary info ===============================================================\r\nFAILED examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller::test_distill_checkpointing_with_teacher - AssertionError: 1 != 0\r\nFAILED examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller::test_distill_mbart - AssertionError: 1 != 0\r\nFAILED examples/seq2seq/test_seq2seq_examples.py::TestSummarizationDistiller::test_distill_no_teacher - AssertionError: assert 0 > 0\r\n\r\nResults (18.02s):\r\n 1 passed\r\n 3 failed\r\n - examples/seq2seq/test_seq2seq_examples.py:161 TestSummarizationDistiller.test_distill_checkpointing_with_teacher\r\n - examples/seq2seq/test_seq2seq_examples.py:206 TestSummarizationDistiller.test_distill_mbart\r\n - examples/seq2seq/test_seq2seq_examples.py:157 TestSummarizationDistiller.test_distill_no_teacher\r\n 3 skipped\r\n```\r\nand it hangs. getting a trace:\r\n```\r\nThread 0x00007fc81efd6700 (most recent call first):\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 306 in wait\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py\", line 179 in get\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py\", line 232 in run\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 932 in _bootstrap_inner\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 890 in _bootstrap\r\n\r\nThread 0x00007fc80e850700 (most recent call first):\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 306 in wait\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py\", line 179 in get\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py\", line 232 in run\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 932 in _bootstrap_inner\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 890 in _bootstrap\r\n\r\nThread 0x00007fc8248b0700 (most recent call first):\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 306 in wait\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py\", line 179 in get\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py\", line 232 in run\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 932 in _bootstrap_inner\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 890 in _bootstrap\r\n\r\nThread 0x00007fc8250b1700 (most recent call first):\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 306 in wait\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py\", line 179 in get\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py\", line 232 in run\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 932 in _bootstrap_inner\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 890 in _bootstrap\r\n\r\nThread 0x00007fc81cbec700 (most recent call first):\r\n<no Python frame>\r\n\r\nThread 0x00007fc81e7d5700 (most recent call first):\r\n<no Python frame>\r\n\r\nThread 0x00007fc826ffd700 (most recent call first):\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 306 in wait\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/queue.py\", line 179 in get\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/tensorboard/summary/writer/event_file_writer.py\", line 232 in run\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 932 in _bootstrap_inner\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/threading.py\", line 890 in _bootstrap\r\n\r\nCurrent thread 0x00007fc91d2e2740 (most recent call first):\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/torch/distributed/rendezvous.py\", line 172 in _env_rendezvous_handler\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py\", line 423 in init_process_group\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pytorch_lightning/core/lightning.py\", line 973 in init_ddp_connection\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pytorch_lightning/trainer/distrib_data_parallel.py\", line 506 in ddp_train\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pytorch_lightning/trainer/distrib_data_parallel.py\", line 462 in spawn_ddp_children\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py\", line 992 in fit\r\n File \"/mnt/nvme1/code/huggingface/transformers-master/examples/lightning_base.py\", line 380 in generic_train\r\n File \"/mnt/nvme1/code/huggingface/transformers-master/examples/seq2seq/finetune.py\", line 420 in main\r\n File \"/mnt/nvme1/code/huggingface/transformers-master/examples/seq2seq/distillation.py\", line 511 in distill_main\r\n File \"/mnt/nvme1/code/huggingface/transformers-master/examples/seq2seq/test_seq2seq_examples.py\", line 267 in _test_distiller_cli\r\n File \"/mnt/nvme1/code/huggingface/transformers-master/examples/seq2seq/test_seq2seq_examples.py\", line 157 in test_multigpu\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/unittest/case.py\", line 633 in _callTestMethod\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/unittest/case.py\", line 676 in run\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/unittest/case.py\", line 736 in __call__\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/unittest.py\", line 278 in runtest\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py\", line 153 in pytest_runtest_call\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/callers.py\", line 187 in _multicall\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py\", line 84 in <lambda>\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py\", line 93 in _hookexec\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/hooks.py\", line 286 in __call__\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py\", line 247 in <lambda>\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py\", line 294 in from_call\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py\", line 246 in call_runtest_hook\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py\", line 207 in call_and_report\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py\", line 117 in runtestprotocol\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/runner.py\", line 100 in pytest_runtest_protocol\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/callers.py\", line 187 in _multicall\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py\", line 84 in <lambda>\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py\", line 93 in _hookexec\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/hooks.py\", line 286 in __call__\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/main.py\", line 321 in