DanielPFlorian/Transformers-Github-Issues · Datasets at Hugging Face

https://api.github.com/repos/huggingface/transformers/issues/24418

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24418/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24418/comments

https://api.github.com/repos/huggingface/transformers/issues/24418/events

https://github.com/huggingface/transformers/issues/24418

1,769,164,033

I_kwDOCUB6oc5pc00B

24,418

Review added

{ "login": "repo-reviews", "id": 135327276, "node_id": "U_kgDOCBDuLA", "avatar_url": "https://avatars.githubusercontent.com/u/135327276?v=4", "gravatar_id": "", "url": "https://api.github.com/users/repo-reviews", "html_url": "https://github.com/repo-reviews", "followers_url": "https://api.github.com/users/repo-reviews/followers", "following_url": "https://api.github.com/users/repo-reviews/following{/other_user}", "gists_url": "https://api.github.com/users/repo-reviews/gists{/gist_id}", "starred_url": "https://api.github.com/users/repo-reviews/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/repo-reviews/subscriptions", "organizations_url": "https://api.github.com/users/repo-reviews/orgs", "repos_url": "https://api.github.com/users/repo-reviews/repos", "events_url": "https://api.github.com/users/repo-reviews/events{/privacy}", "received_events_url": "https://api.github.com/users/repo-reviews/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

### Thank you for building **transformers**! @AuroraCoboAguilera created a review titled: *Hugging face, the accelerator to develop your own NLP* on [repo-reviews.github.io](https://repo-reviews.github.io) to share their experience using **transformers**. [link to review](https://repo-reviews.github.io//reviews/2023-06-21_AuroraCoboAguilera_huggingface_transformers) If you would like to help your super-users share their experiences using your repo, add a [badge](https://github.com/repo-reviews/repo-reviews.github.io#add-badges) to your README.md. We hope that sharing these experiences helps your users **increase their productivity**. -- Please be kind, I’m a human!

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24418/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24418/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24417

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24417/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24417/comments

https://api.github.com/repos/huggingface/transformers/issues/24417/events

https://github.com/huggingface/transformers/pull/24417

1,769,121,791

PR_kwDOCUB6oc5ToHMm

24,417

Skip `test_conditional_generation_pt_pix2struct` in Past CI (torch < 1.11)

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? Same as in #24270, but for a test inside pipeline test in `ImageToTextPipelineTests`.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24417/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24417/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24417", "html_url": "https://github.com/huggingface/transformers/pull/24417", "diff_url": "https://github.com/huggingface/transformers/pull/24417.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24417.patch", "merged_at": 1687440853000 }

https://api.github.com/repos/huggingface/transformers/issues/24416

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24416/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24416/comments

https://api.github.com/repos/huggingface/transformers/issues/24416/events

https://github.com/huggingface/transformers/pull/24416

1,769,058,737

PR_kwDOCUB6oc5Tn5bb

24,416

[`bnb`] Fix bnb serialization issue with new release

{ "login": "younesbelkada", "id": 49240599, "node_id": "MDQ6VXNlcjQ5MjQwNTk5", "avatar_url": "https://avatars.githubusercontent.com/u/49240599?v=4", "gravatar_id": "", "url": "https://api.github.com/users/younesbelkada", "html_url": "https://github.com/younesbelkada", "followers_url": "https://api.github.com/users/younesbelkada/followers", "following_url": "https://api.github.com/users/younesbelkada/following{/other_user}", "gists_url": "https://api.github.com/users/younesbelkada/gists{/gist_id}", "starred_url": "https://api.github.com/users/younesbelkada/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/younesbelkada/subscriptions", "organizations_url": "https://api.github.com/users/younesbelkada/orgs", "repos_url": "https://api.github.com/users/younesbelkada/repos", "events_url": "https://api.github.com/users/younesbelkada/events{/privacy}", "received_events_url": "https://api.github.com/users/younesbelkada/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24416). All of your documentation changes will be reflected on that endpoint." ]

1,687

CONTRIBUTOR

null

# What does this PR do? Fixes the failing bnb tests in: https://github.com/huggingface/transformers/actions/runs/5329457026/jobs/9655258851 The recent release of bitsandbytes slightly broke the serialization mechanism for int8 models. In the new release, bitsandbytes has introduced a new way of serializing int8 weights that are more memory efficient and avoid OOM issues when saving for instance PEFT models. https://github.com/TimDettmers/bitsandbytes/pull/503 That PR introduced a new paradigm, when saving int8 state dict, [that state dict will contain some string values](https://github.com/TimDettmers/bitsandbytes/pull/503/files#diff-4d235c7e595546c6656c229dfa139298ce6602b356c2d0bafcb2352eb2cfae79R360-R363) to store some metadata information related to the quantized format. Therefore the fix should be to slightly adapt the `shard_checkpoint` method (that is called regardless of if the model is sharded or not) by adding a new argument `state_dict_contains_metadata` that will skip manipulating the `weight` variable that is no longer a tensor but a string. We constraint `state_dict_contains_metadata` to be applicable only in the int8 case to ensure we don't break anything else cc @amyeroberts

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24416/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24416/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24416", "html_url": "https://github.com/huggingface/transformers/pull/24416", "diff_url": "https://github.com/huggingface/transformers/pull/24416.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24416.patch", "merged_at": 1687441239000 }

https://api.github.com/repos/huggingface/transformers/issues/24415

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24415/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24415/comments

https://api.github.com/repos/huggingface/transformers/issues/24415/events

https://github.com/huggingface/transformers/pull/24415

1,768,989,268

PR_kwDOCUB6oc5TnqSi

24,415

fix the grad_acc issue at epoch boundaries

{ "login": "pacman100", "id": 13534540, "node_id": "MDQ6VXNlcjEzNTM0NTQw", "avatar_url": "https://avatars.githubusercontent.com/u/13534540?v=4", "gravatar_id": "", "url": "https://api.github.com/users/pacman100", "html_url": "https://github.com/pacman100", "followers_url": "https://api.github.com/users/pacman100/followers", "following_url": "https://api.github.com/users/pacman100/following{/other_user}", "gists_url": "https://api.github.com/users/pacman100/gists{/gist_id}", "starred_url": "https://api.github.com/users/pacman100/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/pacman100/subscriptions", "organizations_url": "https://api.github.com/users/pacman100/orgs", "repos_url": "https://api.github.com/users/pacman100/repos", "events_url": "https://api.github.com/users/pacman100/events{/privacy}", "received_events_url": "https://api.github.com/users/pacman100/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "I believe this PR also fixes #24245", "@amyeroberts we coordinate Accelerate releases to be a day or two before `transformers`, so there shouldn't be an issue there :) \r\n\r\n(Though @pacman100 we should do the version check like we've done before with these fixes 😬 )", "Hello,\r\n\r\nHow do i install this? I expected that PR means some kind of transformers update, in this case there should be install link such as git+https://github.com/huggingface/transformers@de9255de27abfcae4a1f816b904915f0b1e23cd9\r\n\r\n\r\n\r\n\r\n \r\n", "Hello @Oxi84, you can install this once it gets merged via `pip install git+https://github.com/huggingface/transformers` and `pip install git+https://github.com/huggingface/accelerate`", "Just completed a training run with this PR and can confirm that the issue didn't occur. Thanks for the fix!" ]

1,687

CONTRIBUTOR

null

# What does this PR do? Should solve accumulating via epoch when using Accelerate (seen in https://github.com/huggingface/transformers/issues/23935#issuecomment-1588134562). Requires https://github.com/huggingface/accelerate/pull/1624 Fixes # (issue)

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24415/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24415/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24415", "html_url": "https://github.com/huggingface/transformers/pull/24415", "diff_url": "https://github.com/huggingface/transformers/pull/24415.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24415.patch", "merged_at": 1687522387000 }

https://api.github.com/repos/huggingface/transformers/issues/24414

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24414/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24414/comments

https://api.github.com/repos/huggingface/transformers/issues/24414/events

https://github.com/huggingface/transformers/issues/24414

1,768,977,236

I_kwDOCUB6oc5pcHNU

24,414

Trouble fine-tuning zero-shot image classification model

{ "login": "moon001light", "id": 137312482, "node_id": "U_kgDOCC844g", "avatar_url": "https://avatars.githubusercontent.com/u/137312482?v=4", "gravatar_id": "", "url": "https://api.github.com/users/moon001light", "html_url": "https://github.com/moon001light", "followers_url": "https://api.github.com/users/moon001light/followers", "following_url": "https://api.github.com/users/moon001light/following{/other_user}", "gists_url": "https://api.github.com/users/moon001light/gists{/gist_id}", "starred_url": "https://api.github.com/users/moon001light/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/moon001light/subscriptions", "organizations_url": "https://api.github.com/users/moon001light/orgs", "repos_url": "https://api.github.com/users/moon001light/repos", "events_url": "https://api.github.com/users/moon001light/events{/privacy}", "received_events_url": "https://api.github.com/users/moon001light/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @moon001light, thanks for raising this issue.\r\n\r\nModels that can be loaded with the `AutoXxx` API will have a shared input and output structure. In the example notebook, the model is loaded using `AutoModelForImageClassification`. Models loaded with `AutoModelForZeroShotImageClassification` have a different set of expected inputs, in particular, they don't accept `labels` and expect `input_ids`. Here's a guide on performing the zero-shot image classification task using transformers: https://huggingface.co/docs/transformers/tasks/zero_shot_image_classification. \r\n\r\nAdd more example scripts is on my to-do list. I'll make sure to include this task! ", "Thanks for the quick reply @amyeroberts , the link you gave seems to be a high level overview. Is there anything else besides replacing `labels` with `input_ids`? Could you point me to the code so that I can see the full list of expected inputs for `AutoModelForZeroShotImageClassification`? Appreciate your patience thanks :)", "@moon001light If all you want is to train this particular checkpoint, then I would look directly at the architecture it loads, which in this [case is CLIPModel](https://github.com/huggingface/transformers/blob/ea91c2adca842da3d2f87e094504fa7d66a7008a/src/transformers/models/clip/modeling_clip.py#L1082). \r\n\r\nMore generally, the auto groups are [defined in `modeling_auto.py`](https://github.com/huggingface/transformers/blob/main/src/transformers/models/auto/modeling_auto.py). For [AutoModelForZeroShotClassification](https://github.com/huggingface/transformers/blob/ea91c2adca842da3d2f87e094504fa7d66a7008a/src/transformers/models/auto/modeling_auto.py#L1256C7-L1256C46) the model architectures that can be used are listed under [MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING_NAMES](https://github.com/huggingface/transformers/blob/ea91c2adca842da3d2f87e094504fa7d66a7008a/src/transformers/models/auto/modeling_auto.py#L980). \r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n", "A bit of a newbie here, after looking into it I am still not sure how to properly set `input_ids`. This is what I tried:\r\n\r\n```\r\nfrom transformers import AutoTokenizer\r\ntokenizer = AutoTokenizer.from_pretrained(model_checkpoint)\r\n\r\ndef collate_fn(examples):\r\n pixel_values = torch.stack([example[\"pixel_values\"] for example in examples])\r\n input_ids = tokenizer(examples[\"label\"], padding=\"max_length\", truncation=True)\r\n return {\"pixel_values\": pixel_values, \"input_ids\": input_ids} \r\n```\r\n\r\nBut my `input_ids` is wrong :( Not really sure what `input_ids` should be. Looking forward to that example script of yours so I can learn @amyeroberts . Appreciate any help I can get :)", "@moon001light I suggest inspecting the objects at each step to see what they are and contain. For example, the output of \r\n\r\n```python\r\ntokenizer(examples[\"label\"], padding=\"max_length\", truncation=True)\r\n```\r\n\r\nis not `input_ids`, but rather a dictionary that contains `input_ids`.", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,691

NONE

null

### System Info transformers 4.30.2 python 3.11.4 ### Who can help? @amyeroberts @sgugger ### Information - [X] The official example scripts - [X] My own modified scripts ### Tasks - [X] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction I am trying to fine-tune a **zero-shot** image classifier, I am following this example for reference: https://github.com/huggingface/notebooks/blob/main/examples/image_classification.ipynb What I changed from the notebook: The checkpoint I am starting from: `laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K` For the model I am using `AutoModelForZeroShotImageClassification` like this: ``` model = AutoModelForZeroShotImageClassification.from_pretrained( model_checkpoint, label2id=label2id, id2label=id2label, ignore_mismatched_sizes = True, ) ``` When I run `trainer.train()`, I get this error: > TypeError: CLIPModel.forward() got an unexpected keyword argument 'labels' ### Expected behavior With my changes to the notebook, it should fine-tune for zero-shot pre-trained model `laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K`

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24414/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24414/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24413

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24413/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24413/comments

https://api.github.com/repos/huggingface/transformers/issues/24413/events

https://github.com/huggingface/transformers/pull/24413

1,768,891,339

PR_kwDOCUB6oc5TnVML

24,413

Create Pretrained module

{ "login": "Ntrystan", "id": 95559349, "node_id": "U_kgDOBbIetQ", "avatar_url": "https://avatars.githubusercontent.com/u/95559349?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Ntrystan", "html_url": "https://github.com/Ntrystan", "followers_url": "https://api.github.com/users/Ntrystan/followers", "following_url": "https://api.github.com/users/Ntrystan/following{/other_user}", "gists_url": "https://api.github.com/users/Ntrystan/gists{/gist_id}", "starred_url": "https://api.github.com/users/Ntrystan/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Ntrystan/subscriptions", "organizations_url": "https://api.github.com/users/Ntrystan/orgs", "repos_url": "https://api.github.com/users/Ntrystan/repos", "events_url": "https://api.github.com/users/Ntrystan/events{/privacy}", "received_events_url": "https://api.github.com/users/Ntrystan/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

1,687

NONE

null

# What does this PR do?   Fixes # (issue) ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [x] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24413/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24413/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24413", "html_url": "https://github.com/huggingface/transformers/pull/24413", "diff_url": "https://github.com/huggingface/transformers/pull/24413.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24413.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24412

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24412/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24412/comments

https://api.github.com/repos/huggingface/transformers/issues/24412/events

https://github.com/huggingface/transformers/pull/24412

1,768,846,875

PR_kwDOCUB6oc5TnL29

24,412

Removed @torch.no_grad() and in-place operations in optimizers for backwards

{ "login": "shirayu", "id": 963961, "node_id": "MDQ6VXNlcjk2Mzk2MQ==", "avatar_url": "https://avatars.githubusercontent.com/u/963961?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shirayu", "html_url": "https://github.com/shirayu", "followers_url": "https://api.github.com/users/shirayu/followers", "following_url": "https://api.github.com/users/shirayu/following{/other_user}", "gists_url": "https://api.github.com/users/shirayu/gists{/gist_id}", "starred_url": "https://api.github.com/users/shirayu/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/shirayu/subscriptions", "organizations_url": "https://api.github.com/users/shirayu/orgs", "repos_url": "https://api.github.com/users/shirayu/repos", "events_url": "https://api.github.com/users/shirayu/events{/privacy}", "received_events_url": "https://api.github.com/users/shirayu/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "I consider this pull request is insufficient because tests have failed.\r\nI welcome comments on the fix!", "Transformers is primarily a library of models, not optimizers. I would recommend not using the AdamW/Adafactor from the library (which are going to be removed in the next major version) and use another implementation :-)", "Thank you for the comment!\r\n\r\nAll right, I found an implementation `fairseq.optim.adafactor.Adafactor` and I will use it.\r\n\r\nhttps://github.com/pytorch/pytorch/issues/30446" ]

1,687

CONTRIBUTOR

null

# What does this PR do? In #23417, two``@torch.no_grad()`` lines were added before ```python def step(self, closure: Callable = None): ``` in class `AdamW` and `Adafactor`. However, I think this should not be because this causes errors in backward. Other optimizers in Pytorch have ``@_use_grad_for_differentiable`` lines before `def step`. ```python @_use_grad_for_differentiable def step(self, closure=None): ``` - Examples - https://github.com/pytorch/pytorch/blob/430cb3e1600e0aca742105a2cdf4a01d901955dd/torch/optim/adam.py#L122-L123 - https://github.com/pytorch/pytorch/blob/430cb3e1600e0aca742105a2cdf4a01d901955dd/torch/optim/adamw.py#L149-L150 I also replaced in-place operations into assignments for backwards. ## Question Should the following line also be replaced? https://github.com/huggingface/transformers/blob/6ce6d62b6f20040129ec9831e7c4f6576402ea42/src/transformers/optimization.py#L728 ## Context I faced this problem when using PyTorch Lightning 2.0.3 with `transformers.optimization.Adafactor` as an optimizer. With 3cf01b206 (one previous commit), this error did not occur. ```txt File "/path/to/lib/python3.10/site-packages/pytorch_lightning/strategies/strategy.py", line 225, in optimizer_step return self.precision_plugin.optimizer_step(optimizer, model=model, closure=closure, **kwargs) File "/path/to/lib/python3.10/site-packages/pytorch_lightning/plugins/precision/precision_plugin.py", line 114, in optimizer_step return optimizer.step(closure=closure, **kwargs) File "/path/to/lib/python3.10/site-packages/torch/optim/optimizer.py", line 280, in wrapper out = func(*args, **kwargs) File "/path/to/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context return func(*args, **kwargs) File "/path/to/lib/python3.10/site-packages/transformers/optimization.py", line 649, in step loss = closure() File "/path/to/lib/python3.10/site-packages/pytorch_lightning/plugins/precision/precision_plugin.py", line 101, in _wrap_closure closure_result = closure() File "/path/to/lib/python3.10/site-packages/pytorch_lightning/loops/optimization/automatic.py", line 140, in __call__ self._result = self.closure(*args, **kwargs) File "/path/to/lib/python3.10/site-packages/pytorch_lightning/loops/optimization/automatic.py", line 135, in closure self._backward_fn(step_output.closure_loss) File "/path/to/lib/python3.10/site-packages/pytorch_lightning/loops/optimization/automatic.py", line 232, in backward_fn call._call_strategy_hook(self.trainer, "backward", loss, optimizer) File "/path/to/lib/python3.10/site-packages/pytorch_lightning/trainer/call.py", line 287, in _call_strategy_hook output = fn(*args, **kwargs) File "/path/to/lib/python3.10/site-packages/pytorch_lightning/strategies/strategy.py", line 200, in backward self.precision_plugin.backward(closure_loss, self.lightning_module, optimizer, *args, **kwargs) File "/path/to/lib/python3.10/site-packages/pytorch_lightning/plugins/precision/precision_plugin.py", line 67, in backward model.backward(tensor, *args, **kwargs) File "/path/to/lib/python3.10/site-packages/pytorch_lightning/core/module.py", line 1046, in backward loss.backward(*args, **kwargs) File "/path/to/lib/python3.10/site-packages/torch/_tensor.py", line 487, in backward torch.autograd.backward( File "/path/to/lib/python3.10/site-packages/torch/autograd/__init__.py", line 200, in backward Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass RuntimeError: element 0 of tensors does not require grad and does not have a grad_fn Epoch 0: 0%| | 0/2852 [00:01<?, ?it/s] ``` ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.  - PyTorch: @sgugger

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24412/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24412/timeline

null

true

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24412", "html_url": "https://github.com/huggingface/transformers/pull/24412", "diff_url": "https://github.com/huggingface/transformers/pull/24412.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24412.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24411

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24411/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24411/comments

https://api.github.com/repos/huggingface/transformers/issues/24411/events

https://github.com/huggingface/transformers/issues/24411

1,768,666,636

I_kwDOCUB6oc5pa7YM

24,411

GPTJForCausalLM with instruction provided on tutorial doesn't load on 4090

{ "login": "km5ar", "id": 54015474, "node_id": "MDQ6VXNlcjU0MDE1NDc0", "avatar_url": "https://avatars.githubusercontent.com/u/54015474?v=4", "gravatar_id": "", "url": "https://api.github.com/users/km5ar", "html_url": "https://github.com/km5ar", "followers_url": "https://api.github.com/users/km5ar/followers", "following_url": "https://api.github.com/users/km5ar/following{/other_user}", "gists_url": "https://api.github.com/users/km5ar/gists{/gist_id}", "starred_url": "https://api.github.com/users/km5ar/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/km5ar/subscriptions", "organizations_url": "https://api.github.com/users/km5ar/orgs", "repos_url": "https://api.github.com/users/km5ar/repos", "events_url": "https://api.github.com/users/km5ar/events{/privacy}", "received_events_url": "https://api.github.com/users/km5ar/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @km5ar \r\nThanks for the issue\r\n\r\nThis is because google colab instances have a relatively low CPU RAM (<24GB) and the GPT-J model stored on the Hub at `EleutherAI/gpt-j-6b` are actually in float32 (24GB). Therefore from_pretrained will try to download that large file and crashed. Moreover, for large models it is recommended to load a model either with `low_cpu_mem_usage=True` or `device_map=\"auto\"` as by default from_pretrained will initialize a random model with the same number of paramaters then try to populate the model with the dowloaded weights. Using `low_cpu_mem_usage=True` will avoid that step and not create a dummy random model at the beginning of the from_pretrained call.\r\n\r\nTo run gpt-j 6B on google colab consider using this repo: [`ybelkada/gpt-j-6b-sharded-bf16`](https://huggingface.co/ybelkada/gpt-j-6b-sharded-bf16) if you want to load the model in bf16 (by passing `torch_dtype=torch.bfloat16` ) or this repo: [`philschmid/gpt-j-6B-fp16-sharded`](https://huggingface.co/philschmid/gpt-j-6B-fp16-sharded) if you want to run the model in `float16` (by passing `torch_dtype=torch.float16`).", "@younesbelkada \r\nThanks for the answer\r\n\r\nHowever, my question is \r\nif you see code, I did use torch.float16.\r\nand please see the following screenshot,\r\nI was copy the code from the tutorial from official guide, which clearly said \"The model should fit on 16GB GPU for inference.\"\r\n\r\nI was using a 4090 in my local machine.\r\n![image](https://github.com/huggingface/transformers/assets/54015474/57cdf8b4-0281-4afa-a106-46852d880fbf)\r\n![1687434430321](https://github.com/huggingface/transformers/assets/54015474/e88bd673-671c-4461-a9e0-3a6ef70ec560)\r\n\r\n", "@km5ar \r\nThanks ! \r\nThe statement \"The model should fit on 16GB GPU\" is still true. As explained above, the culprit is that `low_cpu_mem_uage` is set by default to `False` therefore blows up the Google Colab's CPU memory that is relatively low for that model due to the reasons I have detailed. Also, loading a checkpoint that is sharded helps to not getting those errors as the shards are processed one by one and deleted afterwards. \r\n\r\n<img width=\"560\" alt=\"Screenshot 2023-06-22 at 14 06 53\" src=\"https://github.com/huggingface/transformers/assets/49240599/d919490b-e915-431b-8d64-08c78762adb9\">\r\n\r\nAs you can see from the screen shot above, `low_cpu_mem_usage=False` combined with that checkpoint will force the program to allocate 12+12=24GB CPU memory before moving the weights on GPU, hence your error.\r\n\r\nCan you try to call `from_pretrained` with `low_cpu_mem_usage=True`, and use [philschmid/gpt-j-6B-fp16-sharded](https://huggingface.co/philschmid/gpt-j-6B-fp16-sharded) instead of the original repository?\r\n\r\nThanks", "Thank you!!! that's very helpful!" ]

1,687

NONE

null

### System Info GPU 4090 - `transformers` version: 4.30.2 - Platform: Linux-5.10.16.3-microsoft-standard-WSL2-x86_64-with-glibc2.35 - Python version: 3.9.16 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): GPU 2.0.1+cu117 (True) - Using GPU in script?: <4090> - Using distributed or parallel set-up in script?: <no> ### Who can help? @ArthurZucker @younesbelkada ### Information - [X] The official example scripts - [ ] My own modified scripts ### Tasks - [X] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction ![image](https://github.com/huggingface/transformers/assets/54015474/74568eac-b7c5-40c3-ba4e-b118758d4761) ![image](https://github.com/huggingface/transformers/assets/54015474/530a6f53-a379-4acb-a6c9-72bf1381be30) ![image](https://github.com/huggingface/transformers/assets/54015474/f8aa219f-00e9-47b6-ad38-35da7ba8c7b5) ### Expected behavior from the official tutorial, it said it should work with 16 GB of Ram, but 4090 has 24G of ram. the kernel is died. Can anyone help me out? https://huggingface.co/docs/transformers/model_doc/gptj error: Kernel Restarting The kernel for gpt-j-6b/1.ipynb appears to have died. It will restart automatically.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24411/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24411/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24410

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24410/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24410/comments

https://api.github.com/repos/huggingface/transformers/issues/24410/events

https://github.com/huggingface/transformers/issues/24410

1,768,659,684

I_kwDOCUB6oc5pa5rk

24,410

Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cuda:3!

{ "login": "karths8", "id": 47289950, "node_id": "MDQ6VXNlcjQ3Mjg5OTUw", "avatar_url": "https://avatars.githubusercontent.com/u/47289950?v=4", "gravatar_id": "", "url": "https://api.github.com/users/karths8", "html_url": "https://github.com/karths8", "followers_url": "https://api.github.com/users/karths8/followers", "following_url": "https://api.github.com/users/karths8/following{/other_user}", "gists_url": "https://api.github.com/users/karths8/gists{/gist_id}", "starred_url": "https://api.github.com/users/karths8/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/karths8/subscriptions", "organizations_url": "https://api.github.com/users/karths8/orgs", "repos_url": "https://api.github.com/users/karths8/repos", "events_url": "https://api.github.com/users/karths8/events{/privacy}", "received_events_url": "https://api.github.com/users/karths8/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "You cannot use `device_map=\"auto\"` with the `to` method afterward as you do. The model will be split up on the GPUs already.\r\nAlso, how are you launching your training script after?", "Getting the same error when i remove the `to` method. Traceback given below. Also, I launch the training script using `CUDA_VISIBLE_DEVICES=0,1,2,3 python training.py`\r\n\r\n```\r\nTraceback (most recent call last):\r\n File \"/root/Custom-LLM/new_train.py\", line 91, in <module>\r\n main()\r\n File \"/root/Custom-LLM/new_train.py\", line 88, in main\r\n trainer.train()\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/trainer.py\", line 1645, in train\r\n return inner_training_loop(\r\n ^^^^^^^^^^^^^^^^^^^^\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/trainer.py\", line 1938, in _inner_training_loop\r\n tr_loss_step = self.training_step(model, inputs)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/trainer.py\", line 2759, in training_step\r\n loss = self.compute_loss(model, inputs)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/trainer.py\", line 2784, in compute_loss\r\n outputs = model(**inputs)\r\n ^^^^^^^^^^^^^^^\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/accelerate/hooks.py\", line 165, in new_forward\r\n output = old_forward(*args, **kwargs)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/root/.cache/huggingface/modules/transformers_modules/instructcodet5p-16b/modeling_codet5p.py\", line 932, in forward\r\n loss = loss_fct(logits.reshape(-1, self.decoder.config.vocab_size), labels.view(-1))\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/torch/nn/modules/loss.py\", line 1174, in forward\r\n return F.cross_entropy(input, target, weight=self.weight,\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/torch/nn/functional.py\", line 3029, in cross_entropy\r\n return torch._C._nn.cross_entropy_loss(input, target, weight, _Reduction.get_enum(reduction), ignore_index, label_smoothing)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\nRuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:3 and cuda:0! (when checking argument for argument target in method wrapper_CUDA_nll_loss_forward)\r\n 0%| | 0/17409 [00:02<?, ?it/s]\r\n```", "Ah, the problem lies in the custom code of this model. You need to move the `labels` to the device of the logits [here](https://huggingface.co/Salesforce/instructcodet5p-16b/blob/main/modeling_codet5p.py#L930) by adding `labels = labels.to(logits.device)`. Your logits are on the last GPU but your labels are still on the first one.", "I made a [PR](https://huggingface.co/Salesforce/instructcodet5p-16b/discussions/4) with this suggestion on the repo. You can check it out locally by adding the `revision=\"pr_4\"` argument when loading the model.", "Thanks a lot! That solved the issue", "I have a similar error but I could not figure out the reason.\r\nI am using a pre-trained ESM-V2 models from huggingface using QLoRa technique.\r\nHere is my encoder:\r\n\r\n```python\r\n def __init__(self):\r\n # QLoRa fine-tuning:\r\n quantization_config = BitsAndBytesConfig(\r\n load_in_4bit=True,\r\n bnb_4bit_quant_type=\"nf4\",\r\n bnb_4bit_use_double_quant=True,\r\n bnb_4bit_compute_dtype=torch.float16\r\n )\r\n self.model = EsmModel.from_pretrained(model_name, quantization_config=quantization_config)\r\n self.model = prepare_model_for_kbit_training(self.model,\r\n use_gradient_checkpointing=False)\r\n\r\n config = LoraConfig(\r\n r=8,\r\n lora_alpha=32,\r\n target_modules=[\r\n \"query\", \"key\", \"value\",\r\n \"dense\"\r\n ],\r\n lora_dropout=0.05,\r\n bias=\"none\",\r\n )\r\n self.model = get_peft_model(self.model, config)\r\n .\r\n .\r\n .\r\n\r\n def forward(self, x):\r\n x_sequence = {key: value for key, value in x[\"sequence\"].items()}\r\n features = self.model(**x_sequence)\r\n .\r\n .\r\n .\r\n```\r\n\r\nHere is my decoder forward code:\r\n```python\r\n def forward(self, encoder_out, target_input):\r\n tgt_mask, tgt_padding_mask = create_mask(target_input, self.pad_idx, self.device)\r\n tgt_embedding = self.embedding(target_input)\r\n tgt_embedding = self.decoder_pos_drop(tgt_embedding + self.decoder_pos_embed)\r\n\r\n encoder_out = self.encoder_pos_drop(encoder_out + self.encoder_pos_embed)\r\n\r\n encoder_out = encoder_out.transpose(0, 1)\r\n tgt_embedding = tgt_embedding.transpose(0, 1)\r\n\r\n preds = self.decoder(memory=encoder_out,\r\n tgt=tgt_embedding,\r\n tgt_mask=tgt_mask,\r\n tgt_key_padding_mask=tgt_padding_mask)\r\n preds = preds.transpose(0, 1)\r\n return self.output(preds)\r\n```\r\n\r\nThis is my training loop:\r\n\r\n```python\r\n accelerator = Accelerator(\r\n mixed_precision='fp16',\r\n gradient_accumulation_steps=8\r\n )\r\n\r\n net, optimizer, dataloaders_dict[\"train\"], scheduler = accelerator.prepare(\r\n net, optimizer, dataloaders_dict[\"train\"], scheduler\r\n )\r\n\r\n for i, data in enumerate(tools['train_loader']):\r\n with accelerator.accumulate(tools['net']):\r\n embeddings, task_num, sequence, target = data\r\n\r\n target_input = target[:, :-1]\r\n target_expected = target[:, 1:]\r\n\r\n batch = {\"sequence\": sequence, \"embedding\": embeddings, \"target_input\": target_input}\r\n\r\n preds = tools['net'](batch)\r\n loss = tools['loss_function'](preds.reshape(-1, preds.shape[-1]), target_expected.reshape(-1))\r\n loss = torch.mean(loss)\r\n\r\n avg_loss = accelerator.gather(loss.repeat(tools[\"train_batch_size\"])).mean()\r\n train_loss += avg_loss.item() / tools['accum_iter']\r\n\r\n accelerator.backward(loss)\r\n if accelerator.sync_gradients:\r\n accelerator.clip_grad_norm_(tools['net'].parameters(), tools['grad_clip'])\r\n\r\n tools['optimizer'].step()\r\n tools['scheduler'].step()\r\n tools['optimizer'].zero_grad()\r\n```\r\n\r\nI connected an autoregressive decoder to it to create a seq2seq model. My code works pretty well when I use one GPU, but when I set the accelerate config to use two GPUs, I got this error:\r\n \r\n**RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0! (when checking argument for argument index in method wrapper_CUDA__index_select)**\r\n\r\nThis is my error log:\r\n```\r\nFile \"/home/mpngf/projects/JointTraining/train.py\", line 90, in train\r\n preds = tools['net'](batch, mode=0)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/parallel/distributed.py\", line 1156, in forward\r\n output = self._run_ddp_forward(*inputs, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/parallel/distributed.py\", line 1110, in _run_ddp_forward\r\n return module_to_run(*inputs[0], **kwargs[0]) # type: ignore[index]\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/accelerate/utils/operations.py\", line 632, in forward\r\n return model_forward(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/accelerate/utils/operations.py\", line 620, in __call__\r\n return convert_to_fp32(self.model_forward(*args, **kwargs))\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/amp/autocast_mode.py\", line 14, in decorate_autocast\r\n return func(*args, **kwargs)\r\n File \"/home/mpngf/projects/JointTraining/model.py\", line 192, in forward\r\n encoder_out = self.encoder(batch)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n File \"/home/mpngf/projects/JointTraining/model.py\", line 89, in forward\r\n features = self.model(**x[\"sequence\"])\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/peft/peft_model.py\", line 322, in forward\r\n return self.get_base_model()(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/accelerate/hooks.py\", line 165, in new_forward\r\n output = old_forward(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/transformers/models/esm/modeling_esm.py\", line 917, in forward\r\n embedding_output = self.embeddings(\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/accelerate/hooks.py\", line 165, in new_forward\r\n output = old_forward(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/transformers/models/esm/modeling_esm.py\", line 203, in forward\r\n inputs_embeds = self.word_embeddings(input_ids)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/modules/module.py\", line 1501, in _call_impl\r\n return forward_call(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/accelerate/hooks.py\", line 165, in new_forward\r\n output = old_forward(*args, **kwargs)\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/modules/sparse.py\", line 162, in forward\r\n return F.embedding(\r\n File \"/home/mpngf/environments/joint_training/lib/python3.10/site-packages/torch/nn/functional.py\", line 2210, in embedding\r\n return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse)\r\nRuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0! (when checking argument for argument index in method wrapper_CUDA__index_select)\r\n```\r\n\r\nIt seams that it is related to this line:\r\n```python\r\nfeatures = self.model(**x_sequence)\r\n```\r\n\r\n@sgugger, would you be so kind as to assist me with this matter? I would greatly appreciate anyone who can offer their expertise to help me ensure the functionality of my code across multiple GPUs. ", "Hey 🤗 We try to keep the github issues for bugs/feature requests and not for custom code debugging 😅 \r\nCould you ask your question on the [forum](https://discuss.huggingface.co/) instead? I'm sure the community will be of help!\r\n\r\nThanks!", "\r\n[modeling_falcon.zip](https://github.com/huggingface/transformers/files/13403101/modeling_falcon.zip)\r\nI had the same problem by using falcon model for sequence classification with more the one gpu and device auto option.\r\nI added in the modelling_falcon.py on line 1402 and 1507 \"labels = labels.to(logits.device)\"\r\nand from then on it worked. I updated the transformers and accelerater too.\r\nPerhaps this helps to somebody.\r\nRegards\r\nDragan" ]

1,687

1,700

1,688

NONE

null

### System Info - `transformers` version: 4.29.2 - Platform: Linux-5.4.0-137-generic-x86_64-with-glibc2.31 - Python version: 3.11.4 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): 2.0.1 (True) - Tensorflow version (GPU?): not installed (NA) - Flax version (CPU?/GPU?/TPU?): not installed (NA) - Jax version: not installed - JaxLib version: not installed - Using GPU in script?: Yes - Using distributed or parallel set-up in script?: Yes ### Who can help? _No response_ ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction I am trying to finetune an [InstructCodeT5+](https://huggingface.co/Salesforce/instructcodet5p-16b) model on some training data using a multi-GPU setup. The same code (given further below) seems to work in a single-GPU setting (when i set `CUDA_VISIBLE_DEVICES=0`): ``` CUDA_VISIBLE_DEVICES=0,1,2,3 python training.py ===================================BUG REPORT=================================== Welcome to bitsandbytes. For bug reports, please run python -m bitsandbytes and submit this information together with your error trace to: https://github.com/TimDettmers/bitsandbytes/issues ================================================================================ bin /root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/bitsandbytes/libbitsandbytes_cuda118.so CUDA SETUP: CUDA runtime path found: /root/anaconda3/envs/datachat_env/lib/libcudart.so.11.0 CUDA SETUP: Highest compute capability among GPUs detected: 8.0 CUDA SETUP: Detected CUDA version 118 CUDA SETUP: Loading binary /root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/bitsandbytes/libbitsandbytes_cuda118.so... [nltk_data] Downloading package punkt to /root/nltk_data... [nltk_data] Package punkt is already up-to-date! The model weights are not tied. Please use the `tie_weights` method before using the `infer_auto_device` function. Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:24<00:00, 4.92s/it] /root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/optimization.py:407: FutureWarning: This implementation of AdamW is deprecated and will be removed in a future version. Use the PyTorch implementation torch.optim.AdamW instead, or set `no_deprecation_warning=True` to disable this warning warnings.warn( 0%| | 0/17409 [00:00<?, ?it/s]You're using a CodeGenTokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding. Traceback (most recent call last): File "/root/Custom-LLM/training.py", line 364, in <module> main() File "/root/Custom-LLM/training.py", line 336, in main trainer.train() File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/trainer.py", line 1664, in train return inner_training_loop( ^^^^^^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/trainer.py", line 1940, in _inner_training_loop tr_loss_step = self.training_step(model, inputs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/trainer.py", line 2735, in training_step loss = self.compute_loss(model, inputs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/transformers/trainer.py", line 2767, in compute_loss outputs = model(**inputs) ^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward output = old_forward(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/.cache/huggingface/modules/transformers_modules/instructcodet5p-16b/modeling_codet5p.py", line 904, in forward encoder_hidden_states = self.enc_to_dec_proj(encoder_hidden_states) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward output = old_forward(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/datachat_env/lib/python3.11/site-packages/torch/nn/modules/linear.py", line 114, in forward return F.linear(input, self.weight, self.bias) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cuda:3! (when checking argument for argument mat1 in method wrapper_CUDA_addmm) ``` Code for the above error is given below: ``` from transformers import AutoModelForSeq2SeqLM, AutoTokenizer import pandas as pd import os import torch from peft import TaskType from transformers import DataCollatorForSeq2Seq, Seq2SeqTrainer, Seq2SeqTrainingArguments, BitsAndBytesConfig import evaluate import nltk import numpy as np from nltk.tokenize import sent_tokenize from datasets import Dataset, DatasetDict import argparse import pickle import json import statistics import ast from copy import deepcopy device = 'cuda' parser = argparse.ArgumentParser(description='Options') parser.add_argument('--dataset_dir', default='data', type=str, help="folder in which the dataset is stored") parser.add_argument('--output_dir', default="lora-instructcodet5p", type=str, help="output directory for the model") parser.add_argument('--results_dir', default="results", type=str, help="where the results should be stored") args = parser.parse_args() tokenized_dataset = DatasetDict.load_from_disk(args.dataset_dir) pad_tok = 50256 token_id="Salesforce/instructcodet5p-16b" tokenizer = AutoTokenizer.from_pretrained(token_id) def main(): # huggingface hub model id model_id="instructcodet5p-16b" if not os.path.exists(model_id): model_id=token_id # load model from the hub model = AutoModelForSeq2SeqLM.from_pretrained(model_id, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, trust_remote_code=True, decoder_start_token_id=1, pad_token_id=pad_tok, device_map="auto").to(device) # we want to ignore tokenizer pad token in the loss label_pad_token_id = pad_tok # Data collator data_collator = DataCollatorForSeq2Seq( tokenizer, model=model, label_pad_token_id=label_pad_token_id, pad_to_multiple_of=8 ) output_dir=args.output_dir training_args = Seq2SeqTrainingArguments( output_dir=output_dir, per_device_train_batch_size=1, predict_with_generate=True, weight_decay=0.05, warmup_steps=100, fp16=False, learning_rate=1e-3, num_train_epochs=3, logging_dir=f"{output_dir}/logs", logging_strategy="epoch", save_strategy="no", report_to="tensorboard", push_to_hub=False, generation_max_length=200, include_inputs_for_metrics = True, lr_scheduler_type = 'cosine' ) # Create Trainer instance trainer = Seq2SeqTrainer( model=model, args=training_args, data_collator=data_collator, train_dataset=tokenized_dataset["train"] ) # train model trainer.train() if __name__ == '__main__': main() ``` ### Expected behavior Expected behavior is that the model should train in a multi-GPU setting without throwing any errors. The same script works in single-GPU setting but throws the above error in a multi-GPU setting

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24410/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24410/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24409

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24409/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24409/comments

https://api.github.com/repos/huggingface/transformers/issues/24409/events

https://github.com/huggingface/transformers/issues/24409

1,768,337,059

I_kwDOCUB6oc5pZq6j

24,409

wandb metric argument is weird

{ "login": "edmcman", "id": 1017189, "node_id": "MDQ6VXNlcjEwMTcxODk=", "avatar_url": "https://avatars.githubusercontent.com/u/1017189?v=4", "gravatar_id": "", "url": "https://api.github.com/users/edmcman", "html_url": "https://github.com/edmcman", "followers_url": "https://api.github.com/users/edmcman/followers", "following_url": "https://api.github.com/users/edmcman/following{/other_user}", "gists_url": "https://api.github.com/users/edmcman/gists{/gist_id}", "starred_url": "https://api.github.com/users/edmcman/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/edmcman/subscriptions", "organizations_url": "https://api.github.com/users/edmcman/orgs", "repos_url": "https://api.github.com/users/edmcman/repos", "events_url": "https://api.github.com/users/edmcman/events{/privacy}", "received_events_url": "https://api.github.com/users/edmcman/received_events", "type": "User", "site_admin": false }

[]

open

false

null

[]

[ "Thanks @edmcman for the ping.\r\n\r\n@ayulockin @morganmcg1 in case this is relevant to the integrations team. tbh, I don't remember much about this integration but I implemented it so happy to help in case there's a blocker.", "Thanks for the detailed issue @edmcman; we are taking a look at this. Will connect with you @AyushExel in case of blocker.", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "Ping\r\n\r\nOn Mon, Jul 31, 2023, 11:03 AM github-actions[bot] -\r\n***@***.*** <github.edmcman.99c9f1b9d0.notifications#\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1658554442>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZPH3FQDLIUC7DSYYRLXS7CKXANCNFSM6AAAAAAZPGOQMA>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "Ping\r\n\r\nOn Fri, Aug 25, 2023, 4:03 AM github-actions[bot] - ***@***.***\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1692940561>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZPWTBMM362D7HGYQGTXXBL3JANCNFSM6AAAAAAZPGOQMA>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "Ping\r\n\r\nOn Tue, Sep 19, 2023, 4:03 AM github-actions[bot] - ***@***.***\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1725018773>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZKLSBIWBWDK3UV6WGDX3FGVJANCNFSM6AAAAAAZPGOQMA>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "Hey @edmcman, apologies for the delay. I am taking a look at this. Will let you know soon.", "No worries, I was just keeping the issue open :)", "Boop\r\n\r\nOn Sat, Oct 14, 2023, 4:06 AM github-actions[bot] - ***@***.***\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1762708379>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZNRWB7DQT2CAGH3BZ3X7JBZVANCNFSM6AAAAAAZPGOQMA>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "Boop\r\n\r\nOn Fri, Nov 10, 2023, 3:07 AM github-actions[bot] - ***@***.***\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1805270715>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZLZLJX6TNBZJ2LOJG3YDXOFZAVCNFSM6AAAAAAZPGOQMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMBVGI3TANZRGU>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "Boop\r\n\r\nOn Tue, Dec 5, 2023, 3:06 AM github-actions[bot] - ***@***.***\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1840213033>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZJANX7W6GOKOVQQOWLYH3IZZAVCNFSM6AAAAAAZPGOQMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNBQGIYTGMBTGM>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "Boop\r\n\r\nOn Sat, Dec 30, 2023, 3:06 AM github-actions[bot] - ***@***.***\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1872479380>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZMEV7QFVIKKLL7VA3TYL7DP7AVCNFSM6AAAAAAZPGOQMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNZSGQ3TSMZYGA>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "Bump\r\n\r\nOn Wed, Jan 24, 2024, 3:07 AM github-actions[bot] - ***@***.***\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1907600067>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZJIJOVQKD7V5UOB2GDYQC6NPAVCNFSM6AAAAAAZPGOQMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBXGYYDAMBWG4>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "Boop\r\n\r\nOn Sun, Feb 18, 2024, 3:07 AM github-actions[bot] - ***@***.***\r\n***@***.***> wrote:\r\n\r\n> This issue has been automatically marked as stale because it has not had\r\n> recent activity. If you think this still needs to be addressed please\r\n> comment on this thread.\r\n>\r\n> Please note that issues that do not follow the contributing guidelines\r\n> <https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md>\r\n> are likely to be ignored.\r\n>\r\n> —\r\n> Reply to this email directly, view it on GitHub\r\n> <https://github.com/huggingface/transformers/issues/24409#issuecomment-1950997682>,\r\n> or unsubscribe\r\n> <https://github.com/notifications/unsubscribe-auth/AAHYKZKVITBHJR65I2F5KTLYUGZCFAVCNFSM6AAAAAAZPGOQMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJQHE4TONRYGI>\r\n> .\r\n> You are receiving this because you were mentioned.Message ID:\r\n> ***@***.***>\r\n>\r\n" ]