pytest_runtestloop\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/callers.py\", line 187 in _multicall\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py\", line 84 in <lambda>\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py\", line 93 in _hookexec\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/hooks.py\", line 286 in __call__\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/main.py\", line 296 in _main\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/main.py\", line 240 in wrap_session\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/main.py\", line 289 in pytest_cmdline_main\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/callers.py\", line 187 in _multicall\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py\", line 84 in <lambda>\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/manager.py\", line 93 in _hookexec\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/pluggy/hooks.py\", line 286 in __call__\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/config/__init__.py\", line 157 in main\r\n File \"/home/stas/anaconda3/envs/main-38/lib/python3.8/site-packages/_pytest/config/__init__.py\", line 180 in console_main\r\n File \"/home/stas/anaconda3/envs/main-38/bin/pytest\", line 8 in <module>\r\n```\r\n\r\nIf I run just these 3 tests - all is green. So its the multigpu test.\r\n\r\nAh, it somehow starts running pytest second time half-way through.\r\n\r\nRechecked with py37 and py38 - same, same.\r\n\r\np.s. checkout `pip install pytest-sugar` - makes running tests a brighter experience.", "I will try to make a small reproducible test where the problem still happens, and then we can try to sort it out.", "OK, so first there is some strange issue with the properties - seems to be related to https://github.com/pytorch/pytorch/issues/43542, the multigpu test fails with:\r\n```\r\n\r\n―――――――――――――――――――――――――――――――――――――――――――――――――――――― TestSummarizationDistiller.test_multigpu ――――――――――――――――――――――――――――――――――――――――――――――――――――――\r\n\r\nself = <seq2seq.test_seq2seq_examples.TestSummarizationDistiller testMethod=test_multigpu>\r\n\r\n @require_multigpu\r\n def test_multigpu(self):\r\n updates = dict(\r\n no_teacher=True,\r\n freeze_encoder=True,\r\n gpus=2,\r\n sortish_sampler=True,\r\n )\r\n> self._test_distiller_cli(updates, check_contents=False)\r\n\r\nexamples/seq2seq/test_seq2seq_examples.py:154: \r\n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \r\nexamples/seq2seq/test_seq2seq_examples.py:264: in _test_distiller_cli\r\n model = distill_main(argparse.Namespace(**args_d))\r\nexamples/seq2seq/distillation.py:495: in distill_main\r\n return ft_main(args, model=model)\r\nexamples/seq2seq/finetune.py:399: in main\r\n trainer: pl.Trainer = generic_train(\r\nexamples/lightning_base.py:380: in generic_train\r\n trainer.fit(model)\r\n../../github/00pytorch/pytorch-lightning/pytorch_lightning/trainer/trainer.py:301: in fit\r\n results = self.accelerator_backend.train()\r\n../../github/00pytorch/pytorch-lightning/pytorch_lightning/accelerators/ddp_backend.py:133: in train\r\n self.ddp_train_tmp(process_idx=self.task_idx, mp_queue=None, model=model)\r\n../../github/00pytorch/pytorch-lightning/pytorch_lightning/accelerators/ddp_base_backend.py:148: in ddp_train_tmp\r\n optimizers, lr_schedulers, optimizer_frequencies = self.trainer.init_optimizers(model)\r\n../../github/00pytorch/pytorch-lightning/pytorch_lightning/trainer/optimizers.py:32: in init_optimizers\r\n optim_conf = model.configure_optimizers()\r\nexamples/lightning_base.py:152: in configure_optimizers\r\n scheduler = self.get_lr_scheduler()\r\nexamples/lightning_base.py:122: in get_lr_scheduler\r\n self.opt, num_warmup_steps=self.hparams.warmup_steps, num_training_steps=self.total_steps\r\n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \r\n\r\nself = SummarizationModule(\r\n (model): BartForConditionalGeneration(\r\n (model): BartModel(\r\n (shared): Embedding(50265,... )\r\n )\r\n (layernorm_embedding): LayerNorm((24,), eps=1e-05, elementwise_affine=True)\r\n )\r\n )\r\n )\r\n)\r\nname = 'total_steps'\r\n\r\n def __getattr__(self, name: str) -> Union[Tensor, 'Module']:\r\n if '_parameters' in self.__dict__:\r\n _parameters = self.__dict__['_parameters']\r\n if name in _parameters:\r\n return _parameters[name]\r\n if '_buffers' in self.__dict__:\r\n _buffers = self.__dict__['_buffers']\r\n if name in _buffers:\r\n return _buffers[name]\r\n if '_modules' in self.__dict__:\r\n modules = self.__dict__['_modules']\r\n if name in modules:\r\n return modules[name]\r\n> raise ModuleAttributeError(\"'{}' object has no attribute '{}'\".format(\r\n type(self).__name__, name))\r\nE torch.nn.modules.module.ModuleAttributeError: 'SummarizationModule' object has no attribute 'total_steps'\r\n```\r\nI fixed it with:\r\n```\r\ndiff --git a/examples/lightning_base.py b/examples/lightning_base.py\r\nindex e7c41a3e..f5e858c7 100644\r\n--- a/examples/lightning_base.py\r\n+++ b/examples/lightning_base.py\r\n@@ -119,8 +119,7 @@ class BaseTransformer(pl.LightningModule):\r\n def get_lr_scheduler(self):\r\n get_schedule_func = arg_to_scheduler[self.hparams.lr_scheduler]\r\n scheduler = get_schedule_func(\r\n- self.opt, num_warmup_steps=self.hparams.warmup_steps, num_training_steps=self.total_steps\r\n- )\r\n+ self.opt, num_warmup_steps=self.hparams.warmup_steps, num_training_steps=self.total_steps())\r\n scheduler = {\"scheduler\": scheduler, \"interval\": \"step\", \"frequency\": 1}\r\n return scheduler\r\n\r\n@@ -159,12 +158,11 @@ class BaseTransformer(pl.LightningModule):\r\n def test_epoch_end(self, outputs):\r\n return self.validation_end(outputs)\r\n\r\n- @property\r\n def total_steps(self) -> int:\r\n \"\"\"The number of total training steps that will be run. Used for lr scheduler purposes.\"\"\"\r\n num_devices = max(1, self.hparams.gpus) # TODO: consider num_tpu_cores\r\n effective_batch_size = self.hparams.train_batch_size * self.hparams.accumulate_grad_batches * num_devices\r\n- dataset_size = len(self.train_loader.dataset)\r\n+ dataset_size = len(self.train_dataloader().dataset)\r\n return (dataset_size / effective_batch_size) * self.hparams.max_epochs\r\n\r\n def setup(self, mode):\r\n```\r\n\r\nit works fine the first time, but when it re-reruns itself, it behaves differently - so the above fix was needed.", "WIP: https://github.com/huggingface/transformers/pull/7281\r\n" ]

[]

[]

[ "Actually never mind. At least GPT2 does the shifting internally. " ]

[]

[ { "login": "julien-c", "id": 326577, "node_id": "MDQ6VXNlcjMyNjU3Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/326577?v=4", "gravatar_id": "", "url": "https://api.github.com/users/julien-c", "html_url": "https://github.com/julien-c", "followers_url": "https://api.github.com/users/julien-c/followers", "following_url": "https://api.github.com/users/julien-c/following{/other_user}", "gists_url": "https://api.github.com/users/julien-c/gists{/gist_id}", "starred_url": "https://api.github.com/users/julien-c/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/julien-c/subscriptions", "organizations_url": "https://api.github.com/users/julien-c/orgs", "repos_url": "https://api.github.com/users/julien-c/repos", "events_url": "https://api.github.com/users/julien-c/events{/privacy}", "received_events_url": "https://api.github.com/users/julien-c/received_events", "type": "User", "site_admin": false }, { "login": "mfuntowicz", "id": 2241520, "node_id": "MDQ6VXNlcjIyNDE1MjA=", "avatar_url": "https://avatars.githubusercontent.com/u/2241520?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mfuntowicz", "html_url": "https://github.com/mfuntowicz", "followers_url": "https://api.github.com/users/mfuntowicz/followers", "following_url": "https://api.github.com/users/mfuntowicz/following{/other_user}", "gists_url": "https://api.github.com/users/mfuntowicz/gists{/gist_id}", "starred_url": "https://api.github.com/users/mfuntowicz/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/mfuntowicz/subscriptions", "organizations_url": "https://api.github.com/users/mfuntowicz/orgs", "repos_url": "https://api.github.com/users/mfuntowicz/repos", "events_url": "https://api.github.com/users/mfuntowicz/events{/privacy}", "received_events_url": "https://api.github.com/users/mfuntowicz/received_events", "type": "User", "site_admin": false }, { "login": "sshleifer", "id": 6045025, "node_id": "MDQ6VXNlcjYwNDUwMjU=", "avatar_url": "https://avatars.githubusercontent.com/u/6045025?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sshleifer", "html_url": "https://github.com/sshleifer", "followers_url": "https://api.github.com/users/sshleifer/followers", "following_url": "https://api.github.com/users/sshleifer/following{/other_user}", "gists_url": "https://api.github.com/users/sshleifer/gists{/gist_id}", "starred_url": "https://api.github.com/users/sshleifer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sshleifer/subscriptions", "organizations_url": "https://api.github.com/users/sshleifer/orgs", "repos_url": "https://api.github.com/users/sshleifer/repos", "events_url": "https://api.github.com/users/sshleifer/events{/privacy}", "received_events_url": "https://api.github.com/users/sshleifer/received_events", "type": "User", "site_admin": false } ]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=h1) Report\n> Merging [#5885](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/42fddacd1cac3cc57c3326aa51a409f5090b1261?el=desc) will **decrease** coverage by `0.88%`.\n> The diff coverage is `75.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5885 +/- ##\n==========================================\n- Coverage 78.47% 77.58% -0.89% \n==========================================\n Files 157 157 \n Lines 28569 28571 +2 \n==========================================\n- Hits 22420 22168 -252 \n- Misses 6149 6403 +254 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/pipelines.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9waXBlbGluZXMucHk=) | `80.50% <75.00%> (+0.55%)` | :arrow_up: |\n| [src/transformers/modeling\\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19yZWZvcm1lci5weQ==) | `16.87% <0.00%> (-79.30%)` | :arrow_down: |\n| [src/transformers/configuration\\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3JlZm9ybWVyLnB5) | `21.62% <0.00%> (-78.38%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `22.58% <0.00%> (-72.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_flaubert.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9mbGF1YmVydC5weQ==) | `24.53% <0.00%> (-63.81%)` | :arrow_down: |\n| [src/transformers/modeling\\_marian.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19tYXJpYW4ucHk=) | `60.00% <0.00%> (-30.00%)` | :arrow_down: |\n| [src/transformers/activations.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9hY3RpdmF0aW9ucy5weQ==) | `85.00% <0.00%> (-5.00%)` | :arrow_down: |\n| [src/transformers/tokenization\\_dpr.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fZHByLnB5) | `53.15% <0.00%> (-4.51%)` | :arrow_down: |\n| [src/transformers/modeling\\_bart.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19iYXJ0LnB5) | `95.05% <0.00%> (-0.35%)` | :arrow_down: |\n| [src/transformers/tokenization\\_utils\\_base.py](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfYmFzZS5weQ==) | `93.49% <0.00%> (-0.28%)` | :arrow_down: |\n| ... and [13 more](https://codecov.io/gh/huggingface/transformers/pull/5885/diff?src=pr&el=tree-more) | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=footer). Last update [42fddac...00efd7d](https://codecov.io/gh/huggingface/transformers/pull/5885?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Another solution would be to just allow `prefix` element like in `TranslationPipeline`. The main difference will be that it becomes part of the display and it should not affect too much future plans such as suggested above with `padding_text`, `prefix` and `suffix`.\r\n\r\nI still think that `padding_text` as proposed here will be useful in the future to easily provide \"GPT-3 type\" widgets.", "I updated my [example](https://gist.github.com/borisdayma/89aaf1587390340add44c3c7081bffcd) of the feature with the case when it loads model config.", "Let me know if I can provide any additional information or make any changes", "Just checking if you need anything else from me @julien-c", "I’m off for a couple of days so I’ll let others chime in", "Enjoy your time off!", "@julien-c @mfuntowicz sorry I'm just pinging.\r\nI have a [demo](https://huggingface.co/huggingtweets/julien_c) I'm excited about and this feature just makes predictions so much better.\r\n*Note: not on master branch - dev version [here](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/feat-artifacts/huggingtweets-demo.ipynb)*", "Thoughts @patrickvonplaten @sshleifer? (text generation pipeline option)", "This PR adds the padding functionality from `run_generation` to pipelines (and allow for custom text).\r\n\r\nHowever the main idea would be to eventually allow \"GPT-3 style\" API's.\r\n\r\n\r\n\r\nIt could be renamed \"prefix\" but I suggest to make a distinction between padding and prefix where you would potentially want to display prefix & suffix in an interface (for example huggingface widget) but not necessarily the \"padding\".\r\n\r\nThis kind of API could be used for T5 and can let people create more modular inference examples.", "@sshleifer I gave more thoughts to it and I agree we should just name it `prefix` and maybe add a \"deprecated warning\" on `run_generation`.\r\nIt should be sufficient for now and we will always be able to add a more complex behavior in the future if necessary.\r\n\r\nNote: I don't understand the CI errors, quality is good on my side and the other tests errors seem unrelated to this PR", "style: you need to rebase and then `pip install black --upgrade`\r\nOther CI failures are spurious, you can ignore them.