1,687

1,708

null

NONE

null

### System Info - `transformers` version: 4.30.2 - Platform: Linux-5.7.19-050719-generic-x86_64-with-glibc2.29 - Python version: 3.8.10 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): 2.0.1+cu117 (True) - Tensorflow version (GPU?): not installed (NA) - Flax version (CPU?/GPU?/TPU?): not installed (NA) - Jax version: not installed - JaxLib version: not installed - Using GPU in script?: <yes> - Using distributed or parallel set-up in script?: <no> ### Who can help? @AyushExel ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction There are a few bugs/weirdness when specifying the `metric` in `wandb` hyperparameter search. 1. If a `metric` is specified via the `hp_space` argument but there is not a `metric` argument to `hyperparameter_search`, the metric in the `hp_space` is ignored and changed to `eval/loss`. See: https://github.com/huggingface/transformers/blame/6ce6d62b6f20040129ec9831e7c4f6576402ea42/src/transformers/integrations.py#L497 2. If a custom `hp_space` is provided that does not define `metric` at all, and the `metric` argument is specified to `hyperparameter_search`, this code throws an exception: https://github.com/huggingface/transformers/blame/6ce6d62b6f20040129ec9831e7c4f6576402ea42/src/transformers/integrations.py#L500 ### Expected behavior 1. If a `hp_space` defines a metric, use that instead of overwriting it. 2. Don't throw an exception is the `hp_space` lacks a `metric` key.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24409/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24409/timeline

null

https://api.github.com/repos/huggingface/transformers/issues/24408

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24408/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24408/comments

https://api.github.com/repos/huggingface/transformers/issues/24408/events

https://github.com/huggingface/transformers/pull/24408

1,768,040,493

PR_kwDOCUB6oc5Tkicz

24,408

[WIP] Add Restormer

{ "login": "susnato", "id": 56069179, "node_id": "MDQ6VXNlcjU2MDY5MTc5", "avatar_url": "https://avatars.githubusercontent.com/u/56069179?v=4", "gravatar_id": "", "url": "https://api.github.com/users/susnato", "html_url": "https://github.com/susnato", "followers_url": "https://api.github.com/users/susnato/followers", "following_url": "https://api.github.com/users/susnato/following{/other_user}", "gists_url": "https://api.github.com/users/susnato/gists{/gist_id}", "starred_url": "https://api.github.com/users/susnato/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/susnato/subscriptions", "organizations_url": "https://api.github.com/users/susnato/orgs", "repos_url": "https://api.github.com/users/susnato/repos", "events_url": "https://api.github.com/users/susnato/events{/privacy}", "received_events_url": "https://api.github.com/users/susnato/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @amyeroberts for information.", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

CONTRIBUTOR

null

# What does this PR do?   Adds Restormer to HF and closes #22372 ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [x] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [x] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24408/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24408/timeline

null

true

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24408", "html_url": "https://github.com/huggingface/transformers/pull/24408", "diff_url": "https://github.com/huggingface/transformers/pull/24408.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24408.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24407

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24407/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24407/comments

https://api.github.com/repos/huggingface/transformers/issues/24407/events

https://github.com/huggingface/transformers/pull/24407

1,767,996,485

PR_kwDOCUB6oc5TkaOv

24,407

🚨🚨 Fix group beam search

{ "login": "hukuda222", "id": 21185928, "node_id": "MDQ6VXNlcjIxMTg1OTI4", "avatar_url": "https://avatars.githubusercontent.com/u/21185928?v=4", "gravatar_id": "", "url": "https://api.github.com/users/hukuda222", "html_url": "https://github.com/hukuda222", "followers_url": "https://api.github.com/users/hukuda222/followers", "following_url": "https://api.github.com/users/hukuda222/following{/other_user}", "gists_url": "https://api.github.com/users/hukuda222/gists{/gist_id}", "starred_url": "https://api.github.com/users/hukuda222/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/hukuda222/subscriptions", "organizations_url": "https://api.github.com/users/hukuda222/orgs", "repos_url": "https://api.github.com/users/hukuda222/repos", "events_url": "https://api.github.com/users/hukuda222/events{/privacy}", "received_events_url": "https://api.github.com/users/hukuda222/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "## Summary of the problem and corresponding fix (for the core maintainer and our future selves)\r\n\r\n### Problem\r\nThe generation loop in `group_beam_search` is correct, and it builds `num_beam_groups` distinct groups of sequences. However, the [`beam_scorer.finalize()` step](https://github.com/huggingface/transformers/blob/8e164c5400b7b413c7b8fb32e35132001effc970/src/transformers/generation/utils.py#L3711) was not taking `num_beam_groups` into consideration and the beam selection therein, when appending the last tokens, was free to write across groups. This should not happen at all, and it could entirely flush out the diversity in the different groups (when `num_beam_groups >= num_beams/2`), as we see in the example in the PR header.\r\n\r\n### Fix\r\nTwo different paths were possible: a) add logic to `finalize` to handle groups correctly; b) treat each group as an independent set of hypotheses. From the [paper](https://arxiv.org/pdf/1610.02424.pdf), we can read \"we divide the beam budget B into G groups and greedily optimize each group using beam search\", so option b), kindly implemented by @hukuda222, is closer to the reference. ", "@gante \r\nThanks for the review, CI now passes, and I confirmed that `RUN_SLOW=1 py.test tests/generation/test_utils.py -vv` also passes.", "@sgugger [this comment](https://github.com/huggingface/transformers/pull/24407#issuecomment-1605624032) summarizes the problem and the fix", "@sgugger the breaking changes here in the generated outputs from `group_beam_search`, which are inevitable due to the bug fix. The method was underperforming (measured in log scores AND beam diversity, which is the point of the method) before these changes.\r\n\r\nSince it is a bug fix, there is no need to ensure retro compatibility, correct?" ]

1,687

1,688

1,687

CONTRIBUTOR

null

# What does this PR do? Diverse beam search is a method that generates `num_beams//num_beam_groups` sentences for each group independently. However, the current code uses one BeamHypotheses shared by all groups. Therefore, group A will generate two sentences before group B outputs a sentence. So, I created BeamHypotheses for each group so that inferences can be made independently. Changes are as follows. inference code: ```python from transformers import AutoTokenizer, AutoModelForSeq2SeqLM model = AutoModelForSeq2SeqLM.from_pretrained("facebook/bart-large-xsum") tokenizer = AutoTokenizer.from_pretrained("facebook/bart-large-xsum") text = "The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration." outputs = model.generate( tokenizer.encode(text, return_tensors="pt", max_length=512), num_beam_groups=2, num_beams=2, diversity_penalty=1000000.0, num_return_sequences=2, ) print("\n".join(tokenizer.batch_decode(outputs, skip_special_tokens=True))) ``` before: ``` A number Of research projects have investigated the role of the brain's encoder and decoder in the control of the encoded sequences. A number Of research projects have investigated the role of the brain's encoder and decoder in the control of the encoded sequences.. ``` after: ``` The study of the activity of the brain's encoders and decoders has revealed a range of different models of how the brain processes information. A number Of research projects have investigated the role of the brain's encoder and decoder in the control of the encoded sequences. ```   Fixes #24369 ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [x] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [x] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? @gante

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24407/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24407/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24407", "html_url": "https://github.com/huggingface/transformers/pull/24407", "diff_url": "https://github.com/huggingface/transformers/pull/24407.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24407.patch", "merged_at": 1687858991000 }

https://api.github.com/repos/huggingface/transformers/issues/24406

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24406/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24406/comments

https://api.github.com/repos/huggingface/transformers/issues/24406/events

https://github.com/huggingface/transformers/issues/24406

1,767,926,667

I_kwDOCUB6oc5pYGuL

24,406

Potential Memory Leakage during inference using DistilBert/Bert

{ "login": "TOP-RX", "id": 103393767, "node_id": "U_kgDOBimp5w", "avatar_url": "https://avatars.githubusercontent.com/u/103393767?v=4", "gravatar_id": "", "url": "https://api.github.com/users/TOP-RX", "html_url": "https://github.com/TOP-RX", "followers_url": "https://api.github.com/users/TOP-RX/followers", "following_url": "https://api.github.com/users/TOP-RX/following{/other_user}", "gists_url": "https://api.github.com/users/TOP-RX/gists{/gist_id}", "starred_url": "https://api.github.com/users/TOP-RX/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/TOP-RX/subscriptions", "organizations_url": "https://api.github.com/users/TOP-RX/orgs", "repos_url": "https://api.github.com/users/TOP-RX/repos", "events_url": "https://api.github.com/users/TOP-RX/events{/privacy}", "received_events_url": "https://api.github.com/users/TOP-RX/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @ArthurZucker and @younesbelkada ", "Hi @TOP-RX \r\nCan you try to add the `torch.cuda_empty_cache()` before `\"final\"`? Also you might need to combine it with \r\n```python\r\nimport gc\r\n\r\ngc.collect()\r\n```\r\nAfter that call. See this comment from a previous issue for reference: https://github.com/huggingface/transformers/issues/21094#issuecomment-1396951333", "Hello @sgugger @younesbelkada @ArthurZucker,\r\n\r\nThanks for your reply! I tried to include the `torch.cuda.empty_cache()` before \"final\" and also include `gc.collect()` whether before or after `torch.cuda.empty_cache()`, the issue still happened. And I also use `print(torch.cuda.memory_allocated(\"cuda:4\")/ 1024 / 1024)` to check the allocated memory:\r\n\r\n```\r\n @torch.no_grad()\r\n def inference(self, hidden_state, mask, id):\r\n self.eval()\r\n print(\"begin_max\", torch.cuda.max_memory_allocated(\"cuda:4\")/ 1024 / 1024)\r\n print(\"begin\", torch.cuda.memory_allocated(\"cuda:4\")/ 1024 / 1024)\r\n distilbert_output = self.bert(inputs_embeds=hidden_state, attention_mask=mask, return_dict=False)\r\n print(\"middle_max\", torch.cuda.max_memory_allocated(\"cuda:4\")/ 1024 / 1024)\r\n print(\"middle\", torch.cuda.memory_allocated(\"cuda:4\")/ 1024 / 1024)\r\n\r\n gc.collect()\r\n \r\n torch.cuda.empty_cache()\r\n\r\n gc.collect()\r\n\r\n hidden_state = distilbert_output[0] \r\n pooled_output = hidden_state[:, 0] \r\n x = pooled_output\r\n x = F.dropout(x, p=args.dropout, training=self.training)\r\n\r\n del hidden_state, pooled_output\r\n \r\n print(\"final_max\", torch.cuda.max_memory_allocated(\"cuda:4\")/ 1024 / 1024)\r\n print(\"final\", torch.cuda.memory_allocated(\"cuda:4\")/ 1024 / 1024)\r\n for i, lin in enumerate(self.lins[:-1]):\r\n x = lin(x)\r\n #x = self.bns[i](x)\r\n x = F.relu(x)\r\n x = F.dropout(x, p=args.dropout, training=self.training)\r\n x = self.lins[-1](x)\r\n self.z_mlp = self.z_mlp.to(device)\r\n self.z_mlp[id] = x.clone().detach()\r\n print(\"final2\", torch.cuda.max_memory_allocated(\"cuda:4\")/ 1024 / 1024)\r\n torch.cuda.empty_cache()\r\n return x\r\n```\r\nHere are the results:\r\n\r\nbegin_max 4217.28662109375\r\nbegin 4217.28662109375\r\nmiddle_max 39844.28662109375\r\nmiddle 7967.28662109375\r\nfinal_max 39844.28662109375\r\nfinal 7996.58349609375\r\n\r\nthere is also a big gap between the `max_memory_allocated` and `memory_allocated`, could I have some further advices? Thanks.", "I have tired several different ways which I found, but the problem is still existing, is this normal?", "So one thing to note, you are using `gc.collect();torch.cuda.empty_cache()` but you the `del hidden_state, pooled_output` is after. You should first delete, then call `gc.collect();torch.cuda.empty_cache()`. ", "Note: regarding ` max_memory_allocated`\r\n\r\n> By default, this returns the peak allocated memory since the beginning of this program. [reset_peak_memory_stats()](https://pytorch.org/docs/stable/generated/torch.cuda.reset_peak_memory_stats.html#torch.cuda.reset_peak_memory_stats) can be used to reset the starting point in tracking this metric.\r\n\r\nI don't see a problem with those larger values for `max_memory_allocated`. They are just the peak values.", "Hello @ydshieh @ArthurZucker,\r\n\r\nThanks for your help! Before I call the function as shown above, I use:\r\n\r\n```\r\nnum_layers = 1 # Specify the number of layers to remove\r\nencoder_layers = model.bert.transformer.layer[-num_layers:]\r\nmodel.bert.transformer.layer = nn.ModuleList(encoder_layers)\r\n```\r\n\r\nto control the number of transformer layers in my model. However, if I include `@torch.no_grad()` in the function I showed above, here are the results:\r\n\r\n> 1 transformer layer: allocated memory: 800MB, max allocated: 2300MB\r\n> 2 transformer layers: allocated memory: 840MB, max allocated: 2560MB\r\n\r\nif I just comment out the `@torch.no_grad()` in the above code to do a comparison:\r\n\r\n> 1 transformer layer: allocated memory: 4107MB, max allocated: 4299MB\r\n> 2 transformer layers: allocated memory: 7564MB, max allocated: 7756MB\r\n\r\nFor the first case with `@torch.no_grad()`, we don't need to store the intermediate value for backward, it's reasonable the GPU memory is less than the second case. In the second case, the GPU usage is proportional to the number of layers I used which is consistent with my intuition. What makes me confused is no matter how many transformer layers I used in the first case(with `@torch.no_grad()`), the GPU memory usage is almost same. I am wondering if I misunderstand something?\r\n\r\nAny help would be appreciated!", "Hi @TOP-RX \r\n\r\nIt's necessary to provide a self-complete code snippet. With only the definition of `def inference` but not the inputs you passed to it, nothing we can help. Also please import everything necessary in the code snippet.\r\n\r\n> What makes me confused is no matter how many transformer layers I used in the first case(with @torch.no_grad()),\r\n\r\nCould you let us know where do you put the print statements in the function `inference` to get:\r\n\r\n> transformer layer: allocated memory\r\n> transformer layers: allocated memory", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

### System Info transformer: 4.24.0 python: 3.8.13 ### Who can help? _No response_ ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction Hello, I am potentially facing the memory leakage problem when using DistilBert or Bert to do inference, the following is my code: ``` @torch.no_grad() def inference(self, hidden_state, mask, id): self.eval() print("begin", torch.cuda.max_memory_allocated("cuda:4")/ 1024 / 1024) distilbert_output = self.bert(inputs_embeds=hidden_state, attention_mask=mask, return_dict=False) print("middle", torch.cuda.max_memory_allocated("cuda:4")/ 1024 / 1024) hidden_state = distilbert_output[0] pooled_output = hidden_state[:, 0] x = pooled_output x = F.dropout(x, p=args.dropout, training=self.training) del hidden_state, pooled_output print("final", torch.cuda.max_memory_allocated("cuda:4")/ 1024 / 1024) for i, lin in enumerate(self.lins[:-1]): x = lin(x) x = F.relu(x) x = F.dropout(x, p=args.dropout, training=self.training) x = self.lins[-1](x) print("final2", torch.cuda.max_memory_allocated("cuda:4")/ 1024 / 1024) torch.cuda.empty_cache() return x ``` And the result of each printing of memory usage is: begin : 4200 middle : 38000 final : 38000 final 2 : 38000 It seems `distilbert_output = self.bert(inputs_embeds=hidden_state, attention_mask=mask, return_dict=False)` will not release the memory as MLP does, there is a huge increase between "begin" and "middle", but no increase between "final" and "final 2". Could I have some ideas about this issue? Thanks. ### Expected behavior GPU should release memory after transformer inference

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24406/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24406/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24405

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24405/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24405/comments

https://api.github.com/repos/huggingface/transformers/issues/24405/events

https://github.com/huggingface/transformers/pull/24405

1,767,811,650

PR_kwDOCUB6oc5TjyiF

24,405

Fix accumulation by epoch with Accelerate

{ "login": "muellerzr", "id": 7831895, "node_id": "MDQ6VXNlcjc4MzE4OTU=", "avatar_url": "https://avatars.githubusercontent.com/u/7831895?v=4", "gravatar_id": "", "url": "https://api.github.com/users/muellerzr", "html_url": "https://github.com/muellerzr", "followers_url": "https://api.github.com/users/muellerzr/followers", "following_url": "https://api.github.com/users/muellerzr/following{/other_user}", "gists_url": "https://api.github.com/users/muellerzr/gists{/gist_id}", "starred_url": "https://api.github.com/users/muellerzr/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/muellerzr/subscriptions", "organizations_url": "https://api.github.com/users/muellerzr/orgs", "repos_url": "https://api.github.com/users/muellerzr/repos", "events_url": "https://api.github.com/users/muellerzr/events{/privacy}", "received_events_url": "https://api.github.com/users/muellerzr/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Solved the problem for me with small naming fix in trainer.py:\r\n```python\r\n accumulation_plugin = GradientAccumulationPlugin(\r\n num_steps=self.args.gradient_accumulation_steps, sync_with_dataloader=False\r\n )\r\n```\r\n(field name is num_steps not gradient_accumulation_steps)", "We can close this one, as I have added a missing edge case to #24415 " ]

1,687

CONTRIBUTOR

null

# What does this PR do? Should solve accumulating via epoch when using Accelerate (seen in https://github.com/huggingface/transformers/issues/23935#issuecomment-1588134562). Requires https://github.com/huggingface/accelerate/pull/1624 Fixes # (issue) ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. @pacman100

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24405/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24405/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24405", "html_url": "https://github.com/huggingface/transformers/pull/24405", "diff_url": "https://github.com/huggingface/transformers/pull/24405.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24405.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24404

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24404/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24404/comments

https://api.github.com/repos/huggingface/transformers/issues/24404/events

https://github.com/huggingface/transformers/pull/24404

1,767,708,853

PR_kwDOCUB6oc5Tjbys

24,404

TF safetensors reduced mem usage

{ "login": "Rocketknight1", "id": 12866554, "node_id": "MDQ6VXNlcjEyODY2NTU0", "avatar_url": "https://avatars.githubusercontent.com/u/12866554?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Rocketknight1", "html_url": "https://github.com/Rocketknight1", "followers_url": "https://api.github.com/users/Rocketknight1/followers", "following_url": "https://api.github.com/users/Rocketknight1/following{/other_user}", "gists_url": "https://api.github.com/users/Rocketknight1/gists{/gist_id}", "starred_url": "https://api.github.com/users/Rocketknight1/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Rocketknight1/subscriptions", "organizations_url": "https://api.github.com/users/Rocketknight1/orgs", "repos_url": "https://api.github.com/users/Rocketknight1/repos", "events_url": "https://api.github.com/users/Rocketknight1/events{/privacy}", "received_events_url": "https://api.github.com/users/Rocketknight1/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "This looks good in my testing! cc @Narsil for review and @amyeroberts for core maintainer review to save Sylvain from having to read code while he's distracted by the hallucinatory dental anaesthetic goblins \r\n\r\nTo explain one detail: In general, all TF ops can take NumPy inputs. Often the \"idiomatic\" way to write TF code is to do preprocessing/once-off stuff in NumPy and just pass the result to TF. Since Safetensors have very efficient Numpy loading, I generally open them in `\"np\"` format and let TF handle any necessary conversions.\r\n\r\nThe one exception is when loading a PyTorch state dict archive. The reason here is that PyTorch weights often need to be transposed to load them into TF, and NumPy transposes on CPU are much slower than TF transposes on GPU. Model loading was several seconds slower when I passed `\"np\"` format archives to the PyTorch state dict crossloading function, but with GPU transposes this PR has almost no impact on performance, while hugely reducing peak memory usage.", "Yeah, actually, I thought about it a little more - using `np` is idiomatic for TF, but it's a bit messy if the PyTorch crossloading uses `tf` for fast GPU transposes and the others use `np`, especially when it doesn't really matter. I'll use `tf` globally for consistency!", "Just ran the slow tests for a few models that load safetensor checkpoints and this still seems good", "Finished slow testing - performance and memory usage look good!", "Everything looks good at this point - any objections to merging?" ]

1,687

MEMBER

null

When we load safetensors files in TF, the entire safetensors state dict is materialized on GPU alongside the randomly initialized weights. This inflates our memory usage during loading a lot, up to about 2X - 2.5X the amount we actually need. This PR grabs tensors iteratively from the underlying safetensors archives and assigns them. ~It's working for TF-formatted safetensors archives right now, and I'll add torch-format support next.~ Now supports PT and TF formatted archives! Load times still seem very quick for me locally, so I don't think this negatively impacts anything! Fixes #24393

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24404/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24404/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24404", "html_url": "https://github.com/huggingface/transformers/pull/24404", "diff_url": "https://github.com/huggingface/transformers/pull/24404.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24404.patch", "merged_at": 1687439176000 }

https://api.github.com/repos/huggingface/transformers/issues/24403

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24403/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24403/comments

https://api.github.com/repos/huggingface/transformers/issues/24403/events

https://github.com/huggingface/transformers/pull/24403

1,767,681,243

PR_kwDOCUB6oc5TjVrq

24,403

Update activations.py with nn.GELU

{ "login": "nikitakapitan", "id": 101126304, "node_id": "U_kgDOBgcQoA", "avatar_url": "https://avatars.githubusercontent.com/u/101126304?v=4", "gravatar_id": "", "url": "https://api.github.com/users/nikitakapitan", "html_url": "https://github.com/nikitakapitan", "followers_url": "https://api.github.com/users/nikitakapitan/followers", "following_url": "https://api.github.com/users/nikitakapitan/following{/other_user}", "gists_url": "https://api.github.com/users/nikitakapitan/gists{/gist_id}", "starred_url": "https://api.github.com/users/nikitakapitan/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/nikitakapitan/subscriptions", "organizations_url": "https://api.github.com/users/nikitakapitan/orgs", "repos_url": "https://api.github.com/users/nikitakapitan/repos", "events_url": "https://api.github.com/users/nikitakapitan/events{/privacy}", "received_events_url": "https://api.github.com/users/nikitakapitan/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

use nn.GELU # What does this PR do?   Fixes # (issue) ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24403/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24403/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24403", "html_url": "https://github.com/huggingface/transformers/pull/24403", "diff_url": "https://github.com/huggingface/transformers/pull/24403.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24403.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24402

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24402/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24402/comments

https://api.github.com/repos/huggingface/transformers/issues/24402/events

https://github.com/huggingface/transformers/pull/24402

1,767,664,827

PR_kwDOCUB6oc5TjSEt

24,402

Clean up dist import

{ "login": "muellerzr", "id": 7831895, "node_id": "MDQ6VXNlcjc4MzE4OTU=", "avatar_url": "https://avatars.githubusercontent.com/u/7831895?v=4", "gravatar_id": "", "url": "https://api.github.com/users/muellerzr", "html_url": "https://github.com/muellerzr", "followers_url": "https://api.github.com/users/muellerzr/followers", "following_url": "https://api.github.com/users/muellerzr/following{/other_user}", "gists_url": "https://api.github.com/users/muellerzr/gists{/gist_id}", "starred_url": "https://api.github.com/users/muellerzr/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/muellerzr/subscriptions", "organizations_url": "https://api.github.com/users/muellerzr/orgs", "repos_url": "https://api.github.com/users/muellerzr/repos", "events_url": "https://api.github.com/users/muellerzr/events{/privacy}", "received_events_url": "https://api.github.com/users/muellerzr/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