\r\n\r\nMy last request would be to add test coverage for the new argument in test_pipelines.py.\r\nsee:\r\n```\r\ntest_torch_text_generation\r\ntest_tf_text_generation\r\n```\r\n\r\n", "Changes:\r\n\r\n* renamed `padding_text` to `prefix` in pipelines and `run_generation.py` (with warning of deprecated argument)\r\n* since PR was made, we cannot tokenize empty text anymore so we now check for empty \"prefix\"\r\n* since PR was made, we cannot pass extra args to `generate` so we explicitly add `prefix` to pipeline `__call__` method\r\n* added tests for pipeline using prefix\r\n\r\nNotes:\r\n\r\n* we set default `prefix` to `None` (vs `\"\"`) as we could pass `\"\"` to overwrite model default value.\r\n* sorry I should have done a merge instead of rebase as it changes the history. New commits start at \"feat: rename padding_text to prefix\"", "Thanks @LysandreJik. Yes I noticed later that my results were better without the initial space as well.\r\n\r\nIntuitively it seemed to me like a waste of encoding differentiating between having a space or not (for example even after new lines) as the start of a document can simply be identified by the special token.\r\nI thought there were more examples without the space within the original training set which would help my fine-tuning, as long as I do the same both for training and inference.\r\nIn the end, I get my best results with just `<|endoftext|>` between each sample and stripping every extra space.", "@LysandreJik let me know if anything else is required on my side.\r\nSince it's a relatively large PR it can be hard to remain in sync with master.", "I think we can merge this: @LysandreJik I leave it to you to merge :-) ", "+ @mfuntowicz can you update the transformers dependency in the inference API when this lands on master, so that @borisdayma can showcase his magic ;)", "@mfuntowicz can you let me know when you update the inference API so I do some tests on the widget?", "@borisdayma Everything has been deployed on the inference API, servers are up. \r\n\r\nThanks for taking care of this 💪 🙏 ", "That was fast! Thanks it looks to be working great!" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=h1) Report\n> Merging [#5884](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/4b506a37e3e0ff679235961ba14dd9397843ef3a&el=desc) will **increase** coverage by `1.39%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5884 +/- ##\n==========================================\n+ Coverage 77.27% 78.67% +1.39% \n==========================================\n Files 146 146 \n Lines 26210 26210 \n==========================================\n+ Hits 20254 20620 +366 \n+ Misses 5956 5590 -366 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `43.98% <0.00%> (-49.38%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+1.25%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `95.53% <0.00%> (+69.51%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5884/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=footer). Last update [4b506a3...d67656b](https://codecov.io/gh/huggingface/transformers/pull/5884?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "This is really cool, thanks for sharing" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=h1) Report\n> Merging [#5883](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/3653d01f2af0389207f2239875a8ceae41bf0598&el=desc) will **increase** coverage by `1.36%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5883 +/- ##\n==========================================\n+ Coverage 77.26% 78.62% +1.36% \n==========================================\n Files 146 146 \n Lines 25948 25958 +10 \n==========================================\n+ Hits 20048 20409 +361 \n+ Misses 5900 5549 -351 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/configuration\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3V0aWxzLnB5) | `96.52% <ø> (ø)` | |\n| [src/transformers/configuration\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3hsbmV0LnB5) | `94.33% <ø> (ø)` | |\n| [src/transformers/modeling\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `81.75% <100.00%> (+0.26%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5883/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=footer). Last update [3653d01...0c73c6d](https://codecov.io/gh/huggingface/transformers/pull/5883?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Did @LysandreJik @patrickvonplaten or @sgugger get a chance to review this before merging, @TevenLeScao?", "@patrickvonplaten and @joeddav reviewed it (@LysandreJik was off), there's quite a bit of discussion in #5770 if you wanna take a look!" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=h1) Report\n> Merging [#5882](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/13be4872123094c37eb5fab939b38967b0ad2cd0&el=desc) will **increase** coverage by `0.21%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5882 +/- ##\n==========================================\n+ Coverage 78.27% 78.48% +0.21% \n==========================================\n Files 146 146 \n Lines 26210 26200 -10 \n==========================================\n+ Hits 20515 20563 +48 \n+ Misses 5695 5637 -58 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/configuration\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3V0aWxzLnB5) | `96.52% <ø> (ø)` | |\n| [src/transformers/configuration\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3hsbmV0LnB5) | `94.33% <ø> (ø)` | |\n| [src/transformers/modeling\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `81.49% <100.00%> (-0.27%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `43.98% <0.00%> (-49.38%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (+0.25%)` | :arrow_up: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `82.31% <0.00%> (+1.28%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5882/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `95.18% <0.00%> (+74.91%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=footer). Last update [13be487...1be083d](https://codecov.io/gh/huggingface/transformers/pull/5882?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=h1) Report\n> Merging [#5881](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/3653d01f2af0389207f2239875a8ceae41bf0598&el=desc) will **increase** coverage by `0.54%`.\n> The diff coverage is `100.00%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5881 +/- ##\n==========================================\n+ Coverage 77.26% 77.81% +0.54% \n==========================================\n Files 146 146 \n Lines 25948 25958 +10 \n==========================================\n+ Hits 20048 20198 +150 \n+ Misses 5900 5760 -140 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/configuration\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3V0aWxzLnB5) | `96.52% <ø> (ø)` | |\n| [src/transformers/configuration\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3hsbmV0LnB5) | `94.33% <ø> (ø)` | |\n| [src/transformers/modeling\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ194bG5ldC5weQ==) | `81.75% <100.00%> (+0.26%)` | :arrow_up: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5881/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=footer). Last update [3653d01...cfd7c66](https://codecov.io/gh/huggingface/transformers/pull/5881?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=h1) Report\n> Merging [#5879](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `1.11%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5879 +/- ##\n==========================================\n+ Coverage 77.54% 78.66% +1.11% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n+ Hits 20318 20609 +291 \n+ Misses 5882 5591 -291 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5879/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5879/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5879/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=footer). Last update [eae6d8d...0c9ff72](https://codecov.io/gh/huggingface/transformers/pull/5879?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=h1) Report\n> Merging [#5878](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `1.11%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5878 +/- ##\n==========================================\n+ Coverage 77.54% 78.66% +1.11% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n+ Hits 20318 20610 +292 \n+ Misses 5882 5590 -292 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5878/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5878/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.46% <0.00%> (ø)` | |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5878/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=footer). Last update [eae6d8d...e9b1654](https://codecov.io/gh/huggingface/transformers/pull/5878?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=h1) Report\n> Merging [#5877](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `0.78%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5877 +/- ##\n==========================================\n+ Coverage 77.54% 78.33% +0.78% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n+ Hits 20318 20524 +206 \n+ Misses 5882 5676 -206 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5877/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5877/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5877/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.21% <0.00%> (-1.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5877/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=footer). Last update [eae6d8d...c7a4977](https://codecov.io/gh/huggingface/transformers/pull/5877?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=h1) Report\n> Merging [#5876](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `0.93%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5876 +/- ##\n==========================================\n+ Coverage 77.54% 78.48% +0.93% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n+ Hits 20318 20564 +246 \n+ Misses 5882 5636 -246 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5876/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5876/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5876/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5876/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=footer). Last update [eae6d8d...befb998](https://codecov.io/gh/huggingface/transformers/pull/5876?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=h1) Report\n> Merging [#5875](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `0.81%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5875 +/- ##\n==========================================\n+ Coverage 77.54% 78.36% +0.81% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n+ Hits 20318 20531 +213 \n+ Misses 5882 5669 -213 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_gpt2.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9ncHQyLnB5) | `65.42% <0.00%> (-29.91%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `83.98% <0.00%> (-4.91%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_pytorch\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9weXRvcmNoX3V0aWxzLnB5) | `88.05% <0.00%> (-1.26%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.46% <0.00%> (-1.01%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5875/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=footer). Last update [eae6d8d...103b7af](https://codecov.io/gh/huggingface/transformers/pull/5875?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=h1) Report\n> Merging [#5873](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `0.79%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5873 +/- ##\n==========================================\n+ Coverage 77.54% 78.34% +0.79% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n+ Hits 20318 20527 +209 \n+ Misses 5882 5673 -209 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5873/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5873/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5873/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.51%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5873/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=footer). Last update [eae6d8d...f782f75](https://codecov.io/gh/huggingface/transformers/pull/5873?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=h1) Report\n> Merging [#5872](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `1.11%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5872 +/- ##\n==========================================\n+ Coverage 77.54% 78.66% +1.11% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n+ Hits 20318 20609 +291 \n+ Misses 5882 5591 -291 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5872/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5872/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5872/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=footer). Last update [eae6d8d...857c1c2](https://codecov.io/gh/huggingface/transformers/pull/5872?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=h1) Report\n> Merging [#5871](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/eae6d8d14f1d25d62c3fe9e7e410607bbaf69787&el=desc) will **increase** coverage by `1.11%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5871 +/- ##\n==========================================\n+ Coverage 77.54% 78.66% +1.11% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n+ Hits 20318 20609 +291 \n+ Misses 5882 5591 -291 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5871/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5871/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_mobilebert.py](https://codecov.io/gh/huggingface/transformers/pull/5871/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9tb2JpbGViZXJ0LnB5) | `96.77% <0.00%> (+73.38%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=footer). Last update [eae6d8d...823159f](https://codecov.io/gh/huggingface/transformers/pull/5871?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=h1) Report\n> Merging [#5870](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/dad5e12e54bc2cf80a24b3430b5c847fc213a73e&el=desc) will **decrease** coverage by `0.49%`.\n> The diff coverage is `85.71%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5870 +/- ##\n==========================================\n- Coverage 78.48% 77.98% -0.50% \n==========================================\n Files 146 146 \n Lines 26200 26203 +3 \n==========================================\n- Hits 20563 20435 -128 \n- Misses 5637 5768 +131 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ191dGlscy5weQ==) | `87.02% <85.71%> (-0.13%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/tokenization\\_xlnet.