CONTRIBUTOR

null

# What does this PR do? Cleans up the `torch.distributed.X` imports in `training_args` to use the already imported `dist` module, which helps clean up our logic and cases quite a bit. Fixes # (issue) ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. @amyeroberts (cc @sgugger )

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24402/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24402/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24402", "html_url": "https://github.com/huggingface/transformers/pull/24402", "diff_url": "https://github.com/huggingface/transformers/pull/24402.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24402.patch", "merged_at": 1687360783000 }

https://api.github.com/repos/huggingface/transformers/issues/24401

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24401/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24401/comments

https://api.github.com/repos/huggingface/transformers/issues/24401/events

https://github.com/huggingface/transformers/pull/24401

1,767,648,328

PR_kwDOCUB6oc5TjOcK

24,401

Remove redundant code from TrainingArgs

{ "login": "muellerzr", "id": 7831895, "node_id": "MDQ6VXNlcjc4MzE4OTU=", "avatar_url": "https://avatars.githubusercontent.com/u/7831895?v=4", "gravatar_id": "", "url": "https://api.github.com/users/muellerzr", "html_url": "https://github.com/muellerzr", "followers_url": "https://api.github.com/users/muellerzr/followers", "following_url": "https://api.github.com/users/muellerzr/following{/other_user}", "gists_url": "https://api.github.com/users/muellerzr/gists{/gist_id}", "starred_url": "https://api.github.com/users/muellerzr/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/muellerzr/subscriptions", "organizations_url": "https://api.github.com/users/muellerzr/orgs", "repos_url": "https://api.github.com/users/muellerzr/repos", "events_url": "https://api.github.com/users/muellerzr/events{/privacy}", "received_events_url": "https://api.github.com/users/muellerzr/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

CONTRIBUTOR

null

# What does this PR do? Removes some more redundant code that Accelerate can handle directly. Namely: - World size - Process index - `main_process_first` Fixes # (issue) ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. @sgugger @pacman100

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24401/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24401/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24401", "html_url": "https://github.com/huggingface/transformers/pull/24401", "diff_url": "https://github.com/huggingface/transformers/pull/24401.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24401.patch", "merged_at": 1687362687000 }

https://api.github.com/repos/huggingface/transformers/issues/24400

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24400/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24400/comments

https://api.github.com/repos/huggingface/transformers/issues/24400/events

https://github.com/huggingface/transformers/pull/24400

1,767,442,576

PR_kwDOCUB6oc5TihLE

24,400

Check auto mappings could be imported via `from transformers`

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? As shown in #24364, we easily forget to add model mappings like `TF_MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING` to some `__init__` files. Let's add a check to avoid and could detect such issues as early as possible. Along this new check, also add some missing mappings to `__init__` files.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24400/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24400/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24400", "html_url": "https://github.com/huggingface/transformers/pull/24400", "diff_url": "https://github.com/huggingface/transformers/pull/24400.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24400.patch", "merged_at": 1687361518000 }

https://api.github.com/repos/huggingface/transformers/issues/24399

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24399/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24399/comments

https://api.github.com/repos/huggingface/transformers/issues/24399/events

https://github.com/huggingface/transformers/pull/24399

1,767,146,412

PR_kwDOCUB6oc5Thgfo

24,399

byebye Hub connection timeout - Recast

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? It's a bit hard to break up with timeout failure, but @Wauplin was working on it to have 60 seconds instead https://github.com/huggingface/huggingface_hub/pull/1523 We need to change the commit hash to that one in our CircleCI job though.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24399/reactions", "total_count": 1, "+1": 1, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24399/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24399", "html_url": "https://github.com/huggingface/transformers/pull/24399", "diff_url": "https://github.com/huggingface/transformers/pull/24399.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24399.patch", "merged_at": 1687343794000 }

https://api.github.com/repos/huggingface/transformers/issues/24398

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24398/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24398/comments

https://api.github.com/repos/huggingface/transformers/issues/24398/events

https://github.com/huggingface/transformers/pull/24398

1,767,105,979

PR_kwDOCUB6oc5ThXf1

24,398

feat: add support for protobuf 4

{ "login": "jose-turintech", "id": 93319775, "node_id": "U_kgDOBY_yXw", "avatar_url": "https://avatars.githubusercontent.com/u/93319775?v=4", "gravatar_id": "", "url": "https://api.github.com/users/jose-turintech", "html_url": "https://github.com/jose-turintech", "followers_url": "https://api.github.com/users/jose-turintech/followers", "following_url": "https://api.github.com/users/jose-turintech/following{/other_user}", "gists_url": "https://api.github.com/users/jose-turintech/gists{/gist_id}", "starred_url": "https://api.github.com/users/jose-turintech/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/jose-turintech/subscriptions", "organizations_url": "https://api.github.com/users/jose-turintech/orgs", "repos_url": "https://api.github.com/users/jose-turintech/repos", "events_url": "https://api.github.com/users/jose-turintech/events{/privacy}", "received_events_url": "https://api.github.com/users/jose-turintech/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "As you can see from all the red crosses above, this sadly requires more work than just unpinning protobuf.", "@sgugger yes, thanks for taking the time to pointing that out. Currently trying to identify the scope of needed changes and viability. If I can't work on providing right/complete support I'll close the PR." ]

1,687

CONTRIBUTOR

null

# What does this PR do? add support for protobuf 4

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24398/reactions", "total_count": 3, "+1": 3, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24398/timeline

null

true

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24398", "html_url": "https://github.com/huggingface/transformers/pull/24398", "diff_url": "https://github.com/huggingface/transformers/pull/24398.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24398.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24397

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24397/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24397/comments

https://api.github.com/repos/huggingface/transformers/issues/24397/events

https://github.com/huggingface/transformers/pull/24397

1,767,096,326

PR_kwDOCUB6oc5ThVay

24,397

Add `ffmpeg` for `doc_test_job` on CircleCI

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? Need this at least for `docs/source/en/task_summary.md`. Otherwise, [job](https://app.circleci.com/pipelines/github/huggingface/transformers/66845/workflows/ae6bcd25-5071-4f48-a9ba-d446ae6e060f/jobs/833148) fails

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24397/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24397/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24397", "html_url": "https://github.com/huggingface/transformers/pull/24397", "diff_url": "https://github.com/huggingface/transformers/pull/24397.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24397.patch", "merged_at": 1687338759000 }

https://api.github.com/repos/huggingface/transformers/issues/24396

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24396/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24396/comments

https://api.github.com/repos/huggingface/transformers/issues/24396/events

https://github.com/huggingface/transformers/pull/24396

1,766,976,918

PR_kwDOCUB6oc5Tg7HR

24,396

[`pipeline`] Fix str device issue

{ "login": "younesbelkada", "id": 49240599, "node_id": "MDQ6VXNlcjQ5MjQwNTk5", "avatar_url": "https://avatars.githubusercontent.com/u/49240599?v=4", "gravatar_id": "", "url": "https://api.github.com/users/younesbelkada", "html_url": "https://github.com/younesbelkada", "followers_url": "https://api.github.com/users/younesbelkada/followers", "following_url": "https://api.github.com/users/younesbelkada/following{/other_user}", "gists_url": "https://api.github.com/users/younesbelkada/gists{/gist_id}", "starred_url": "https://api.github.com/users/younesbelkada/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/younesbelkada/subscriptions", "organizations_url": "https://api.github.com/users/younesbelkada/orgs", "repos_url": "https://api.github.com/users/younesbelkada/repos", "events_url": "https://api.github.com/users/younesbelkada/events{/privacy}", "received_events_url": "https://api.github.com/users/younesbelkada/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Also \r\n\r\n```\r\npython -c 'from transformers import pipeline; pipe = pipeline(model=\"gpt2\", device=\"cuda\")'\r\n```\r\n\r\nWorks on `main`.. So I'm not sure what's the issue\r\n", "@Narsil what you shared works on main but it should throw an error if you try to run an example with it (I attached a reproducible snippet above)\r\n\r\nAlternatively, this fails on main and this PR fixes it\r\n\r\n```bash\r\npython -c 'from transformers import pipeline; pipe = pipeline(model=\"gpt2\", device=\"cuda\"); pipe(\"hello\")'\r\n```", "Can we remove the `set_device` instead then ? Seems better:\r\n\r\n```patch\r\ndiff --git a/src/transformers/pipelines/base.py b/src/transformers/pipelines/base.py\r\nindex 510c07cf5..b5975d081 100644\r\n--- a/src/transformers/pipelines/base.py\r\n+++ b/src/transformers/pipelines/base.py\r\n@@ -901,10 +901,8 @@ class Pipeline(_ScikitCompat):\r\n with tf.device(\"/CPU:0\" if self.device == -1 else f\"/device:GPU:{self.device}\"):\r\n yield\r\n else:\r\n- if self.device.type == \"cuda\":\r\n- torch.cuda.set_device(self.device)\r\n-\r\n- yield\r\n+ with torch.cuda.device(self.device):\r\n+ yield\r\n```", "The initial thing fails indeed, and seems to be linked to the fact that there are multiple `set_device` happening causing issues.\r\n\r\nBy removing it the issue is indeed removed (but the test you added in the test suite isn't failing on main, and since this is what supposed to catch the regression, this is what I tried :) )\r\n", "I am happy to revert some of the changes I proposed and add yours, it looks much better. However I have few questions \r\n1- is it ok to call that context manager if `self.device` is CPU? I think we need a check on top of that to make sure we're not on CPU (similarly as what we had before)\r\n\r\n```python\r\nimport torch\r\ndevice = torch.device(\"cpu\")\r\n\r\nwith torch.cuda.device(device):\r\n print(torch.randn(1))\r\n```\r\nThrows:\r\n```bash\r\n raise ValueError('Expected a cuda device, but got: {}'.format(device))\r\nValueError: Expected a cuda device, but got: cpu\r\n```\r\nEDIT: just `with torch.device(self.device)` seems to work\r\n\r\n2- I am not sure but I think the `with device` context manager is only available since PT2.0 no?\r\n", "> 2- I am not sure but I think the with device context manager is only available since PT2.0 no?\r\n\r\nI don't know, all those are very good questions for which I don't have the answer to. I just know that now `set_device` is strongly discouraged so it's probably the source of our issues.", "Thanks ! \r\nI can confirm the context manager doesn't work for PT==1.9 which is [should be supported by us](https://github.com/huggingface/transformers/blob/4c6e42958951ca66a6b498b1afce8d8ad4ac2274/setup.py#L178): \r\n\r\n```python\r\nTraceback (most recent call last):\r\n File \"scratch.py\", line 203, in <module>\r\n with torch.device(device):\r\nAttributeError: __enter__\r\n```\r\n\r\nTherefore I just added some changes to ensure backward compatibility with older PT versions. WDYT?", "Hi @Narsil \r\nLet me know if the changes look all good to you, happy to address any additional comments you have ", "May I attempt a different thing ?\r\n\r\nI think the fix is correct, but I'm wondering if simply relying on `torch.cuda.device` context manager couldn't help remove the need for the compat layer.", "Sure yes! ", "Cannot push\r\n\r\n```patch\r\ndiff --git a/src/transformers/pipelines/base.py b/src/transformers/pipelines/base.py\r\nindex 626d33a3d..ee117e62a 100644\r\n--- a/src/transformers/pipelines/base.py\r\n+++ b/src/transformers/pipelines/base.py\r\n@@ -50,7 +50,6 @@ if is_torch_available():\r\n from torch.utils.data import DataLoader, Dataset\r\n\r\n from ..models.auto.modeling_auto import AutoModel\r\n- from ..pytorch_utils import is_torch_greater_or_equal_than_2_0\r\n\r\n # Re-export for backward compatibility\r\n from .pt_utils import KeyDataset\r\n@@ -794,16 +793,11 @@ class Pipeline(_ScikitCompat):\r\n if isinstance(device, torch.device):\r\n self.device = device\r\n elif isinstance(device, str):\r\n- if device == \"cuda\" and not is_torch_greater_or_equal_than_2_0:\r\n- # for backward compatiblity if using `set_device` and `cuda`\r\n- device = f\"cuda:{torch.cuda.current_device()}\"\r\n self.device = torch.device(device)\r\n elif device < 0:\r\n self.device = torch.device(\"cpu\")\r\n- elif isinstance(device, int):\r\n- self.device = torch.device(f\"cuda:{device}\")\r\n else:\r\n- raise ValueError(f\"Device type not supported. Got {device}\")\r\n+ self.device = torch.device(f\"cuda:{device}\")\r\n else:\r\n self.device = device if device is not None else -1\r\n self.torch_dtype = torch_dtype\r\n@@ -908,13 +902,10 @@ class Pipeline(_ScikitCompat):\r\n with tf.device(\"/CPU:0\" if self.device == -1 else f\"/device:GPU:{self.device}\"):\r\n yield\r\n else:\r\n- if is_torch_greater_or_equal_than_2_0:\r\n- with torch.device(self.device):\r\n+ if self.device.type == \"cuda\":\r\n+ with torch.cuda.device(self.device):\r\n yield\r\n- # for backward compatibility\r\n else:\r\n- if self.device.type == \"cuda\":\r\n- torch.cuda.set_device(self.device)\r\n yield\r\n```", "`torch.cuda.device` is defined for torch==1.9 so it should work.\r\n\r\nAnd `torch.device(\"cpu\")` ... well it's the default there's no need to context manage it.", "Hi @Narsil \r\nI am not sure if `with torch.cuda.device(self.device):` is supported for torch<2.0 \r\n\r\nhttps://pytorch.org/tutorials/recipes/recipes/changing_default_device.html\r\n\r\nMaybe we should merge this PR for now to unblock also @thomasw21 & @NouamaneTazi . what do you think?", "I don't think we're blocked by this. \r\n\r\n> And torch.device(\"cpu\") ... well it's the default there's no need to context manage it.\r\n\r\nNot sure of the context of this sentence, but we're overriding the default to `cuda`, so having a context manager to switch back to `cpu` makes sense to me.", "\r\nhttps://pytorch.org/docs/1.9.0/generated/torch.cuda.device.html?highlight=torch%20cuda%20device#torch.cuda.device\r\n\r\nIt is supported from 1.9.0+, at least in the docs.", "Great ! agreed with those changes " ]

1,687

CONTRIBUTOR

null

# What does this PR do? Addresses: https://github.com/huggingface/transformers/pull/24140#issuecomment-1584617146 Currently passing `device="cuda"` is not supported when creating a pipeline. This is because `torch.cuda.set_device(self.device)` expects the device to have an explicit index. The fix is to create an indexed device when initializing a pipeline with a str device Handy reproducible snippet: ```python from transformers import pipeline # this works pipe = pipeline("text-generation", device=0) pipe("Hello") # this works pipe = pipeline("text-generation", device="cuda:0") pipe("Hello") # this fails pipe = pipeline("text-generation", device="cuda") pipe("Hello") ``` cc @amyeroberts @Narsil

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24396/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24396/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24396", "html_url": "https://github.com/huggingface/transformers/pull/24396", "diff_url": "https://github.com/huggingface/transformers/pull/24396.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24396.patch", "merged_at": 1687780716000 }

https://api.github.com/repos/huggingface/transformers/issues/24395

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24395/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24395/comments

https://api.github.com/repos/huggingface/transformers/issues/24395/events

https://github.com/huggingface/transformers/issues/24395

1,766,907,749

I_kwDOCUB6oc5pUN9l

24,395

load_in_4bit seems dosen't work as experted which actually increase GPU memory usage when using zero3 via accelerate

{ "login": "tqjack", "id": 38412243, "node_id": "MDQ6VXNlcjM4NDEyMjQz", "avatar_url": "https://avatars.githubusercontent.com/u/38412243?v=4", "gravatar_id": "", "url": "https://api.github.com/users/tqjack", "html_url": "https://github.com/tqjack", "followers_url": "https://api.github.com/users/tqjack/followers", "following_url": "https://api.github.com/users/tqjack/following{/other_user}", "gists_url": "https://api.github.com/users/tqjack/gists{/gist_id}", "starred_url": "https://api.github.com/users/tqjack/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/tqjack/subscriptions", "organizations_url": "https://api.github.com/users/tqjack/orgs", "repos_url": "https://api.github.com/users/tqjack/repos", "events_url": "https://api.github.com/users/tqjack/events{/privacy}", "received_events_url": "https://api.github.com/users/tqjack/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hello, QLoRA and DeepSpeed can't be used together. DeepSpeed doesn't work with quantized parameters.", "> Hello, QLoRA and DeepSpeed can't be used together. DeepSpeed doesn't work with quantized parameters.\r\n\r\nis that means I can't do zero optimization while using qlora, at least for now? ddp is the only parallelism method compatible with qlora?", "Hello, yes", "cc @younesbelkada for adding more context just in case", "Hi there!\r\nIndeed 4bit + 8bit is not supported with DS, let's maybe add that check on accelerate.preprare (I can work on that)", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

### System Info transformers==4.31.0 deepspeed==0.9.2 peft==0.4.0 bitsandbytes==0.39.0 torch==1.13.0 CUDA Version 11.8 GPUS 8x A100 80gb ### Who can help? @pacman100 ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction I compared the from_pretrained method w/o load_in_4bit passed in and I found out that after passing load_in_4bit in from_pretrained method I can't load model with same hardware and same accelerate config. accelerate config: ds_zero3.yaml ``` compute_environment: LOCAL_MACHINE deepspeed_config: gradient_accumulation_steps: 1 gradient_clipping: 1.0 offload_optimizer_device: 'none' offload_param_device: 'cpu' zero3_init_flag: true zero3_save_16bit_model: true zero_stage: 3 distributed_type: DEEPSPEED downcast_bf16: 'no' machine_rank: 0 main_training_function: main mixed_precision: bf16 num_machines: 1 num_processes: 8 rdzv_backend: static same_network: true tpu_env: [] tpu_use_cluster: false tpu_use_sudo: false use_cpu: false ``` load in normal way: test.py ```import torch from accelerate import Accelerator from transformers import ( AutoModelForCausalLM, ) def main(): accelerator = Accelerator() model_name_or_path = "local path of sambanovasystems/BLOOMChat-176B-v1" model = AutoModelForCausalLM.from_pretrained( model_name_or_path, trust_remote_code=True) if __name__ == "__main__": main() ``` run command: ``` accelerate launch --config_file ds_zero3.yaml test.py ``` it works just fine, GPU memory usage: 22GB x 8 peak CPU memory usage: ~500GB then, I try to use load_in_4bit=True test_4bit.py ```import torch from accelerate import Accelerator from transformers import ( AutoModelForCausalLM, ) def main(): accelerator = Accelerator() model_name_or_path = "local path of sambanovasystems/BLOOMChat-176B-v1" model = AutoModelForCausalLM.from_pretrained( model_name_or_path, trust_remote_code=True, load_in_4bit=True) if __name__ == "__main__": main() ``` run command: ``` accelerate launch --config_file ds_zero3.yaml test_4bit.py ``` OOM error, it seems nether zero3 nor parameters offlaod working as experted. peak cpu usage ~500GB in this case. also tried ``` quantization_config = BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_quant_type="nf4", bnb_4bit_use_double_quant=True, bnb_4bit_compute_dtype=torch.bfloat16 ) ``` OOM ### Expected behavior I am trying to finetune bloomchat176b using qlora and qlora should using less hardware resource. Thanks for help

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24395/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24395/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24394

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24394/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24394/comments

https://api.github.com/repos/huggingface/transformers/issues/24394/events

https://github.com/huggingface/transformers/pull/24394

1,766,661,521

PR_kwDOCUB6oc5Tf-of

24,394

[WIP] Add SPTSv2

{ "login": "JesseSilverberg", "id": 35343284, "node_id": "MDQ6VXNlcjM1MzQzMjg0", "avatar_url": "https://avatars.githubusercontent.com/u/35343284?v=4", "gravatar_id": "", "url": "https://api.github.com/users/JesseSilverberg", "html_url": "https://github.com/JesseSilverberg", "followers_url": "https://api.github.com/users/JesseSilverberg/followers", "following_url": "https://api.github.com/users/JesseSilverberg/following{/other_user}", "gists_url": "https://api.github.com/users/JesseSilverberg/gists{/gist_id}", "starred_url": "https://api.github.com/users/JesseSilverberg/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/JesseSilverberg/subscriptions", "organizations_url": "https://api.github.com/users/JesseSilverberg/orgs", "repos_url": "https://api.github.com/users/JesseSilverberg/repos", "events_url": "https://api.github.com/users/JesseSilverberg/events{/privacy}", "received_events_url": "https://api.github.com/users/JesseSilverberg/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @alaradirik for information. Please let us know when your model is ready for review or if you need any help :-)", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

# What does this PR do? This PR adds SPTSv2. Per the docs, I am opening this PR immediately after generating the boilerplate. Fixes #24235

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24394/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24394/timeline

null

true

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24394", "html_url": "https://github.com/huggingface/transformers/pull/24394", "diff_url": "https://github.com/huggingface/transformers/pull/24394.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24394.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24393

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24393/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24393/comments

https://api.github.com/repos/huggingface/transformers/issues/24393/events

https://github.com/huggingface/transformers/issues/24393

1,766,652,619

I_kwDOCUB6oc5pTPrL

24,393

Increased peak memory usage when upgrading to `transformers` v4.30 and inclusion of `safetensors`

{ "login": "mariecwhite", "id": 5143063, "node_id": "MDQ6VXNlcjUxNDMwNjM=", "avatar_url": "https://avatars.githubusercontent.com/u/5143063?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mariecwhite", "html_url": "https://github.com/mariecwhite", "followers_url": "https://api.github.com/users/mariecwhite/followers", "following_url": "https://api.github.com/users/mariecwhite/following{/other_user}", "gists_url": "https://api.github.com/users/mariecwhite/gists{/gist_id}", "starred_url": "https://api.github.com/users/mariecwhite/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/mariecwhite/subscriptions", "organizations_url": "https://api.github.com/users/mariecwhite/orgs", "repos_url": "https://api.github.com/users/mariecwhite/repos", "events_url": "https://api.github.com/users/mariecwhite/events{/privacy}", "received_events_url": "https://api.github.com/users/mariecwhite/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hey @mariecwhite 👋 \r\n\r\nThe script you shared with us is quite large -- any chance you could help us narrow it down? Ideally, we'd start from a short stand-alone script, unless we are unable to reproduce the issue :)\r\n\r\nBTW, my immediate suspicion goes towards the `.from_pretrained()` function, as (de)serialization functions should be the only ones using `safetensors`.", "Investigating and trying to make a minimal reproducer - in preliminary testing I do see some differences and I'd also guess that `safetensors` is the cause.", "Confirmed that the issue is caused by loading `safetensors` weights with `from_pretrained` as @gante suspected. The difference in memory usage appears before training has even begun. It is transient and only occurs during weight loading - so unless your GPU goes OOM during weight loading itself, the rest of training will not be affected by this.\r\n\r\nMy guess is that the `safetensors` loading is creating two (or more?) copies of the weights on the GPU during the loading process before eventually cleaning them up. The most likely cause is that tensors in TF are created on-device by default, whereas in torch they are created on CPU and must be moved, so probably some of the code is accidentally creating some temporary variables on GPU? \r\n\r\nTo test memory usage for loading in TF I used the following:\r\n```python\r\nimport tensorflow as tf\r\nfrom transformers import TFAutoModel\r\n\r\nmodel = TFAutoModel.from_pretrained(repo_dir)\r\nprint(tf.config.experimental.get_memory_info(\"GPU:0\"))\r\n```\r\n\r\nIn my testing, peak GPU memory usage when loading `bert-large-cased` was 1.5GB when loading from TF `.h5` weights and 4.1GB when loading from `safetensors`, which matches @mariecwhite's benchmark.\r\n\r\ncc @Narsil ", "Further investigation: I think the cause is in the `safetensors` code [here](https://github.com/huggingface/safetensors/blob/main/bindings/python/py_src/safetensors/tensorflow.py#L130) - `tf.convert_to_tensor()` creates the tensor on the GPU by default if one is present, so the entire state dict is materialized on the GPU alongside the randomly initialized weights during loading.", "Hi @mariecwhite, thanks again for the bug report! This is a significant issue and we really appreciate the warning. The PR to fix it is open at #24404 and will hopefully be merged soon. If you'd like to try using the PR branch before then, you can install it with\r\n```python\r\npip install git+https://github.com/huggingface/transformers.git@tf_safetensors_reduced_mem_usage\r\n```", "Thank you for the quick follow-up!", "No probs - it's our fault for missing this issue! The PR has now been merged, so you can just install from `main` to use it.\r\n```\r\npip install git+https://github.com/huggingface/transformers.git\r\n```\r\nIt'll be included in the next patch or full release of `transformers`, at which point you can go back to just `pip install transformers`. Thanks again for the clear bug report and the work you did tracing memory usage in different scenarios!" ]

1,687

CONTRIBUTOR

null

### System Info - `transformers` version: 4.30.2 - Platform: Linux-5.19.0-1022-gcp-x86_64-with-glibc2.35 - Python version: 3.10.6 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): not installed (NA) - Tensorflow version (GPU?): 2.12.0 (True) - Flax version (CPU?/GPU?/TPU?): not installed (NA) - Jax version: not installed - JaxLib version: not installed - Using GPU in script?: <fill in> - Using distributed or parallel set-up in script?: <fill in> ### Who can help? @gante @Rocketknight1 From June 8, after upgrading to `transformers` version 4.30 which automatically installs `safetensors`, peak memory usage in BertLarge and T5Large (and possibly other models that we have not measured) increased to what appears to be a fixed value for smaller batch sizes. ![image](https://github.com/huggingface/transformers/assets/5143063/90197b6c-51e9-4479-9b74-74833883a484) ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [X] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction Run the commands below on a CUDA enabled Linux machine: ``` # Clone benchmarking repo git clone https://github.com/iree-org/iree-samples.git cd iree-samples/iree-tf/benchmark # Setup output file. OUTPUT_PATH=/tmp/tf.json echo "{\"trigger\": { \"timestamp\": \"$(date +'%s')\" }, \"benchmarks\": []}" > "${OUTPUT_PATH}" # Setup virtual environment. TENSORFLOW_VERSION=2.12.0 VENV_DIR=tf.venv ./setup_venv.sh source tf.venv/bin/activate # Run benchmark. BENCHMARK_ID=47cb0d3a-5eb7-41c7-9d7c-97aae7023ecf-MODEL_BERT_LARGE-fp32-TF-384xi32-batch1 python benchmark_model.py --benchmark_id="${BENCHMARK_ID}" --device=gpu --output_path="${OUTPUT_PATH}" --iterations=5 ``` Benchmark output will show `"device_memory_peak_mb": 4157.236992` Now remove `safetensors` (which was installed with v4.30): ``` pip uninstall safetensors python benchmark_model.py --benchmark_id="${BENCHMARK_ID}" --device=gpu --output_path="${OUTPUT_PATH}" --iterations=5 ``` Benchmark output will show `"device_memory_peak_mb": 1591.090432` ### Expected behavior Device peak memory usage should not have increased by 2.5x.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24393/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24393/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24392

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24392/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24392/comments

https://api.github.com/repos/huggingface/transformers/issues/24392/events

https://github.com/huggingface/transformers/issues/24392

1,766,630,109

I_kwDOCUB6oc5pTKLd

24,392

Allow `TextClassificationPipeline` to handle input longer than `model_max_length` tokens

{ "login": "boyleconnor", "id": 6520892, "node_id": "MDQ6VXNlcjY1MjA4OTI=", "avatar_url": "https://avatars.githubusercontent.com/u/6520892?v=4", "gravatar_id": "", "url": "https://api.github.com/users/boyleconnor", "html_url": "https://github.com/boyleconnor", "followers_url": "https://api.github.com/users/boyleconnor/followers", "following_url": "https://api.github.com/users/boyleconnor/following{/other_user}", "gists_url": "https://api.github.com/users/boyleconnor/gists{/gist_id}", "starred_url": "https://api.github.com/users/boyleconnor/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/boyleconnor/subscriptions", "organizations_url": "https://api.github.com/users/boyleconnor/orgs", "repos_url": "https://api.github.com/users/boyleconnor/repos", "events_url": "https://api.github.com/users/boyleconnor/events{/privacy}", "received_events_url": "https://api.github.com/users/boyleconnor/received_events", "type": "User", "site_admin": false }

[ { "id": 2648621985, "node_id": "MDU6TGFiZWwyNjQ4NjIxOTg1", "url": "https://api.github.com/repos/huggingface/transformers/labels/Feature%20request", "name": "Feature request", "color": "FBCA04", "default": false, "description": "Request for a new feature" } ]

open

false

null

[]

[ "I don't have a lot of time to review this atm.\r\n\r\n@amyeroberts Do you know someone that could ?\r\n\r\nOverall I'm hesitant to think it's a good idea. Maintenance is much higher for those `ChunkPipeline` and splitting a document into bits is relatively easy to do outside of the pipeline.\r\nThe merge strategies are also not entirely obvious to me.\r\nThat being said, it's definitely very convenient if implemented directly in the pipeline.", "@boyleconnor Thanks for opening this feature request and for opening an example PR! As @Narsil mentions, there's a maintenance cost to adding this and the set of people who could review this are all pretty busy. \r\n\r\nWhat I suggest is leaving the PR as an example for anyone who might wish to see how to implement this. If this issue gets a lot of attention (we'll measure with 👍 on the feature description) then we can revisit. \r\n\r\n", "@amyeroberts, is there any particular threshold for when HF would consider adding this feature? As I write this comment, there are 13 👍's on the feature description.\r\n\r\nI should also note that I posted about this feature request on social media, and so most of these 👍's are from NLP professionals & academics whom I personally know.", "WDYT @Rocketknight1 @ArthurZucker?", "I think the solution is valid and the PR is well-implemented, but my intuition is not to add it!\r\n\r\nMy reasoning is:\r\n- Chunking inputs in the pipeline makes the pipeline more awkward, and will make it harder to add other features.\r\n- Models are getting longer and longer context lengths, including the addition of things like alibi/rope position embeddings that let them scale up to arbitrarily long contexts. That means this feature will be less useful over time.\r\n\r\nThus, I feel this feature is a workaround for a very temporary limitation that some models have, but with more modern models the workaround shouldn't be needed, and so we shouldn't hardcode it into our pipelines in 2023. Instead, we should consider replacing some of the default models for tasks like `sentiment-analysis` with newer models that have better performance and that support longer inputs." ]

1,687

1,696

null

CONTRIBUTOR

null

### Problem Running a `TextClassificationPipeline` on a text with more tokens than its model's maximum position embeddings (e.g. 512 for BERT) like so: ```python from transformers import pipeline classifier = pipeline('sentiment-analysis') classifier("Hello, world! " * 1000) ``` will lead to this error: ``` RuntimeError: The size of tensor a (4002) must match the size of tensor b (512) at non-singleton dimension 1 ``` Note: the numbers (`4002` and `512`, above) will vary depending on the max length of the model in use and the length (in tokens) of the text that triggered the error. (_**If you found this issue through web-searching for the above error or some other means, look at the linked PR for an implemented code fix to this problem, and consider giving a thumbs-up to this comment if you think it should be merged into the main codebase**_) ### Feature request We should add "chunking"/"sliding window" functionality to `TextClassificationPipeline`, allowing it to process documents longer than the `model_max_length` of its `.model`. Specifically, this would run an instance of the model on each of several "sliding window" views of each input sequence, then take the mean, similar to (but somewhat simpler than) how [`TokenClassificationPipeline`](https://github.com/huggingface/transformers/blob/ad78d9597b224443e9fe65a94acc8c0bc48cd039/src/transformers/pipelines/token_classification.py#L96) does so in part by subclassing from `ChunkPipeline`. ### Motivation It would be nice to easily do, e.g., sentiment analysis on documents longer than the `model_max_length` of the given model/tokenizer. I have in the past tried to do this in a time-sensitive context and was unable to do so. ### Your contribution I have already opened a draft PR: #24312. I would be happy to finish the missing parts (e.g. documentation) if someone on the Huggingface team (I believe @Narsil is the appropriate person to tag) can confirm that they would accept this feature as I plan to implement it.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24392/reactions", "total_count": 13, "+1": 13, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24392/timeline

null

https://api.github.com/repos/huggingface/transformers/issues/24391

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24391/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24391/comments

https://api.github.com/repos/huggingface/transformers/issues/24391/events

https://github.com/huggingface/transformers/issues/24391

1,766,600,745

I_kwDOCUB6oc5pTDAp

24,391

Bug on Gather all remaining tensors and put them back on the CPU

{ "login": "jinmang2", "id": 37775784, "node_id": "MDQ6VXNlcjM3Nzc1Nzg0", "avatar_url": "https://avatars.githubusercontent.com/u/37775784?v=4", "gravatar_id": "", "url": "https://api.github.com/users/jinmang2", "html_url": "https://github.com/jinmang2", "followers_url": "https://api.github.com/users/jinmang2/followers", "following_url": "https://api.github.com/users/jinmang2/following{/other_user}", "gists_url": "https://api.github.com/users/jinmang2/gists{/gist_id}", "starred_url": "https://api.github.com/users/jinmang2/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/jinmang2/subscriptions", "organizations_url": "https://api.github.com/users/jinmang2/orgs", "repos_url": "https://api.github.com/users/jinmang2/repos", "events_url": "https://api.github.com/users/jinmang2/events{/privacy}", "received_events_url": "https://api.github.com/users/jinmang2/received_events", "type": "User", "site_admin": false }

[]

closed

false

{ "login": "muellerzr", "id": 7831895, "node_id": "MDQ6VXNlcjc4MzE4OTU=", "avatar_url": "https://avatars.githubusercontent.com/u/7831895?v=4", "gravatar_id": "", "url": "https://api.github.com/users/muellerzr", "html_url": "https://github.com/muellerzr", "followers_url": "https://api.github.com/users/muellerzr/followers", "following_url": "https://api.github.com/users/muellerzr/following{/other_user}", "gists_url": "https://api.github.com/users/muellerzr/gists{/gist_id}", "starred_url": "https://api.github.com/users/muellerzr/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/muellerzr/subscriptions", "organizations_url": "https://api.github.com/users/muellerzr/orgs", "repos_url": "https://api.github.com/users/muellerzr/repos", "events_url": "https://api.github.com/users/muellerzr/events{/privacy}", "received_events_url": "https://api.github.com/users/muellerzr/received_events", "type": "User", "site_admin": false }

[ { "login": "muellerzr", "id": 7831895, "node_id": "MDQ6VXNlcjc4MzE4OTU=", "avatar_url": "https://avatars.githubusercontent.com/u/7831895?v=4", "gravatar_id": "", "url": "https://api.github.com/users/muellerzr", "html_url": "https://github.com/muellerzr", "followers_url": "https://api.github.com/users/muellerzr/followers", "following_url": "https://api.github.com/users/muellerzr/following{/other_user}", "gists_url": "https://api.github.com/users/muellerzr/gists{/gist_id}", "starred_url": "https://api.github.com/users/muellerzr/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/muellerzr/subscriptions", "organizations_url": "https://api.github.com/users/muellerzr/orgs", "repos_url": "https://api.github.com/users/muellerzr/repos", "events_url": "https://api.github.com/users/muellerzr/events{/privacy}", "received_events_url": "https://api.github.com/users/muellerzr/received_events", "type": "User", "site_admin": false } ]

[ "@jinmang2 can you provide the example code you ran? Would be an excellent way for us to write a test-case around it :)", "@muellerzr sure!\r\n\r\n### dialogue state tracking results (my own code)\r\n- colab link: https://colab.research.google.com/drive/1afT8O5OrUaTaZ07xvi_3AMW07P2nnRuC?usp=sharing\r\n- code link: https://github.com/jinmang2/KLUE-DST/blob/main/run.py\r\n- results\r\n\r\n```python\r\n...\r\neval_results = trainer.evaluation_loop(\r\n trainer.get_eval_dataloader(),\r\n description=\"Evaluation\",\r\n prediction_loss_only=False,\r\n ignore_keys=None,\r\n metric_key_prefix=\"eval\",\r\n)\r\nlen(trainer.eval_dataset), eval_results.predictions[0].shape\r\n```\r\n```\r\n(7224, (824, 9, 71))\r\n```\r\n- expected shape: `(7224, 9, 71)`\r\n\r\n### glue mnli results (huggingface's example code)\r\n- colab link: https://colab.research.google.com/drive/1Yfoh4-Pl5LqGUWBZZqbc3OGN1R3x3O_w?usp=sharing\r\n- code link: https://github.com/huggingface/transformers/blob/ba695c1efd55091e394eb59c90fb33ac3f9f0d41/examples/pytorch/text-classification/run_glue.py\r\n- results\r\n\r\n```python\r\n...\r\neval_results = trainer.evaluation_loop(\r\n trainer.get_eval_dataloader(),\r\n description=\"Evaluation\",\r\n prediction_loss_only=False,\r\n ignore_keys=None,\r\n metric_key_prefix=\"eval\",\r\n)\r\n# The total number of samples in the eval example is 9815.\r\n# However, it can be seen that only 3 samples of prediction used for evaluation remain.\r\nlen(trainer.eval_dataset), eval_results.predictions[0].shape\r\n```\r\n```\r\n(9815, (3,))\r\n```\r\n- expected shape: `(9815,)`", "Since the `evaluate` method does not know how many eval samples have been evaluated (internally, only necessary values are loaded into the metrics dictionary), the `evaluation_loop` method directly receives `eval_results` and checks the eval samples.", "Thanks @jinmang2, this indeed was a big from the integration and the original logic should have been maintained. A PR will be opened shortly with the solution (and also solves a failing test!) thanks again for your solution and thorough analysis", "Thanks for fixing it! :-)" ]

1,687

1,688

NONE

null

### System Info - `transformers` version: 4.31.0.dev0 - Platform: Linux-5.15.107+-x86_64-with-glibc2.31 - Python version: 3.10.12 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): 2.0.1+cu118 (True) - Tensorflow version (GPU?): 2.12.0 (True) - Flax version (CPU?/GPU?/TPU?): 0.6.9 (gpu) - Jax version: 0.4.10 - JaxLib version: 0.4.10 - Using GPU in script?: <fill in> - Using distributed or parallel set-up in script?: <fill in> ### Who can help? @sgugger @muellerzr @ArthurZucker ### Information - [X] The official example scripts - [x] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction ### Colab Link - https://colab.research.google.com/drive/1afT8O5OrUaTaZ07xvi_3AMW07P2nnRuC?usp=sharing ### Expected behavior ### What is problem? In the trainer's `evaluation_loop`, the huggingface trainer collects all tensors and puts them into the `compute_metrics` method all at once. If the tensor is too large, a CUDA OOM error occurs, so it's prevented by sending it to the cpu in advance (using `nested_numpify` method) in the middle with `eval_accumulation_step`. For samples smaller than `eval_accumulation_step`, metrics were calculated after additional summing using the code below. https://github.com/huggingface/transformers/blob/66fd3a8d626a32989f4569260db32785c6cbf42a/src/transformers/trainer.py#L3304-L3318 However, since the code changes as below in PR #24028, the problem as occurred. https://github.com/huggingface/transformers/blob/ebd94b0f6f215f6bc0f70e61eba075eb9196f9ef/src/transformers/trainer.py#L3184-L3192 The code above doesn't merge the remaining tensors into the final container, it just allocates them. In fact, in the example code I ran, even though `len(eval_dataset)` was `7224`, with `per_device_eval_batch_size=16`, `eval_accumulation_steps=100`, gpu-cpu communication was performed 4 times, and only `824` eval samples remained. Please check it out and hope the PR will be corrected. Thank you!

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24391/reactions", "total_count": 1, "+1": 1, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24391/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24390

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24390/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24390/comments

https://api.github.com/repos/huggingface/transformers/issues/24390/events

https://github.com/huggingface/transformers/issues/24390

1,766,385,065

I_kwDOCUB6oc5pSOWp

24,390

AttributeError: 'AutoformerModel' object has no attribute 'embedder'

{ "login": "pourmatin", "id": 19475339, "node_id": "MDQ6VXNlcjE5NDc1MzM5", "avatar_url": "https://avatars.githubusercontent.com/u/19475339?v=4", "gravatar_id": "", "url": "https://api.github.com/users/pourmatin", "html_url": "https://github.com/pourmatin", "followers_url": "https://api.github.com/users/pourmatin/followers", "following_url": "https://api.github.com/users/pourmatin/following{/other_user}", "gists_url": "https://api.github.com/users/pourmatin/gists{/gist_id}", "starred_url": "https://api.github.com/users/pourmatin/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/pourmatin/subscriptions", "organizations_url": "https://api.github.com/users/pourmatin/orgs", "repos_url": "https://api.github.com/users/pourmatin/repos", "events_url": "https://api.github.com/users/pourmatin/events{/privacy}", "received_events_url": "https://api.github.com/users/pourmatin/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @kashif ", "While waiting our expert @kashif , the code in \r\n\r\n```python\r\n if config.num_static_categorical_features > 0:\r\n self.embedder = AutoformerFeatureEmbedder(\r\n cardinalities=config.cardinality, embedding_dims=config.embedding_dimension\r\n )\r\n```\r\ntogether\r\n```\r\n if static_categorical_features is not None:\r\n embedded_cat = self.embedder(static_categorical_features)\r\n```\r\n\r\nSince you pass `static_categorical_features` to the model's forward, you can check your config's `num_static_categorical_features ` attribute. Probably it is 0 and `self.embedder` is not created. In this case, I think we should not pass `static_categorical_features` to the model.\r\n", "thanks @ydshieh having a look!", "@pourmatin that is correct, if you do not specify any categorical features, then you should not pass the model a list of categorical features... I believe we had a check for this, let me confirm!", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

### System Info Darwin Kernel Version 22.4.0: Mon Mar 6 21:00:41 PST 2023; root:xnu-8796.101.5~3/RELEASE_ARM64_T8103 ### Who can help? @ydshieh ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction 1. declare the model: `config = AutoformerConfig() config.prediction_length = 5 config.context_length = 55 config.lags_sequence = [1,2, 3, 4, 5] model = AutoformerModel(config)` 2. invoke the forward method of the model by calling it: `outputs = model( past_values=batches["past_values"][:16], past_time_features=batches["past_time_features"][:16], past_observed_mask=batches["past_observed_mask"][:16], static_categorical_features=batches["static_categorical_features"][:16], future_values=batches["future_values"][:16], future_time_features=batches["future_time_features"][:16], )` ### Expected behavior I'd expect the model to run forward method successfully. Instead, I get the following error: ``` --------------------------------------------------------------------------- AttributeError Traceback (most recent call last) /var/folders/2n/p53ld4x51l1_yhdkj7q4y3t00000gr/T/ipykernel_49714/3104772270.py in <module> ----> 1 outputs = model( 2 past_values=batches["past_values"][:16], 3 past_time_features=batches["past_time_features"][:16], 4 past_observed_mask=batches["past_observed_mask"][:16], 5 static_categorical_features=batches["static_categorical_features"][:16], ~/opt/anaconda3/lib/python3.9/site-packages/torch/nn/modules/module.py in _call_impl(self, *args, **kwargs) 1499 or _global_backward_pre_hooks or _global_backward_hooks 1500 or _global_forward_hooks or _global_forward_pre_hooks): -> 1501 return forward_call(*args, **kwargs) 1502 # Do not call functions when jit is used 1503 full_backward_hooks, non_full_backward_hooks = [], [] ~/opt/anaconda3/lib/python3.9/site-packages/transformers/models/autoformer/modeling_autoformer.py in forward(self, past_values, past_time_features, past_observed_mask, static_categorical_features, static_real_features, future_values, future_time_features, decoder_attention_mask, head_mask, decoder_head_mask, cross_attn_head_mask, encoder_outputs, past_key_values, output_hidden_states, output_attentions, use_cache, return_dict) 1725 return_dict = return_dict if return_dict is not None else self.config.use_return_dict 1726 -> 1727 transformer_inputs, temporal_features, loc, scale, static_feat = self.create_network_inputs( 1728 past_values=past_values, 1729 past_time_features=past_time_features, ~/opt/anaconda3/lib/python3.9/site-packages/transformers/models/autoformer/modeling_autoformer.py in create_network_inputs(self, past_values, past_time_features, static_categorical_features, static_real_features, past_observed_mask, future_values, future_time_features) 1637 static_feat = torch.cat((static_real_features, static_feat), dim=1) 1638 if static_categorical_features is not None: -> 1639 embedded_cat = self.embedder(static_categorical_features) 1640 static_feat = torch.cat((embedded_cat, static_feat), dim=1) 1641 expanded_static_feat = static_feat.unsqueeze(1).expand(-1, time_feat.shape[1], -1) ~/opt/anaconda3/lib/python3.9/site-packages/torch/nn/modules/module.py in __getattr__(self, name) 1612 if name in modules: 1613 return modules[name] -> 1614 raise AttributeError("'{}' object has no attribute '{}'".format( 1615 type(self).__name__, name)) 1616 AttributeError: 'AutoformerModel' object has no attribute 'embedder' ```

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24390/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24390/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24389

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24389/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24389/comments

https://api.github.com/repos/huggingface/transformers/issues/24389/events

https://github.com/huggingface/transformers/pull/24389

1,766,339,710

PR_kwDOCUB6oc5TfDPZ

24,389

[Trainer] Fix optimizer step on PyTorch TPU

{ "login": "cowanmeg", "id": 6570496, "node_id": "MDQ6VXNlcjY1NzA0OTY=", "avatar_url": "https://avatars.githubusercontent.com/u/6570496?v=4", "gravatar_id": "", "url": "https://api.github.com/users/cowanmeg", "html_url": "https://github.com/cowanmeg", "followers_url": "https://api.github.com/users/cowanmeg/followers", "following_url": "https://api.github.com/users/cowanmeg/following{/other_user}", "gists_url": "https://api.github.com/users/cowanmeg/gists{/gist_id}", "starred_url": "https://api.github.com/users/cowanmeg/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/cowanmeg/subscriptions", "organizations_url": "https://api.github.com/users/cowanmeg/orgs", "repos_url": "https://api.github.com/users/cowanmeg/repos", "events_url": "https://api.github.com/users/cowanmeg/events{/privacy}", "received_events_url": "https://api.github.com/users/cowanmeg/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @muellerzr and @pacman100 ", "_The documentation is not available anymore as the PR was closed or merged._", "Thank you @cowanmeg for the fix!" ]