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25feGxuZXQucHk=) | `66.66% <0.00%> (-23.43%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5870/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=footer). Last update [dad5e12...02f32f1](https://codecov.io/gh/huggingface/transformers/pull/5870?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "This sounds reasonable", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=h1) Report\n> Merging [#5867](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/ba2400189b2242620868096ae49babf93bd9ce00&el=desc) will **decrease** coverage by `0.39%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5867 +/- ##\n==========================================\n- Coverage 78.48% 78.09% -0.40% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n- Hits 20564 20461 -103 \n- Misses 5636 5739 +103 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.51%)` | :arrow_down: |\n| [src/transformers/generation\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `97.11% <0.00%> (+0.28%)` | :arrow_up: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5867/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=footer). Last update [ba24001...4d35935](https://codecov.io/gh/huggingface/transformers/pull/5867?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "Awesome! Thank you @sepal ", "Could someone provide more detailed instructions to update the tokenizers in the version that solves the issue? I have a mac (10.13) and encountered the same problem. I tried through terminal to pip install tokenizers 0.8.1.rs2 but ended up with this error.\r\n\r\nERROR: Failed building wheel for tokenizers\r\nFailed to build tokenizers\r\nERROR: Could not build wheels for tokenizers which use PEP 517 and cannot be installed directly" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=h1) Report\n> Merging [#5866](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/7e6397a7d8e7433aa4c4cafba98e08e5c73f087c?el=desc) will **decrease** coverage by `0.67%`.\n> The diff coverage is `89.47%`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5866 +/- ##\n==========================================\n- Coverage 80.10% 79.42% -0.68% \n==========================================\n Files 156 156 \n Lines 28411 28426 +15 \n==========================================\n- Hits 22758 22578 -180 \n- Misses 5653 5848 +195 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/tokenization\\_t5.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdDUucHk=) | `95.32% <89.47%> (-1.42%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `22.58% <0.00%> (-72.26%)` | :arrow_down: |\n| [src/transformers/tokenization\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fcm9iZXJ0YS5weQ==) | `87.67% <0.00%> (-10.96%)` | :arrow_down: |\n| [src/transformers/tokenization\\_utils\\_base.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfYmFzZS5weQ==) | `86.58% <0.00%> (-7.19%)` | :arrow_down: |\n| [src/transformers/tokenization\\_transfo\\_xl.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdHJhbnNmb194bC5weQ==) | `38.73% <0.00%> (-3.76%)` | :arrow_down: |\n| [src/transformers/tokenization\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fb3BlbmFpLnB5) | `82.57% <0.00%> (-1.52%)` | :arrow_down: |\n| [src/transformers/tokenization\\_utils\\_fast.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fdXRpbHNfZmFzdC5weQ==) | `92.85% <0.00%> (-1.43%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `80.96% <0.00%> (-1.30%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.51%)` | :arrow_down: |\n| [src/transformers/tokenization\\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy90b2tlbml6YXRpb25fYmVydC5weQ==) | `91.07% <0.00%> (-0.45%)` | :arrow_down: |\n| ... and [2 more](https://codecov.io/gh/huggingface/transformers/pull/5866/diff?src=pr&el=tree-more) | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=footer). Last update [7e6397a...d977bff](https://codecov.io/gh/huggingface/transformers/pull/5866?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "FWIW, I get identical results of `27.84` with this branch and the master.", "Happy to eventually remove the check to see if it's already there.", "I think we can keep it like this right now with a warning for future versions. It would create a breaking change to users, and I feel it would be especially hard to debug an unknown drop in performance due to an additional token being added, right?", "Will this behavior cause problems for the unsupervised setting? Per the [docs](https://huggingface.co/transformers/model_doc/t5.html#training), `</s>` is not added during denoising training:\r\n\r\n```\r\ninput_ids = tokenizer.encode('The <extra_id_0> walks in <extra_id_1> park', return_tensors='pt')\r\nlabels = tokenizer.encode('<extra_id_0> cute dog <extra_id_1> the <extra_id_2> </s>', return_tensors='pt')\r\nmodel(input_ids=input_ids, labels=labels)\r\n```\r\n\r\nNot sure if this will cause problems. (Also, Aa a somewhat related question, should the sentinel tokens in the `labels` be excluded from the loss in this setting, as I believe is the case with `[MASK]` in BERT?).", "I'm not sure about either question:\r\nmade an issue verifying the docs: https://github.com/huggingface/transformers/issues/7904\r\nFeel free to make an issue about the sentinel tokens question. I'd tag thomwolf/patrickvonplaten.\r\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=h1) Report\n> Merging [#5865](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/529850ae7bca0ff388778c3c0d66240834cf56c3&el=desc) will **not change** coverage.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5865 +/- ##\n=======================================\n Coverage 78.48% 78.48% \n=======================================\n Files 146 146 \n Lines 26200 26200 \n=======================================\n Hits 20563 20563 \n Misses 5637 5637 \n```\n\n\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=footer). Last update [529850a...259bc29](https://codecov.io/gh/huggingface/transformers/pull/5865?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=h1) Report\n> Merging [#5864](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/ba2400189b2242620868096ae49babf93bd9ce00&el=desc) will **decrease** coverage by `0.39%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5864 +/- ##\n==========================================\n- Coverage 78.48% 78.09% -0.40% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n- Hits 20564 20460 -104 \n- Misses 5636 5740 +104 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.51%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `81.49% <0.00%> (+0.29%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5864/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=footer). Last update [ba24001...84114cc](https://codecov.io/gh/huggingface/transformers/pull/5864?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "Your `requests` syntax is incorrect, you're not actually sending the request body as json.\r\n\r\nThis works:\r\n```python\r\nanswer = requests.post(\r\n \"https://api-inference.huggingface.co/models/distilbert-base-cased-distilled-squad\",\r\n json={\"question\": \"What is my name?\", \"context\": \"My name is Batik\"},\r\n)\r\n```\r\n\r\nRegarding the stack trace, I do not see an issue with returning it (it's json-encoded), happy to discuss more." ]