1,687

CONTRIBUTOR

null

# What does this PR do? Update optimizer step for TPUs to user `self.optimizer.step()` instead of `xm.optimizer_step(self.optimizer)`. AcceleratedOptimizer properly calls `xm.optimizer_step` on the optimizer (https://github.com/huggingface/accelerate/blob/main/src/accelerate/optimizer.py#L129). This fixes a bug in transformers/trainer.py when using Pytorch on TPUs: File "/usr/local/lib/python3.8/dist-packages/torch_xla/core/xla_model.py", line 471, in _fetch_gradients for param_group in optimizer.getstate()['param_groups']: KeyError: 'param_groups' ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. @sgugger

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24389/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24389/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24389", "html_url": "https://github.com/huggingface/transformers/pull/24389", "diff_url": "https://github.com/huggingface/transformers/pull/24389.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24389.patch", "merged_at": 1687346681000 }

https://api.github.com/repos/huggingface/transformers/issues/24388

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24388/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24388/comments

https://api.github.com/repos/huggingface/transformers/issues/24388/events

https://github.com/huggingface/transformers/pull/24388

1,766,329,118

PR_kwDOCUB6oc5TfBzv

24,388

[docs] Fix NLLB-MoE links

{ "login": "stevhliu", "id": 59462357, "node_id": "MDQ6VXNlcjU5NDYyMzU3", "avatar_url": "https://avatars.githubusercontent.com/u/59462357?v=4", "gravatar_id": "", "url": "https://api.github.com/users/stevhliu", "html_url": "https://github.com/stevhliu", "followers_url": "https://api.github.com/users/stevhliu/followers", "following_url": "https://api.github.com/users/stevhliu/following{/other_user}", "gists_url": "https://api.github.com/users/stevhliu/gists{/gist_id}", "starred_url": "https://api.github.com/users/stevhliu/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/stevhliu/subscriptions", "organizations_url": "https://api.github.com/users/stevhliu/orgs", "repos_url": "https://api.github.com/users/stevhliu/repos", "events_url": "https://api.github.com/users/stevhliu/events{/privacy}", "received_events_url": "https://api.github.com/users/stevhliu/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

MEMBER

null

Fixes the broken links raised in #24382.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24388/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24388/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24388", "html_url": "https://github.com/huggingface/transformers/pull/24388", "diff_url": "https://github.com/huggingface/transformers/pull/24388.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24388.patch", "merged_at": 1687307660000 }

https://api.github.com/repos/huggingface/transformers/issues/24387

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24387/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24387/comments

https://api.github.com/repos/huggingface/transformers/issues/24387/events

https://github.com/huggingface/transformers/pull/24387

1,766,255,972

PR_kwDOCUB6oc5Tezp7

24,387

Update deprecated torch.ger

{ "login": "kit1980", "id": 420184, "node_id": "MDQ6VXNlcjQyMDE4NA==", "avatar_url": "https://avatars.githubusercontent.com/u/420184?v=4", "gravatar_id": "", "url": "https://api.github.com/users/kit1980", "html_url": "https://github.com/kit1980", "followers_url": "https://api.github.com/users/kit1980/followers", "following_url": "https://api.github.com/users/kit1980/following{/other_user}", "gists_url": "https://api.github.com/users/kit1980/gists{/gist_id}", "starred_url": "https://api.github.com/users/kit1980/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/kit1980/subscriptions", "organizations_url": "https://api.github.com/users/kit1980/orgs", "repos_url": "https://api.github.com/users/kit1980/repos", "events_url": "https://api.github.com/users/kit1980/events{/privacy}", "received_events_url": "https://api.github.com/users/kit1980/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "@sgugger please take a look.", "_The documentation is not available anymore as the PR was closed or merged._", "A test (I believe unrelated to the change) timed out https://app.circleci.com/pipelines/github/huggingface/transformers/66846/workflows/ae63fb76-7260-490e-8301-7c6cf986e693/jobs/833163", "Yes it's unrelated, merging :-)" ]

1,687

CONTRIBUTOR

null

` torch.ger` was deprecated long time ago and `torch.outer` is a direct replacement: https://pytorch.org/docs/stable/generated/torch.ger.html

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24387/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24387/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24387", "html_url": "https://github.com/huggingface/transformers/pull/24387", "diff_url": "https://github.com/huggingface/transformers/pull/24387.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24387.patch", "merged_at": 1687306873000 }

https://api.github.com/repos/huggingface/transformers/issues/24386

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24386/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24386/comments

https://api.github.com/repos/huggingface/transformers/issues/24386/events

https://github.com/huggingface/transformers/issues/24386

1,766,146,530

I_kwDOCUB6oc5pRUHi

24,386

Loading Trained RAG Model

{ "login": "YichiRockyZhang", "id": 29335344, "node_id": "MDQ6VXNlcjI5MzM1MzQ0", "avatar_url": "https://avatars.githubusercontent.com/u/29335344?v=4", "gravatar_id": "", "url": "https://api.github.com/users/YichiRockyZhang", "html_url": "https://github.com/YichiRockyZhang", "followers_url": "https://api.github.com/users/YichiRockyZhang/followers", "following_url": "https://api.github.com/users/YichiRockyZhang/following{/other_user}", "gists_url": "https://api.github.com/users/YichiRockyZhang/gists{/gist_id}", "starred_url": "https://api.github.com/users/YichiRockyZhang/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/YichiRockyZhang/subscriptions", "organizations_url": "https://api.github.com/users/YichiRockyZhang/orgs", "repos_url": "https://api.github.com/users/YichiRockyZhang/repos", "events_url": "https://api.github.com/users/YichiRockyZhang/events{/privacy}", "received_events_url": "https://api.github.com/users/YichiRockyZhang/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @YichiRockyZhang \r\nThanks for the issue, looking at your environment (transformers == 4.13.0) I would probably give it a try with on of the newest version of transformers. It seems the config didn't saved properly the model identifier for some reason. Would it be possible to use a recent version of the lib for you? ", "Hi @YichiRockyZhang\r\n\r\n\r\nIf @younesbelkada's above suggesion is still not working, it would help a lot if you can provide a short but a bit more complete code example that you:\r\n- **create/load the initialize model(s)**\r\n- **save it to the checkpoint (without the need of training/fine-tuning)**\r\n- _(you already provide this part)_ the way you try to load the saved checkpoint\r\n\r\nThis way, it's easier and fast for us to reproduce and look into the issue. Thank you in advance.", "Hi @younesbelkada . Thanks for the response! This did help as running the finetuning script now results in a more sensible saved checkpoint.\r\n\r\n![image](https://github.com/huggingface/transformers/assets/29335344/06ebb647-8d2d-4522-a3e4-6448d838d6ce)\r\n\r\nI can now load the model with the following:\r\n```py\r\npath = \"/fs/nexus-scratch/yzhang42/rag_end2end/model_checkpoints_MS/checkpoint31\"\r\n\r\nrag_tokenizer = RagTokenizer.from_pretrained(path)\r\nrag_retriever = RagRetriever.from_pretrained(\r\n path,\r\n use_dummy_dataset=False,\r\n indexed_dataset=ds,\r\n index_name=\"compressed\",\r\n)\r\n\r\nrag_model = RagTokenForGeneration.from_pretrained(path, retriever=rag_retriever)\r\n```\r\n\r\nHi @ydshieh ! Unfortunately, I believe my problem is specific to fine-tuning. I'm using the only fine-tuning script for this model that I can find (in huggingface documentation and even on the internet). The script uses pytorch lightning to train and save the model. The below snippet from [`finetune_rag.py`](https://github.com/huggingface/transformers/blob/main/examples/research_projects/rag-end2end-retriever/finetune_rag.py) details how the models is saved.\r\n\r\n```py\r\n @pl.utilities.rank_zero_only\r\n def on_save_checkpoint(self, checkpoint: Dict[str, Any]) -> None:\r\n save_path = self.output_dir.joinpath(\"checkpoint{}\".format(self.step_count))\r\n self.model.config.save_step = self.step_count\r\n # self.model.save_pretrained(save_path)\r\n self.tokenizer.save_pretrained(save_path)\r\n\r\n if self.custom_config.end2end:\r\n modified_state_dict = self.model.state_dict()\r\n for key in self.model.state_dict().keys():\r\n if key.split(\".\")[1] == \"ctx_encoder\":\r\n del modified_state_dict[key]\r\n self.model.save_pretrained(save_directory=save_path, state_dict=modified_state_dict)\r\n\r\n save_path_dpr = os.path.join(self.dpr_ctx_check_dir, \"checkpoint{}\".format(self.step_count))\r\n self.model.rag.ctx_encoder.save_pretrained(save_path_dpr)\r\n self.context_tokenizer.save_pretrained(save_path_dpr)\r\n```\r\n\r\nI understand HF does not maintain these scripts, but for what it's worth, I think retrieval-augmented models are very important and should have a bit more support!", "@YichiRockyZhang \r\n\r\nThanks for sharing more details. What I means is that you can still make a **self-complete** code snippet:\r\n\r\n- how you (or the script create the model)\r\n- then save that model using the logic in the method `on_save_checkpoint` you provided\r\n\r\nYou don't need to go through the training part in the script, just the create/save part. By ` self-complete`, it means we can just run it directly to see the failure you have. Of course y**ou will have to wrap up things in your own way** (not just showing us the definition of `on_save_checkpoint`). I hope this makes my previous comment a bit clear and look forward to see a reproducible code snippet 🤗 ", "@ydshieh Hi, thank you for the quick responses! I've edited my above reply to reflect the fact that upgrading to transformers==4.30.2 seemed to have worked after making sure my data was ASCII encoded. Though it does seem that the fine-tuning script is only saving the whole model after the first epoch. I've adjusted the code to be \r\n\r\n```py\r\n @pl.utilities.rank_zero_only\r\n def on_save_checkpoint(self, checkpoint: Dict[str, Any]) -> None:\r\n save_path = self.output_dir.joinpath(\"checkpoint{}\".format(self.step_count))\r\n self.model.config.save_step = self.step_count\r\n # self.model.save_pretrained(save_path)\r\n self.tokenizer.save_pretrained(save_path)\r\n\r\n if self.custom_config.end2end:\r\n modified_state_dict = self.model.state_dict()\r\n for key in self.model.state_dict().keys():\r\n if key.split(\".\")[1] == \"ctx_encoder\":\r\n del modified_state_dict[key]\r\n self.model.save_pretrained(save_directory=save_path, state_dict=modified_state_dict)\r\n\r\n save_path_dpr = os.path.join(self.dpr_ctx_check_dir, \"checkpoint{}\".format(self.step_count))\r\n self.model.rag.ctx_encoder.save_pretrained(save_path_dpr)\r\n self.context_tokenizer.save_pretrained(save_path_dpr)\r\n else: #NEW\r\n state_dict = self.model.state_dict()\r\n self.model.save_pretrained(save_directory=save_path, state_dict=state_dict)\r\n```\r\n\r\nI will update this thread in the morning once fine-tuning is finished. If my fix doesn't work out, I'll try to put together a more minimal and self-complete script for debugging purposes! 🤗", "Nice and good luck :-) !", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

### System Info Python 3.9.16 Transformers 4.13.0 WSL ### Who can help? @ArthurZucker @younesbelkada @shamanez ### Information - [X] The official example scripts - [ ] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction After finetuning RAG, I'm left with the following directory, and I'm not sure how to load the resulting checkpoint. ![image](https://github.com/huggingface/transformers/assets/29335344/5480c778-7b8e-40e5-a2ed-b543a970e50d) I should note the checkpoint is ~6 GB while the [original huggingface checkpoint](https://huggingface.co/facebook/rag-token-base/tree/main) is 2 GB. I suspect this is because I used the [`finetune_rag_ray_end2end.sh`](https://github.com/huggingface/transformers/tree/main/examples/research_projects/rag-end2end-retriever) script, so it includes all 3 models (reader, retriever, generator). Below are my attempts to load the checkpoint **Attempt 1** ```py ds = load_dataset(path='wiki_dpr', name='psgs_w100.multiset.compressed', split='train') rag_tokenizer = RagTokenizer.from_pretrained("facebook/rag-token-base") rag_retriever = RagRetriever.from_pretrained( "facebook/rag-token-base", use_dummy_dataset=False, indexed_dataset=ds, index_name="embeddings", ) rag_model = RagTokenForGeneration.from_pretrained("facebook/rag-token-base", retriever=rag_retriever) checkpoint_path = "/fs/nexus-scratch/yzhang42/rag_end2end/model_checkpoints_MS/val_avg_em=0.0026-step_count=601.0.ckpt" rag_model.load_state_dict(torch.load(checkpoint_path)) ``` The program runs forever with the following traceback when I interrupt it: ``` Some weights of RagTokenForGeneration were not initialized from the model checkpoint at facebook/rag-token-base and are newly initialized: ['rag.generator.lm_head.weight'] You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference. /fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/ray/_private/services.py:238: UserWarning: Not all Ray Dashboard dependencies were found. To use the dashboard please install Ray using `pip install ray[default]`. To disable this message, set RAY_DISABLE_IMPORT_WARNING env var to '1'. warnings.warn(warning_message) ^CTraceback (most recent call last): File "/nfshomes/yzhang42/rag/notebooks/rag_eval.py", line 37, in <module> rag_model.load_state_dict(torch.load(checkpoint_path)) File "/fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/torch/serialization.py", line 712, in load return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args) File "/fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/torch/serialization.py", line 1049, in _load result = unpickler.load() File "/fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/ray/actor.py", line 1005, in _deserialization_helper return worker.core_worker.deserialize_and_register_actor_handle( File "python/ray/_raylet.pyx", line 1594, in ray._raylet.CoreWorker.deserialize_and_register_actor_handle File "python/ray/_raylet.pyx", line 1563, in ray._raylet.CoreWorker.make_actor_handle File "/fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/ray/_private/function_manager.py", line 402, in load_actor_class actor_class = self._load_actor_class_from_gcs( File "/fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/ray/_private/function_manager.py", line 487, in _load_actor_class_from_gcs time.sleep(0.001) KeyboardInterrupt ``` **Attempt 2** ```py from transformers import AutoConfig, AutoModel, PretrainedConfig, RagTokenizer, RagRetriever, BartForConditionalGeneration, RagTokenForGeneration, RagSequenceForGeneration, RagConfig from transformers import BartModel qe_config = PretrainedConfig( name_or_path=\ "/fs/nexus-scratch/yzhang42/rag_end2end/model_checkpoints_MS/checkpoint601/generator_tokenizer/tokenizer_config.json") gen_config = PretrainedConfig( name_or_path=\ "/fs/nexus-scratch/yzhang42/rag_end2end/model_checkpoints_MS/checkpoint601/question_encoder_tokenizer/tokenizer_config.json") RagConfig.from_question_encoder_generator_configs( question_encoder_config=qe_config, generator_config=gen_config ) ``` Gives the following error: ``` --------------------------------------------------------------------------- ValueError Traceback (most recent call last) Cell In[2], line 11 4 qe_config = PretrainedConfig( 5 name_or_path=\ 6 "/fs/nexus-scratch/yzhang42/rag_end2end/model_checkpoints_MS/checkpoint601/generator_tokenizer/tokenizer_config.json") 7 gen_config = PretrainedConfig( 8 name_or_path=\ 9 "/fs/nexus-scratch/yzhang42/rag_end2end/model_checkpoints_MS/checkpoint601/question_encoder_tokenizer/tokenizer_config.json") ---> 11 RagConfig.from_question_encoder_generator_configs( 12 question_encoder_config=qe_config, 13 generator_config=gen_config 14 ) File /fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/transformers/models/rag/configuration_rag.py:183, in RagConfig.from_question_encoder_generator_configs(cls, question_encoder_config, generator_config, **kwargs) 172 @classmethod 173 def from_question_encoder_generator_configs( 174 cls, question_encoder_config: PretrainedConfig, generator_config: PretrainedConfig, **kwargs 175 ) -> PretrainedConfig: 176 r""" 177 Instantiate a :class:`~transformers.EncoderDecoderConfig` (or a derived class) from a pre-trained encoder model 178 configuration and decoder model configuration. (...) 181 :class:`EncoderDecoderConfig`: An instance of a configuration object 182 """ --> 183 return cls(question_encoder=question_encoder_config.to_dict(), generator=generator_config.to_dict(), **kwargs) File /fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/transformers/models/rag/configuration_rag.py:140, in RagConfig.__init__(self, vocab_size, is_encoder_decoder, prefix, bos_token_id, pad_token_id, eos_token_id, decoder_start_token_id, title_sep, doc_sep, n_docs, max_combined_length, retrieval_vector_size, retrieval_batch_size, dataset, dataset_split, index_name, index_path, passages_path, use_dummy_dataset, reduce_loss, label_smoothing, do_deduplication, exclude_bos_score, do_marginalize, output_retrieved, use_cache, forced_eos_token_id, **kwargs) 136 decoder_model_type = decoder_config.pop("model_type") 138 from ..auto.configuration_auto import AutoConfig --> 140 self.question_encoder = AutoConfig.for_model(question_encoder_model_type, **question_encoder_config) 141 self.generator = AutoConfig.for_model(decoder_model_type, **decoder_config) 143 self.reduce_loss = reduce_loss File /fs/nexus-scratch/yzhang42/miniconda3/envs/qa3/lib/python3.9/site-packages/transformers/models/auto/configuration_auto.py:492, in AutoConfig.for_model(cls, model_type, *args, **kwargs) 490 config_class = CONFIG_MAPPING[model_type] 491 return config_class(*args, **kwargs) --> 492 raise ValueError( 493 f"Unrecognized model identifier: {model_type}. Should contain one of {', '.join(CONFIG_MAPPING.keys())}" 494 ) ValueError: Unrecognized model identifier: . Should contain one of imagegpt, qdqbert, vision-encoder-decoder, trocr, fnet, segformer, vision-text-dual-encoder, perceiver, gptj, layoutlmv2, beit, rembert, visual_bert, canine, roformer, clip, bigbird_pegasus, deit, luke, detr, gpt_neo, big_bird, speech_to_text_2, speech_to_text, vit, wav2vec2, m2m_100, convbert, led, blenderbot-small, retribert, ibert, mt5, t5, mobilebert, distilbert, albert, bert-generation, camembert, xlm-roberta, pegasus, marian, mbart, megatron-bert, mpnet, bart, blenderbot, reformer, longformer, roberta, deberta-v2, deberta, flaubert, fsmt, squeezebert, hubert, bert, openai-gpt, gpt2, transfo-xl, xlnet, xlm-prophetnet, prophetnet, xlm, ctrl, electra, speech-encoder-decoder, encoder-decoder, funnel, lxmert, dpr, layoutlm, rag, tapas, splinter, sew-d, sew, unispeech-sat, unispeech ``` ### Expected behavior I'm not sure what expected behavior is supposed to be.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24386/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24386/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24385

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24385/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24385/comments

https://api.github.com/repos/huggingface/transformers/issues/24385/events

https://github.com/huggingface/transformers/issues/24385

1,766,135,653

I_kwDOCUB6oc5pRRdl

24,385

How to unwrap after auto_wrap in FSDP?

{ "login": "ZN1010", "id": 13196992, "node_id": "MDQ6VXNlcjEzMTk2OTky", "avatar_url": "https://avatars.githubusercontent.com/u/13196992?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ZN1010", "html_url": "https://github.com/ZN1010", "followers_url": "https://api.github.com/users/ZN1010/followers", "following_url": "https://api.github.com/users/ZN1010/following{/other_user}", "gists_url": "https://api.github.com/users/ZN1010/gists{/gist_id}", "starred_url": "https://api.github.com/users/ZN1010/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ZN1010/subscriptions", "organizations_url": "https://api.github.com/users/ZN1010/orgs", "repos_url": "https://api.github.com/users/ZN1010/repos", "events_url": "https://api.github.com/users/ZN1010/events{/privacy}", "received_events_url": "https://api.github.com/users/ZN1010/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "This is more a question for the PyTorch forums as it's purely related to FSDP. Still cc-ing @pacman100 in case he has any idea.", "Hello, as Sylvain mentioned this question is for the PyTorch forums. `summon_full_params` usage can be found in these tests: https://github.com/pytorch/pytorch/blob/main/test/distributed/checkpoint/test_fsdp_optim_state.py#L56-L59\r\n\r\nI am not sure if it contains the information related to the gradients of a given parameter. ", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

I am currently fine-tuning a LLM (LLaMA) and would like to retrieve the gradients of each weight (parameter) after every gradient update. However, I notice that weights are (auto) wrapped into stuff like “_fsdp_wrapped_module._flat_param” during training. I need to map these wrapped weights to the original LLaMA architecture such as “self_attn.v_proj”. Any code examples? I guess “summon_full_params()” might be the function that I look for, but I am not sure if that is correct. I also have difficulty using this function. Thanks a lot for any help!

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24385/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24385/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24384

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24384/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24384/comments

https://api.github.com/repos/huggingface/transformers/issues/24384/events

https://github.com/huggingface/transformers/pull/24384

1,766,129,169

PR_kwDOCUB6oc5TeeWV

24,384

Refactor hyperparameter search backends

{ "login": "alexmojaki", "id": 3627481, "node_id": "MDQ6VXNlcjM2Mjc0ODE=", "avatar_url": "https://avatars.githubusercontent.com/u/3627481?v=4", "gravatar_id": "", "url": "https://api.github.com/users/alexmojaki", "html_url": "https://github.com/alexmojaki", "followers_url": "https://api.github.com/users/alexmojaki/followers", "following_url": "https://api.github.com/users/alexmojaki/following{/other_user}", "gists_url": "https://api.github.com/users/alexmojaki/gists{/gist_id}", "starred_url": "https://api.github.com/users/alexmojaki/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/alexmojaki/subscriptions", "organizations_url": "https://api.github.com/users/alexmojaki/orgs", "repos_url": "https://api.github.com/users/alexmojaki/repos", "events_url": "https://api.github.com/users/alexmojaki/events{/privacy}", "received_events_url": "https://api.github.com/users/alexmojaki/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "> Using abstract classes like this is not really the way the Transformers library is designed\r\n\r\nThat's fine, I was very unsure which approach to take. `abc` offers additional safety and IDE assistance as the standard way to ensure that all abstract methods are implemented, but it's probably overkill here and I also didn't like how heavy it was. I've pushed a much simpler strategy.\r\n\r\n> I recommended to just complete the error message to include wandb.\r\n\r\nThe point of this is that it's difficult to see all the missing bits. The current code isn't just missing wandb in the error message, it's also missing from `default_hp_space` (fixed in this PR) and the docstring (not enforced in this PR, although it could be, I just didn't want to jump there just yet).", "> I also don't see how you would have benefits for IDE as you show in the PR description.\r\n\r\nSorry for the confusion, that's not part of this PR to keep the scope focused, but if this is merged I can follow it up with another which adds constructors to each backend class which accept the precise kwargs that the backend `run` supports.", "Opened https://github.com/huggingface/huggingface_hub/issues/1526 in regards to the unrelated test failure.", "Thanks a lot for your contribution!", "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24384). All of your documentation changes will be reflected on that endpoint." ]

1,687

CONTRIBUTOR

null

Fixes https://github.com/huggingface/transformers/issues/24379 The goal here is to clearly group the essential info/functionality about each backend together to make reading/changing things easier. For example if another backend integration is added it should be less likely for something to be forgotten as apparently happened with wandb. @sgugger sorry I didn't get a full confirmation to go ahead with this, it just seemed easier to show what I meant with code rather than continue explaining in the issue. There's many other ways this could be done and I can change the approach but I hope that the general direction at least is clear from this PR. I also think this would help move towards improving the user facing API since as mentioned in https://github.com/huggingface/transformers/issues/24278#issuecomment-1599189018 (cc @hugocool) the kwargs have no type hints and are not very easy to use. So maybe instead of: ```python best_run = trainer.hyperparameter_search( direction="maximize", backend="ray", # this is just **kwargs, not so clear what's possible... storage_path="...", callbacks=..., ) ``` one could write: ```python best_run = trainer.hyperparameter_search( direction="maximize", backend=RayTuneBackend( # now more assistance is possible storage_path="...", callbacks=..., ), ) ```

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24384/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24384/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24384", "html_url": "https://github.com/huggingface/transformers/pull/24384", "diff_url": "https://github.com/huggingface/transformers/pull/24384.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24384.patch", "merged_at": 1687458505000 }

https://api.github.com/repos/huggingface/transformers/issues/24383

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24383/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24383/comments

https://api.github.com/repos/huggingface/transformers/issues/24383/events

https://github.com/huggingface/transformers/issues/24383

1,766,078,398

I_kwDOCUB6oc5pRDe-

24,383

How BERT 512 limit works?

{ "login": "Bisht9887", "id": 26176160, "node_id": "MDQ6VXNlcjI2MTc2MTYw", "avatar_url": "https://avatars.githubusercontent.com/u/26176160?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Bisht9887", "html_url": "https://github.com/Bisht9887", "followers_url": "https://api.github.com/users/Bisht9887/followers", "following_url": "https://api.github.com/users/Bisht9887/following{/other_user}", "gists_url": "https://api.github.com/users/Bisht9887/gists{/gist_id}", "starred_url": "https://api.github.com/users/Bisht9887/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Bisht9887/subscriptions", "organizations_url": "https://api.github.com/users/Bisht9887/orgs", "repos_url": "https://api.github.com/users/Bisht9887/repos", "events_url": "https://api.github.com/users/Bisht9887/events{/privacy}", "received_events_url": "https://api.github.com/users/Bisht9887/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Please use the [forums](https://discuss.huggingface.co/) for such questions. You did not pass your input the model, just the tokenizer.", "sure! I will take care of that. I am little new to this. Thanks for info" ]

1,687

NONE

null

### System Info I passed a long text of 3000 tokens and it did not give me any error. Does BERT not have 512 token limit? Why it is not giving any error? This is the code I used. You can pass any input with tokens more than 512 ### Who can help? _No response_ ### Information - [X] The official example scripts - [X] My own modified scripts ### Tasks - [X] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction from transformers import BertTokenizer tokenizer = BertTokenizer.from_pretrained('bert-base-uncased') # Long input text with more than 512 tokens text = "This is a very long text with more than 512 tokens..." tokens = tokenizer.tokenize(text) print(len(tokens)) ### Expected behavior error when processing more than 512

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24383/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24383/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24381

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24381/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24381/comments

https://api.github.com/repos/huggingface/transformers/issues/24381/events

https://github.com/huggingface/transformers/pull/24381

1,765,835,534

PR_kwDOCUB6oc5TdgQ1

24,381

fixing layer indexing error when pipeline parallel > 1

{ "login": "xshaun", "id": 8446322, "node_id": "MDQ6VXNlcjg0NDYzMjI=", "avatar_url": "https://avatars.githubusercontent.com/u/8446322?v=4", "gravatar_id": "", "url": "https://api.github.com/users/xshaun", "html_url": "https://github.com/xshaun", "followers_url": "https://api.github.com/users/xshaun/followers", "following_url": "https://api.github.com/users/xshaun/following{/other_user}", "gists_url": "https://api.github.com/users/xshaun/gists{/gist_id}", "starred_url": "https://api.github.com/users/xshaun/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/xshaun/subscriptions", "organizations_url": "https://api.github.com/users/xshaun/orgs", "repos_url": "https://api.github.com/users/xshaun/repos", "events_url": "https://api.github.com/users/xshaun/events{/privacy}", "received_events_url": "https://api.github.com/users/xshaun/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24381). All of your documentation changes will be reflected on that endpoint." ]

1,687

CONTRIBUTOR

null

when applying pipeline parallel, the index of layers converted from transformers to megatron is wrong since there is not offset applied successfully. For example, if having 4 layers, and pipeline parallel is 2, we want to the result looks like `layers.0 + layers.1` and `layers.2 + layers.3`, but now the result is `layers.0 + layers.1` and `layers.0 + layers.1`, since user should use `pp_layer_id` that is calculated with `layer + offset` , instead of `layer` that is only a index of range loop.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24381/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24381/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24381", "html_url": "https://github.com/huggingface/transformers/pull/24381", "diff_url": "https://github.com/huggingface/transformers/pull/24381.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24381.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24380

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24380/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24380/comments

https://api.github.com/repos/huggingface/transformers/issues/24380/events

https://github.com/huggingface/transformers/issues/24380

1,765,798,908

I_kwDOCUB6oc5pP_P8

24,380

WavLM error when running forward

{ "login": "MorenoLaQuatra", "id": 10062811, "node_id": "MDQ6VXNlcjEwMDYyODEx", "avatar_url": "https://avatars.githubusercontent.com/u/10062811?v=4", "gravatar_id": "", "url": "https://api.github.com/users/MorenoLaQuatra", "html_url": "https://github.com/MorenoLaQuatra", "followers_url": "https://api.github.com/users/MorenoLaQuatra/followers", "following_url": "https://api.github.com/users/MorenoLaQuatra/following{/other_user}", "gists_url": "https://api.github.com/users/MorenoLaQuatra/gists{/gist_id}", "starred_url": "https://api.github.com/users/MorenoLaQuatra/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/MorenoLaQuatra/subscriptions", "organizations_url": "https://api.github.com/users/MorenoLaQuatra/orgs", "repos_url": "https://api.github.com/users/MorenoLaQuatra/repos", "events_url": "https://api.github.com/users/MorenoLaQuatra/events{/privacy}", "received_events_url": "https://api.github.com/users/MorenoLaQuatra/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "I think there's a small typo with your codesnippet:\r\n```diff\r\n- input = fe(audio, return_tensor=\"pt\")\r\n+ input = fe(audio, return_tensors=\"pt\")\r\n```\r\n\r\nE.g. running the following works for me:\r\n```python\r\nfrom transformers import AutoModel, AutoFeatureExtractor\r\nimport torch\r\nimport numpy as np\r\n\r\nmodel = AutoModel.from_pretrained(\"microsoft/wavlm-base-plus\")\r\nfe = AutoFeatureExtractor.from_pretrained(\"microsoft/wavlm-base-plus\")\r\n\r\naudio = np.random.randn(16000) # random 1 second input audio\r\n\r\ninput = fe(audio, sampling_rate=16000, return_tensors=\"pt\")\r\n\r\nwith torch.no_grad():\r\n model(input_values=input[\"input_values\"])\r\n```", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "Hey @MorenoLaQuatra - did the above suggestion fix your issue? Feel free to close this thread if so", "Well, actually my problem is more linked with the hidden_states when using output_hidden_states=True (the typo was my fault when reporting the snippet here on GitHub). However, I cannot reproduce it at the moment, so I will close for now.\r\n\r\nThanks @sanchit-gandhi !", "Thanks for clarifying @MorenoLaQuatra! Feel free to open the issue again with a code repro if you find the model isn't working and we can take a deeper dive into it" ]

1,687

1,690

CONTRIBUTOR

null

### System Info **Relevant Libraries** transformers==4.26.1 torchaudio==2.0.2 torch==2.0.1 OS: Ubuntu 20.04 ### Who can help? @sanchit-gandhi ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [X] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction Run the following code: ```python from transformers import AutoModel, AutoFeatureExtractor model = AutoModel.from_pretrained("microsoft/wavlm-base-plus") fe = AutoFeatureExtractor.from_pretrained("microsoft/wavlm-base-plus") audio_path = "..." audio, sr = torchaudio.load(audio_path) input = fe(audio, return_tensor="pt") model(input_values=input["input_values"]) ``` --- When I try running the previous code, I got the following error: ```bash Traceback (most recent call last): File "<stdin>", line 1, in <module> File "PATH_TO_MY_ENV_SITE_PACKAGES/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "PATH_TO_MY_ENV_SITE_PACKAGES/transformers/models/wavlm/modeling_wavlm.py", line 1229, in forward extract_features = self.feature_extractor(input_values) File "PATH_TO_MY_ENV_SITE_PACKAGES/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "PATH_TO_MY_ENV_SITE_PACKAGES/transformers/models/wavlm/modeling_wavlm.py", line 346, in forward hidden_states = input_values[:, None] TypeError: list indices must be integers or slices, not tuple ``` ### Expected behavior get the output with last_hidden_state and others. This is not happening with HuBERT or Wav2Vec2.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24380/reactions", "total_count": 1, "+1": 1, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24380/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24379

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24379/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24379/comments

https://api.github.com/repos/huggingface/transformers/issues/24379/events

https://github.com/huggingface/transformers/issues/24379

1,765,783,928

I_kwDOCUB6oc5pP7l4

24,379

`Trainer.hyperparameter_search` doesn't document `wandb` or offer it as a default backend

{ "login": "alexmojaki", "id": 3627481, "node_id": "MDQ6VXNlcjM2Mjc0ODE=", "avatar_url": "https://avatars.githubusercontent.com/u/3627481?v=4", "gravatar_id": "", "url": "https://api.github.com/users/alexmojaki", "html_url": "https://github.com/alexmojaki", "followers_url": "https://api.github.com/users/alexmojaki/followers", "following_url": "https://api.github.com/users/alexmojaki/following{/other_user}", "gists_url": "https://api.github.com/users/alexmojaki/gists{/gist_id}", "starred_url": "https://api.github.com/users/alexmojaki/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/alexmojaki/subscriptions", "organizations_url": "https://api.github.com/users/alexmojaki/orgs", "repos_url": "https://api.github.com/users/alexmojaki/repos", "events_url": "https://api.github.com/users/alexmojaki/events{/privacy}", "received_events_url": "https://api.github.com/users/alexmojaki/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Those are all integrations maintained by the authors of those libraries, we do not maintain them ourselves. It might be a bug, but it's up to the wandb folks to fix it in this case :-) ", "Even the glue code in `trainer.py` that ties the various backends together?\r\n\r\nWould you accept a PR to refactor this stuff? For example this code:\r\n\r\n```python\r\n if backend is None:\r\n backend = default_hp_search_backend()\r\n if backend is None:\r\n raise RuntimeError(\r\n \"At least one of optuna or ray should be installed. \"\r\n \"To install optuna run `pip install optuna`. \"\r\n \"To install ray run `pip install ray[tune]`. \"\r\n \"To install sigopt run `pip install sigopt`.\"\r\n )\r\n backend = HPSearchBackend(backend)\r\n if backend == HPSearchBackend.OPTUNA and not is_optuna_available():\r\n raise RuntimeError(\"You picked the optuna backend, but it is not installed. Use `pip install optuna`.\")\r\n if backend == HPSearchBackend.RAY and not is_ray_tune_available():\r\n raise RuntimeError(\r\n \"You picked the Ray Tune backend, but it is not installed. Use `pip install 'ray[tune]'`.\"\r\n )\r\n if backend == HPSearchBackend.SIGOPT and not is_sigopt_available():\r\n raise RuntimeError(\"You picked the sigopt backend, but it is not installed. Use `pip install sigopt`.\")\r\n if backend == HPSearchBackend.WANDB and not is_wandb_available():\r\n raise RuntimeError(\"You picked the wandb backend, but it is not installed. Use `pip install wandb`.\")\r\n```\r\n\r\ncontains a lot of repetition that I'd be happy to clean up, and it's easy to see how the wandb integration author missed a place to add a reference to wandb.", "The first bit with the runtime error is fine (though missing wandb). For the rest, it should be done in each integration which normally error very fast if the corresponding lib is not installed." ]

1,687

CONTRIBUTOR

null

`Trainer.hyperparameter_search` seems to prioritise the `optuna/ray/sigopt` backends, while `wandb` almost seems like a second-class citizen in the code. Specifically, the docstring explicitly mentions the first three backends multiple times in different contexts but not `wandb`, and `default_hp_search_backend` won't return `wandb` even if it's available. Is this intentional or accidental?

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24379/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24379/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24382

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24382/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24382/comments

https://api.github.com/repos/huggingface/transformers/issues/24382/events

https://github.com/huggingface/transformers/issues/24382

1,765,922,167

I_kwDOCUB6oc5pQdV3

24,382

Some links in NLLB page are broken.

{ "login": "ranggihwang", "id": 50730045, "node_id": "MDQ6VXNlcjUwNzMwMDQ1", "avatar_url": "https://avatars.githubusercontent.com/u/50730045?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ranggihwang", "html_url": "https://github.com/ranggihwang", "followers_url": "https://api.github.com/users/ranggihwang/followers", "following_url": "https://api.github.com/users/ranggihwang/following{/other_user}", "gists_url": "https://api.github.com/users/ranggihwang/gists{/gist_id}", "starred_url": "https://api.github.com/users/ranggihwang/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ranggihwang/subscriptions", "organizations_url": "https://api.github.com/users/ranggihwang/orgs", "repos_url": "https://api.github.com/users/ranggihwang/repos", "events_url": "https://api.github.com/users/ranggihwang/events{/privacy}", "received_events_url": "https://api.github.com/users/ranggihwang/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @stevhliu ", "Doc in version 4.30.0 still has problems, not only the doc section but other links :( @stevhliu ", "The fix is on main, so it will only be reflected on the main version of the documentation. And if you have found other links to fix, please do tell us or open directly PRs to fix them :-)" ]

1,687

NONE

null

**Is your feature request related to a problem? Please describe.** A clear and concise description of what the problem is. Ex. I'm always frustrated when [...] **Describe the solution you'd like** A clear and concise description of what you want to happen. **Describe alternatives you've considered** A clear and concise description of any alternative solutions or features you've considered. **Additional context** Add any other context or screenshots about the feature request here. Hi, this is just a simple notification for crashed links. https://huggingface.co/docs/transformers/v4.30.0/model_doc/nllb-moe On this page, two links in the "documentation resources" are broken. Thank you.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24382/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24382/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24378

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24378/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24378/comments

https://api.github.com/repos/huggingface/transformers/issues/24378/events

https://github.com/huggingface/transformers/pull/24378

1,765,743,865

PR_kwDOCUB6oc5TdNIy

24,378

Skip a tapas (tokenization) test in past CI

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24378). All of your documentation changes will be reflected on that endpoint." ]

1,687

COLLABORATOR

null

# What does this PR do? Same as in #24251 where 1 test (from the tokenization test file) is missed.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24378/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24378/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24378", "html_url": "https://github.com/huggingface/transformers/pull/24378", "diff_url": "https://github.com/huggingface/transformers/pull/24378.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24378.patch", "merged_at": 1687278945000 }

https://api.github.com/repos/huggingface/transformers/issues/24377

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24377/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24377/comments

https://api.github.com/repos/huggingface/transformers/issues/24377/events

https://github.com/huggingface/transformers/pull/24377

1,765,709,257

PR_kwDOCUB6oc5TdFyF

24,377

Better test name and enable pipeline test for `pix2struct`

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? In #24364, `pix2struct` test file didn't get `pipeline_model_mapping` being updated, due to the `heuristic` of finding a test class not working well (try to find the shortest test class name). Let's try to give the test class a better/short/clear name, (despite we don't really have a base model), `Pix2StructModelTest` instead of `Pix2StructTextImageModelTest`. This enables the script `add_pipeline_model_mapping_to_test.py` works for `pix2struct`, and then we get the pipeline test being run.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24377/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24377/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24377", "html_url": "https://github.com/huggingface/transformers/pull/24377", "diff_url": "https://github.com/huggingface/transformers/pull/24377.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24377.patch", "merged_at": 1687278571000 }

https://api.github.com/repos/huggingface/transformers/issues/24376

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24376/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24376/comments

https://api.github.com/repos/huggingface/transformers/issues/24376/events

https://github.com/huggingface/transformers/pull/24376

1,765,674,035

PR_kwDOCUB6oc5Tc-8s

24,376

Migrate doc files to Markdown.

{ "login": "sgugger", "id": 35901082, "node_id": "MDQ6VXNlcjM1OTAxMDgy", "avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sgugger", "html_url": "https://github.com/sgugger", "followers_url": "https://api.github.com/users/sgugger/followers", "following_url": "https://api.github.com/users/sgugger/following{/other_user}", "gists_url": "https://api.github.com/users/sgugger/gists{/gist_id}", "starred_url": "https://api.github.com/users/sgugger/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sgugger/subscriptions", "organizations_url": "https://api.github.com/users/sgugger/orgs", "repos_url": "https://api.github.com/users/sgugger/repos", "events_url": "https://api.github.com/users/sgugger/events{/privacy}", "received_events_url": "https://api.github.com/users/sgugger/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "I'd add a :warning: emoji as a prefix for the disclaimer, but other than that it looks good to me!", "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? The new UI in GitHub makes MDX pretty hard to read for diffs, so this PR migrates the doc files from mdx to md. This shouldn't break anything in the doc-builder.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24376/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24376/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24376", "html_url": "https://github.com/huggingface/transformers/pull/24376", "diff_url": "https://github.com/huggingface/transformers/pull/24376.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24376.patch", "merged_at": 1687298868000 }

https://api.github.com/repos/huggingface/transformers/issues/24375

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24375/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24375/comments

https://api.github.com/repos/huggingface/transformers/issues/24375/events

https://github.com/huggingface/transformers/pull/24375

1,765,667,213

PR_kwDOCUB6oc5Tc9es

24,375

TF LLaMA Port

{ "login": "Rocketknight1", "id": 12866554, "node_id": "MDQ6VXNlcjEyODY2NTU0", "avatar_url": "https://avatars.githubusercontent.com/u/12866554?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Rocketknight1", "html_url": "https://github.com/Rocketknight1", "followers_url": "https://api.github.com/users/Rocketknight1/followers", "following_url": "https://api.github.com/users/Rocketknight1/following{/other_user}", "gists_url": "https://api.github.com/users/Rocketknight1/gists{/gist_id}", "starred_url": "https://api.github.com/users/Rocketknight1/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Rocketknight1/subscriptions", "organizations_url": "https://api.github.com/users/Rocketknight1/orgs", "repos_url": "https://api.github.com/users/Rocketknight1/repos", "events_url": "https://api.github.com/users/Rocketknight1/events{/privacy}", "received_events_url": "https://api.github.com/users/Rocketknight1/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24375). All of your documentation changes will be reflected on that endpoint.", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

MEMBER

null

This is an autoconversion of the LLaMA code to TF by GPT-4. As always, expect things to be broken until I finish debugging it! TODO list: - [ ] Get tests to pass - [ ] No `MainLayer` - we shouldn't need it! Make sure weight naming can still be controlled. - [ ] Explore full `float16` weights - [ ] Explore passing `DTensor` layouts

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24375/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24375/timeline

null

true

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24375", "html_url": "https://github.com/huggingface/transformers/pull/24375", "diff_url": "https://github.com/huggingface/transformers/pull/24375.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24375.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24374

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24374/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24374/comments

https://api.github.com/repos/huggingface/transformers/issues/24374/events

https://github.com/huggingface/transformers/pull/24374

1,765,665,393

PR_kwDOCUB6oc5Tc9HC

24,374

Rename test to be more accurate

{ "login": "sgugger", "id": 35901082, "node_id": "MDQ6VXNlcjM1OTAxMDgy", "avatar_url": "https://avatars.githubusercontent.com/u/35901082?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sgugger", "html_url": "https://github.com/sgugger", "followers_url": "https://api.github.com/users/sgugger/followers", "following_url": "https://api.github.com/users/sgugger/following{/other_user}", "gists_url": "https://api.github.com/users/sgugger/gists{/gist_id}", "starred_url": "https://api.github.com/users/sgugger/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sgugger/subscriptions", "organizations_url": "https://api.github.com/users/sgugger/orgs", "repos_url": "https://api.github.com/users/sgugger/repos", "events_url": "https://api.github.com/users/sgugger/events{/privacy}", "received_events_url": "https://api.github.com/users/sgugger/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24374). All of your documentation changes will be reflected on that endpoint." ]

1,687

COLLABORATOR

null

# What does this PR do? Tiny fix but this integration test actually tests Finn to English so let's name it accordingly.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24374/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24374/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24374", "html_url": "https://github.com/huggingface/transformers/pull/24374", "diff_url": "https://github.com/huggingface/transformers/pull/24374.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24374.patch", "merged_at": 1687276495000 }

https://api.github.com/repos/huggingface/transformers/issues/24373

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24373/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24373/comments

https://api.github.com/repos/huggingface/transformers/issues/24373/events

https://github.com/huggingface/transformers/pull/24373

1,765,632,291

PR_kwDOCUB6oc5Tc2eu

24,373

Add a check in `ImageToTextPipeline._forward`

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? Inside `ImageToTextPipeline.preprocess`, we have ```python if self.model.config.model_type == "git" and prompt is None: model_inputs["input_ids"] = None ``` So it may happen with a list of `None` (for Git model), and the `_forward` fail. This PR add a check and change the above case to a single `None` value to avoid failure.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24373/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24373/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24373", "html_url": "https://github.com/huggingface/transformers/pull/24373", "diff_url": "https://github.com/huggingface/transformers/pull/24373.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24373.patch", "merged_at": 1687277254000 }

https://api.github.com/repos/huggingface/transformers/issues/24372

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24372/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24372/comments

https://api.github.com/repos/huggingface/transformers/issues/24372/events

https://github.com/huggingface/transformers/issues/24372

1,765,495,469

I_kwDOCUB6oc5pO1Kt

24,372

`resize_token_embeddings` breaks `gpt2` generation

{ "login": "vwxyzjn", "id": 5555347, "node_id": "MDQ6VXNlcjU1NTUzNDc=", "avatar_url": "https://avatars.githubusercontent.com/u/5555347?v=4", "gravatar_id": "", "url": "https://api.github.com/users/vwxyzjn", "html_url": "https://github.com/vwxyzjn", "followers_url": "https://api.github.com/users/vwxyzjn/followers", "following_url": "https://api.github.com/users/vwxyzjn/following{/other_user}", "gists_url": "https://api.github.com/users/vwxyzjn/gists{/gist_id}", "starred_url": "https://api.github.com/users/vwxyzjn/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/vwxyzjn/subscriptions", "organizations_url": "https://api.github.com/users/vwxyzjn/orgs", "repos_url": "https://api.github.com/users/vwxyzjn/repos", "events_url": "https://api.github.com/users/vwxyzjn/events{/privacy}", "received_events_url": "https://api.github.com/users/vwxyzjn/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "You are adding a random line in the model. Without fine-tuning it there is no reason for it to continue working." ]