[]

[]

[ "cc @nateraw ", "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=h1) Report\n> Merging [#5861](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/529850ae7bca0ff388778c3c0d66240834cf56c3&el=desc) will **decrease** coverage by `0.25%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5861 +/- ##\n==========================================\n- Coverage 78.48% 78.23% -0.26% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n- Hits 20563 20497 -66 \n- Misses 5637 5703 +66 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `64.90% <0.00%> (-33.90%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `83.70% <0.00%> (-2.51%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3V0aWxzLnB5) | `97.11% <0.00%> (+0.28%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5861/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `95.53% <0.00%> (+69.51%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=footer). Last update [529850a...c7b897e](https://codecov.io/gh/huggingface/transformers/pull/5861?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "better way in #5865 " ]

[]

[]

[]

[ { "id": 1314768611, "node_id": "MDU6TGFiZWwxMzE0NzY4NjEx", "url": "https://api.github.com/repos/huggingface/transformers/labels/wontfix", "name": "wontfix", "color": "ffffff", "default": true, "description": null } ]

[ { "login": "sshleifer", "id": 6045025, "node_id": "MDQ6VXNlcjYwNDUwMjU=", "avatar_url": "https://avatars.githubusercontent.com/u/6045025?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sshleifer", "html_url": "https://github.com/sshleifer", "followers_url": "https://api.github.com/users/sshleifer/followers", "following_url": "https://api.github.com/users/sshleifer/following{/other_user}", "gists_url": "https://api.github.com/users/sshleifer/gists{/gist_id}", "starred_url": "https://api.github.com/users/sshleifer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sshleifer/subscriptions", "organizations_url": "https://api.github.com/users/sshleifer/orgs", "repos_url": "https://api.github.com/users/sshleifer/repos", "events_url": "https://api.github.com/users/sshleifer/events{/privacy}", "received_events_url": "https://api.github.com/users/sshleifer/received_events", "type": "User", "site_admin": false } ]

[ "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=h1) Report\n> Merging [#5858](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/615be03f9d961c0c9722fe10e7830e011066772e&el=desc) will **decrease** coverage by `0.20%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5858 +/- ##\n==========================================\n- Coverage 78.66% 78.46% -0.21% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n- Hits 20611 20558 -53 \n- Misses 5589 5642 +53 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5858/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5858/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `84.96% <0.00%> (-1.76%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5858/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=footer). Last update [615be03...713f748](https://codecov.io/gh/huggingface/transformers/pull/5858?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "great catch! was fixed on master by #5798 so I don't think we need this. Let us know if further issues!" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=h1) Report\n> Merging [#5857](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/615be03f9d961c0c9722fe10e7830e011066772e&el=desc) will **decrease** coverage by `0.58%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5857 +/- ##\n==========================================\n- Coverage 78.66% 78.08% -0.59% \n==========================================\n Files 146 146 \n Lines 26200 26200 \n==========================================\n- Hits 20611 20459 -152 \n- Misses 5589 5741 +152 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/modeling\\_tf\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9vcGVuYWkucHk=) | `20.27% <0.00%> (-74.92%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `26.02% <0.00%> (-69.52%)` | :arrow_down: |\n| [src/transformers/modeling\\_openai.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19vcGVuYWkucHk=) | `81.02% <0.00%> (-1.29%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `85.96% <0.00%> (-0.76%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_distilbert.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9kaXN0aWxiZXJ0LnB5) | `98.79% <0.00%> (+33.89%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_roberta.py](https://codecov.io/gh/huggingface/transformers/pull/5857/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9yb2JlcnRhLnB5) | `93.36% <0.00%> (+49.37%)` | :arrow_up: |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=footer). Last update [615be03...e7f330a](https://codecov.io/gh/huggingface/transformers/pull/5857?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n" ]

[ { "id": 1838412367, "node_id": "MDU6TGFiZWwxODM4NDEyMzY3", "url": "https://api.github.com/repos/huggingface/transformers/labels/model%20card", "name": "model card", "color": "92d5f4", "default": false, "description": "Related to pretrained model cards" } ]

[]

[ "👍 awesome\r\n", "Even if dataset is not in `nlp` yet, you can still link to it, as it will display a missing dataset page with a call to action to add it to nlp: e.g. https://huggingface.co/datasets/dlkfjldsfjdslf" ]

[]

[]

[ "# [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=h1) Report\n> Merging [#5855](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=desc) into [master](https://codecov.io/gh/huggingface/transformers/commit/8279471506fab5733dab3e2d3a1542010c976d8a?el=desc) will **decrease** coverage by `2.76%`.\n> The diff coverage is `n/a`.\n\n[](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=tree)\n\n```diff\n@@ Coverage Diff @@\n## master #5855 +/- ##\n==========================================\n- Coverage 79.87% 77.11% -2.77% \n==========================================\n Files 181 181 \n Lines 35788 35788 \n==========================================\n- Hits 28587 27597 -990 \n- Misses 7201 8191 +990 \n```\n\n\n| [Impacted Files](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=tree) | Coverage Δ | |\n|---|---|---|\n| [src/transformers/configuration\\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9jb25maWd1cmF0aW9uX3JlZm9ybWVyLnB5) | `21.62% <0.00%> (-78.38%)` | :arrow_down: |\n| [src/transformers/modeling\\_reformer.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19yZWZvcm1lci5weQ==) | `16.71% <0.00%> (-77.89%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_funnel.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9mdW5uZWwucHk=) | `18.53% <0.00%> (-75.51%)` | :arrow_down: |\n| [src/transformers/modeling\\_tf\\_electra.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl9lbGVjdHJhLnB5) | `24.25% <0.00%> (-73.56%)` | :arrow_down: |\n| [src/transformers/activations\\_tf.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9hY3RpdmF0aW9uc190Zi5weQ==) | `54.16% <0.00%> (-20.84%)` | :arrow_down: |\n| [src/transformers/generation\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9nZW5lcmF0aW9uX3RmX3V0aWxzLnB5) | `86.21% <0.00%> (-0.26%)` | :arrow_down: |\n| [src/transformers/file\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9maWxlX3V0aWxzLnB5) | `83.13% <0.00%> (-0.25%)` | :arrow_down: |\n| [src/transformers/modeling\\_bert.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19iZXJ0LnB5) | `88.48% <0.00%> (+0.16%)` | :arrow_up: |\n| [src/transformers/modeling\\_tf\\_utils.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ190Zl91dGlscy5weQ==) | `86.90% <0.00%> (+0.63%)` | :arrow_up: |\n| [src/transformers/modeling\\_auto.py](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree#diff-c3JjL3RyYW5zZm9ybWVycy9tb2RlbGluZ19hdXRvLnB5) | `87.04% <0.00%> (+1.03%)` | :arrow_up: |\n| ... and [5 more](https://codecov.io/gh/huggingface/transformers/pull/5855/diff?src=pr&el=tree-more) | |\n\n------\n\n[Continue to review full report at Codecov](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=continue).\n> **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta)\n> `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data`\n> Powered by [Codecov](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=footer). Last update [8279471...d60525f](https://codecov.io/gh/huggingface/transformers/pull/5855?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments).\n", "It's more general than vocab.txt or vocab.json so I'll let @sgugger chime in as I think he wrote those lines.", "I didn't know that when I wrote it (and the file was vocab.json for me ;-) ). This looks good to me with the suggestion, thanks for clarifying!", "Looks good on my side, the only file removed from my example would be `training_args.bin` which makes sense.", "Let me know if I need to do any other changes", "This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.\n", "Oops, sorry this got lost in my notifications. Don't hesitate to ping me if it happens again @borisdayma, will merge as soon as the CI is green." ]