1,687

CONTRIBUTOR

null

### System Info ``` - `transformers` version: 4.30.1 - Platform: Linux-5.15.0-1023-aws-x86_64-with-glibc2.2.5 - Python version: 3.8.11 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): 2.0.1+cu117 (False) - Tensorflow version (GPU?): not installed (NA) - Flax version (CPU?/GPU?/TPU?): not installed (NA) - Jax version: not installed - JaxLib version: not installed - Using GPU in script?: yes - Using distributed or parallel set-up in script?: no ``` ### Who can help? Maybe @younesbelkada ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction ``` from transformers import AutoModelForCausalLM, AutoTokenizer pretrained_model = AutoModelForCausalLM.from_pretrained("gpt2") tokenizer = AutoTokenizer.from_pretrained("gpt2") device = "cpu" input_ids = tokenizer.encode('I enjoy walking with my cute dog', return_tensors='pt') greedy_output = pretrained_model.generate( input_ids=input_ids.to(device), max_new_tokens=50, temperature=0.7, pad_token_id=tokenizer.pad_token_id, ) print("Output:\n" + 100 * '-') print(tokenizer.decode(greedy_output[0])) ### pretrained_model = AutoModelForCausalLM.from_pretrained("gpt2") tokenizer = AutoTokenizer.from_pretrained("gpt2") tokenizer.add_special_tokens({"pad_token": "[PAD]"}) device = "cpu" input_ids = tokenizer.encode('I enjoy walking with my cute dog', return_tensors='pt') greedy_output = pretrained_model.generate( input_ids=input_ids.to(device), max_new_tokens=50, temperature=0.7, pad_token_id=tokenizer.pad_token_id, ) print("Output2:\n" + 100 * '-') print(tokenizer.decode(greedy_output[0])) ### pretrained_model = AutoModelForCausalLM.from_pretrained("gpt2") tokenizer = AutoTokenizer.from_pretrained("gpt2") tokenizer.add_special_tokens({"pad_token": "[PAD]"}) device = "cpu" pretrained_model.resize_token_embeddings(len(tokenizer)) # encode context the generation is conditioned on input_ids = tokenizer.encode('I enjoy walking with my cute dog', return_tensors='pt') greedy_output = pretrained_model.generate( input_ids=input_ids.to(device), max_new_tokens=50, temperature=0.7, ) print("Output3:\n" + 100 * '-') print(tokenizer.decode(greedy_output[0])) ``` ``` The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results. Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation. Output: ---------------------------------------------------------------------------------------------------- I enjoy walking with my cute dog, but I'm not sure if I'll ever be able to walk with my dog. I'm not sure if I'll ever be able to walk with my dog. I'm not sure if I'll ever be able to walk with my Output2: ---------------------------------------------------------------------------------------------------- I enjoy walking with my cute dog, but I'm not sure if I'll ever be able to walk with my dog. I'm not sure if I'll ever be able to walk with my dog. I'm not sure if I'll ever be able to walk with my Output3: ---------------------------------------------------------------------------------------------------- I enjoy walking with my cute dog[PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD][PAD] ``` ### Expected behavior According to https://stackoverflow.com/a/69194717/6611317, ``` tokenizer.add_special_tokens({"pad_token": "[PAD]"}) pretrained_model.resize_token_embeddings(len(tokenizer)) ``` Should work as expected and does not completely break generation.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24372/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24372/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24371

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24371/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24371/comments

https://api.github.com/repos/huggingface/transformers/issues/24371/events

https://github.com/huggingface/transformers/issues/24371

1,765,441,743

I_kwDOCUB6oc5pOoDP

24,371

Future compatibility with LangChain

{ "login": "LucasMartinCalderon", "id": 25382998, "node_id": "MDQ6VXNlcjI1MzgyOTk4", "avatar_url": "https://avatars.githubusercontent.com/u/25382998?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LucasMartinCalderon", "html_url": "https://github.com/LucasMartinCalderon", "followers_url": "https://api.github.com/users/LucasMartinCalderon/followers", "following_url": "https://api.github.com/users/LucasMartinCalderon/following{/other_user}", "gists_url": "https://api.github.com/users/LucasMartinCalderon/gists{/gist_id}", "starred_url": "https://api.github.com/users/LucasMartinCalderon/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/LucasMartinCalderon/subscriptions", "organizations_url": "https://api.github.com/users/LucasMartinCalderon/orgs", "repos_url": "https://api.github.com/users/LucasMartinCalderon/repos", "events_url": "https://api.github.com/users/LucasMartinCalderon/events{/privacy}", "received_events_url": "https://api.github.com/users/LucasMartinCalderon/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

Is there a specific timeline for future LLM agents compatibility with LangChain? What other current compatibility solutions with LangChain are there?

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24371/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24371/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24369

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24369/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24369/comments

https://api.github.com/repos/huggingface/transformers/issues/24369/events

https://github.com/huggingface/transformers/issues/24369

1,765,063,421

I_kwDOCUB6oc5pNLr9

24,369

Additional option for text generation when setting num_beam_groups

{ "login": "hukuda222", "id": 21185928, "node_id": "MDQ6VXNlcjIxMTg1OTI4", "avatar_url": "https://avatars.githubusercontent.com/u/21185928?v=4", "gravatar_id": "", "url": "https://api.github.com/users/hukuda222", "html_url": "https://github.com/hukuda222", "followers_url": "https://api.github.com/users/hukuda222/followers", "following_url": "https://api.github.com/users/hukuda222/following{/other_user}", "gists_url": "https://api.github.com/users/hukuda222/gists{/gist_id}", "starred_url": "https://api.github.com/users/hukuda222/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/hukuda222/subscriptions", "organizations_url": "https://api.github.com/users/hukuda222/orgs", "repos_url": "https://api.github.com/users/hukuda222/repos", "events_url": "https://api.github.com/users/hukuda222/events{/privacy}", "received_events_url": "https://api.github.com/users/hukuda222/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @gante ", "Hey @hukuda222 \r\n\r\n`generate` is already a configuration behemoth, and we would be adding one more flag. By default, we are reluctant to add more flags unless the benefits are large OR there is demand for the option. As such, I'm going to propose the same as I do in similar issues ([e.g.](https://github.com/huggingface/transformers/issues/22168#issuecomment-1477998997))!\r\n\r\nIf this comment gets 10 reactions/this issue gets mentioned 10 times, then it means that folks have been searching for this feature. In that case, I'll greenlight the suggestion, and let's add it to the codebase. That way, we can balance HF's limited maintenance resources with actual feature demand! (Whoever does the 10th react, plz tag me)\r\n\r\n@hukuda222 does that sound good to you?", "@gante Sounds good. Thank you for your quick response.", "@gante \r\nSorry for the delay. I thought the output I presented earlier might be a bug in the current code. Diverse beam search is a method that generates `num_beams//num_beam_groups` sentences for each group independently. However, the current code uses one BeamHypotheses shared by all groups. Therefore, group A will generate two sentences before group B outputs a sentence.\r\n\r\nhttps://github.com/huggingface/transformers/blob/ad78d9597b224443e9fe65a94acc8c0bc48cd039/src/transformers/generation/beam_search.py#L178-L186\r\n\r\nThis is a problem that can be solved by creating as many BeamHypotheses as there are groups. I would like to address this inconvenience in the form of a modification to the diverse beam search implementation, rather than adding an option. If you don't mind, could you give me your opinion?", "Hey @hukuda222 👋 \r\n\r\nI've had a deeper look at group beam search, and it doesn't not seem to be working properly. For instance, the snippet below produces the same sequence on all beams, and that should not happen (each beam should generate different continuations).\r\n\r\nI don't have the bandwidth to fix it immediately, so if you're able to contribute we'd deeply appreciate 🙌 Since it is broken (and thus no retrocompatibility needs to be respected), feel free to also change the behavior of `num_return_sequences` in group beam search to prioritize returning from different beams, which makes more sense.\r\n\r\n___________________________________\r\n```python\r\nfrom transformers import AutoTokenizer, AutoModelForCausalLM\r\n\r\nmodel = AutoModelForCausalLM.from_pretrained(\"gpt2\")\r\ntokenizer = AutoTokenizer.from_pretrained(\"gpt2\")\r\ninputs = tokenizer([\"The full name of Donald is Donald\"], return_tensors=\"pt\")\r\n\r\noutputs = model.generate(**inputs, num_beams=4, num_beam_groups=4, num_return_sequences=4)\r\nprint(\"\\n\".join(tokenizer.batch_decode(outputs, skip_special_tokens=True)))\r\n# Outputs the following sequence 4 times. Each beam should return different sequences.\r\n# The full name of Donald is Donald J. Trump Jr. The full name of Donald is Donald J\r\n```", "@gante \r\nThanks for doing the research. I will send PR as soon as I can fix it." ]

1,687

CONTRIBUTOR

null

### Feature request I propose to add `num_return_sequences_per_groups` as an argument to `generate` function of `transformers.GenerationMixin`. Setting this will output `num_return_sequences_per_groups` sentences each group. Use cases are as follows: code: ```python outputs = model.generate( tokenizer.encode(text, return_tensors="pt", max_length=512), num_beam_groups=3, num_beams=12, diversity_penalty=1.0, num_return_sequences_per_groups=2, ) for output in outputs: print(tokenizer.decode(output, skip_special_tokens=True)) ``` output: ``` A flock of birds flying over the ocean. A flock of birds flying over a beach. Birds flying over the water in the sun. Birds flying the water near a mountain. Several birds are flying over a body of water. Several birds flying over a body of water. ``` The example referred to https://arxiv.org/abs/1610.02424 . ### Motivation As shown below, the output may have little difference when `num_beam_groups` and `num_beams` have the same values. code: ```python from transformers import ( AutoTokenizer, AutoModelForSeq2SeqLM, ) model = AutoModelForSeq2SeqLM.from_pretrained("facebook/bart-large-xsum") tokenizer = AutoTokenizer.from_pretrained("facebook/bart-large-xsum") text = "The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration." outputs = model.generate( tokenizer.encode(text, return_tensors="pt", max_length=512), num_beam_groups=2, num_beams=2, diversity_penalty=1000000.0, num_return_sequences=2, ) for output in outputs: print(tokenizer.decode(output, skip_special_tokens=True)) ``` output: ``` A number Of research projects have investigated the role of the brain's encoder and decoder in the control of the encoded sequences. A number Of research projects have investigated the role of the brain's encoder and decoder in the control of the encoded sequences.. ``` This problem occurs because num_beam*2 is searched as an implementation of beam search. Such output is undesirable. This example is only for clarity. Even in the general case, the current implementation does not guarantee diversity because the output is ordered by score. Therefore, I would like to enable a variety of outputs with this option. ### Your contribution If it looks good, I will implement it.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24369/reactions", "total_count": 2, "+1": 1, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 1, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24369/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24368

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24368/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24368/comments

https://api.github.com/repos/huggingface/transformers/issues/24368/events

https://github.com/huggingface/transformers/pull/24368

1,764,971,867

PR_kwDOCUB6oc5Tan3Q

24,368

[Tokenizer doc] Clarification about `add_prefix_space`

{ "login": "ArthurZucker", "id": 48595927, "node_id": "MDQ6VXNlcjQ4NTk1OTI3", "avatar_url": "https://avatars.githubusercontent.com/u/48595927?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ArthurZucker", "html_url": "https://github.com/ArthurZucker", "followers_url": "https://api.github.com/users/ArthurZucker/followers", "following_url": "https://api.github.com/users/ArthurZucker/following{/other_user}", "gists_url": "https://api.github.com/users/ArthurZucker/gists{/gist_id}", "starred_url": "https://api.github.com/users/ArthurZucker/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ArthurZucker/subscriptions", "organizations_url": "https://api.github.com/users/ArthurZucker/orgs", "repos_url": "https://api.github.com/users/ArthurZucker/repos", "events_url": "https://api.github.com/users/ArthurZucker/events{/privacy}", "received_events_url": "https://api.github.com/users/ArthurZucker/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "This blog post on constrained decoding also uses `add_prefix_space` in `__call__` https://huggingface.co/blog/constrained-beam-search\r\n", "The blog is not hosted on `transformers` but on `blog`, will open a PR for that too later on, thanks for the catch 😉 " ]

1,687

COLLABORATOR

null

# What does this PR do? Adresses #17391, updates the documentation that suggested to use `add_prefix_space` when calling the tokenizer

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24368/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24368/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24368", "html_url": "https://github.com/huggingface/transformers/pull/24368", "diff_url": "https://github.com/huggingface/transformers/pull/24368.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24368.patch", "merged_at": 1687278120000 }

https://api.github.com/repos/huggingface/transformers/issues/24367

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24367/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24367/comments

https://api.github.com/repos/huggingface/transformers/issues/24367/events

https://github.com/huggingface/transformers/pull/24367

1,764,933,368

PR_kwDOCUB6oc5Tafjx

24,367

[Whisper Docs] Nits

{ "login": "ArthurZucker", "id": 48595927, "node_id": "MDQ6VXNlcjQ4NTk1OTI3", "avatar_url": "https://avatars.githubusercontent.com/u/48595927?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ArthurZucker", "html_url": "https://github.com/ArthurZucker", "followers_url": "https://api.github.com/users/ArthurZucker/followers", "following_url": "https://api.github.com/users/ArthurZucker/following{/other_user}", "gists_url": "https://api.github.com/users/ArthurZucker/gists{/gist_id}", "starred_url": "https://api.github.com/users/ArthurZucker/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ArthurZucker/subscriptions", "organizations_url": "https://api.github.com/users/ArthurZucker/orgs", "repos_url": "https://api.github.com/users/ArthurZucker/repos", "events_url": "https://api.github.com/users/ArthurZucker/events{/privacy}", "received_events_url": "https://api.github.com/users/ArthurZucker/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? Adresses #24342, where it is mentioned that the documentation is conter intuitive. Indeed, after a lot of changes, the default value for the `bos_token` that we use is different, thus no official models (hosted on the hub) use `bos_token = "<startoftranscript>"`

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24367/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24367/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24367", "html_url": "https://github.com/huggingface/transformers/pull/24367", "diff_url": "https://github.com/huggingface/transformers/pull/24367.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24367.patch", "merged_at": 1687281532000 }

https://api.github.com/repos/huggingface/transformers/issues/24366

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24366/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24366/comments

https://api.github.com/repos/huggingface/transformers/issues/24366/events

https://github.com/huggingface/transformers/issues/24366

1,764,835,743

I_kwDOCUB6oc5pMUGf

24,366

Format of the documentation (文档格式和可读性的问题）

{ "login": "KKIverson", "id": 49222488, "node_id": "MDQ6VXNlcjQ5MjIyNDg4", "avatar_url": "https://avatars.githubusercontent.com/u/49222488?v=4", "gravatar_id": "", "url": "https://api.github.com/users/KKIverson", "html_url": "https://github.com/KKIverson", "followers_url": "https://api.github.com/users/KKIverson/followers", "following_url": "https://api.github.com/users/KKIverson/following{/other_user}", "gists_url": "https://api.github.com/users/KKIverson/gists{/gist_id}", "starred_url": "https://api.github.com/users/KKIverson/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/KKIverson/subscriptions", "organizations_url": "https://api.github.com/users/KKIverson/orgs", "repos_url": "https://api.github.com/users/KKIverson/repos", "events_url": "https://api.github.com/users/KKIverson/events{/privacy}", "received_events_url": "https://api.github.com/users/KKIverson/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "sorry，我的浏览器兼容性问题。\r\nMy bad！！！！" ]

1,687

NONE

null

### System Info huggingface的文档格式真的看起来很乱，读起来很不流畅，强烈建议官方优化一下：例如：https://huggingface.co/docs/transformers/installation 1. 段落之间几乎没有分隔符，没有标题，字体大小、行间距之类的也很差； 2. 超链和普通文字的格式几乎一样，容易误触； 3. 代码段也不够美观，看起来就像是从interactive 的IDE写的，不方便初学者复制。 ### Who can help? _No response_ ### Information - [ ] The official example scripts - [ ] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction 希望能改进吧 ### Expected behavior 希望能改进吧

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24366/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24366/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24365

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24365/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24365/comments

https://api.github.com/repos/huggingface/transformers/issues/24365/events

https://github.com/huggingface/transformers/issues/24365

1,764,824,885

I_kwDOCUB6oc5pMRc1

24,365

ValueError: Unexpected result of `train_function` (Empty logs). Please use `Model.compile(..., run_eagerly=True)`, or `tf.config.run_functions_eagerly(True)` for more information of where went wrong, or file a issue/bug to `tf.keras`.

{ "login": "ErHimani", "id": 62006705, "node_id": "MDQ6VXNlcjYyMDA2NzA1", "avatar_url": "https://avatars.githubusercontent.com/u/62006705?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ErHimani", "html_url": "https://github.com/ErHimani", "followers_url": "https://api.github.com/users/ErHimani/followers", "following_url": "https://api.github.com/users/ErHimani/following{/other_user}", "gists_url": "https://api.github.com/users/ErHimani/gists{/gist_id}", "starred_url": "https://api.github.com/users/ErHimani/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ErHimani/subscriptions", "organizations_url": "https://api.github.com/users/ErHimani/orgs", "repos_url": "https://api.github.com/users/ErHimani/repos", "events_url": "https://api.github.com/users/ErHimani/events{/privacy}", "received_events_url": "https://api.github.com/users/ErHimani/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @ErHimani, thanks for raising an issue. \r\n\r\nCould you follow the issue template and provide: \r\n* The running environment: run `transformers-cli env` in the terminal and copy-paste the output\r\n* A minimal code snippet so that was can reproduce the error\r\n\r\nWithout these, we're unable to help.", "hii @amyeroberts I am using the following command:\r\n`python examples\\tensorflow\\contrastive-image-text\\run_clip.py --output_dir .\\clip-roberta-finetuned --vision_model_name_or_path openai/clip-vit-base-patch32 --text_model_name_or_path roberta-base --train_file descriptions.json --image_column image_path --caption_column text --remove_unused_columns=False --do_train --per_device_train_batch_size=\"64\" --per_device_eval_batch_size=\"64\" --learning_rate=\"5e-5\" --warmup_steps=\"0\" --weight_decay 0.1`", "@ErHimani Thanks for providing this. In order for us to be able to reproduce, we'll need `descriptions.json`, or an example sample from the dataset to be able to reproduce. We also require the running environment information, as noted above. ", "@amyeroberts Please find Link of the custom dataset\r\ndescription.json:[https://drive.google.com/file/d/14FGJwXRsxns679-ILGlLcBRpqe8UUmGu/view?usp=sharing](url)\r\nImages:[https://drive.google.com/drive/folders/1yr8zapcCPdxlN-5ZSczOIiIeIyS3K_Vt?usp=sharing](url)", "@ErHimani Apologies for the delay in getting back to this issue. The links to the datasets are no longer working. Is the issue still persisting with the latest transformers release? If so, would you be willing to re-share the example dataset? ", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,701

NONE

null

### System Info I am trying to train the CLIP model with my custom dataset But I am facing the above issue my current version of tensorflow ==2.12.0 ### Who can help? _No response_ ### Information - [ ] The official example scripts - [ ] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction ![image](https://github.com/huggingface/transformers/assets/62006705/a8463187-0a43-4caa-9f0f-d90222434ff4) ### Expected behavior Please provide me the guidance how I can use the code t train on custom dataset

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24365/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24365/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24364

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24364/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24364/comments

https://api.github.com/repos/huggingface/transformers/issues/24364/events

https://github.com/huggingface/transformers/pull/24364

1,764,014,081

PR_kwDOCUB6oc5TXcTT

24,364

Update tiny models for pipeline testing.

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? Update tiny models for pipeline testing.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24364/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24364/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24364", "html_url": "https://github.com/huggingface/transformers/pull/24364", "diff_url": "https://github.com/huggingface/transformers/pull/24364.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24364.patch", "merged_at": 1687264990000 }

https://api.github.com/repos/huggingface/transformers/issues/24363

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24363/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24363/comments

https://api.github.com/repos/huggingface/transformers/issues/24363/events

https://github.com/huggingface/transformers/pull/24363

1,763,940,101

PR_kwDOCUB6oc5TXM8I

24,363

[modelcard] add audio classification to task list

{ "login": "sanchit-gandhi", "id": 93869735, "node_id": "U_kgDOBZhWpw", "avatar_url": "https://avatars.githubusercontent.com/u/93869735?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sanchit-gandhi", "html_url": "https://github.com/sanchit-gandhi", "followers_url": "https://api.github.com/users/sanchit-gandhi/followers", "following_url": "https://api.github.com/users/sanchit-gandhi/following{/other_user}", "gists_url": "https://api.github.com/users/sanchit-gandhi/gists{/gist_id}", "starred_url": "https://api.github.com/users/sanchit-gandhi/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/sanchit-gandhi/subscriptions", "organizations_url": "https://api.github.com/users/sanchit-gandhi/orgs", "repos_url": "https://api.github.com/users/sanchit-gandhi/repos", "events_url": "https://api.github.com/users/sanchit-gandhi/events{/privacy}", "received_events_url": "https://api.github.com/users/sanchit-gandhi/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

CONTRIBUTOR

null

# What does this PR do? Adds audio classification to the modelcard tasks lists, thus enabling model cards to be created for this task (required for https://github.com/huggingface/audio-transformers-course/pull/46)

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24363/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24363/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24363", "html_url": "https://github.com/huggingface/transformers/pull/24363", "diff_url": "https://github.com/huggingface/transformers/pull/24363.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24363.patch", "merged_at": 1687266078000 }

https://api.github.com/repos/huggingface/transformers/issues/24362

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24362/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24362/comments

https://api.github.com/repos/huggingface/transformers/issues/24362/events

https://github.com/huggingface/transformers/issues/24362

1,763,889,516

I_kwDOCUB6oc5pItFs

24,362

For loop support in python interpreter of Transformer agent

{ "login": "dcy0577", "id": 50020414, "node_id": "MDQ6VXNlcjUwMDIwNDE0", "avatar_url": "https://avatars.githubusercontent.com/u/50020414?v=4", "gravatar_id": "", "url": "https://api.github.com/users/dcy0577", "html_url": "https://github.com/dcy0577", "followers_url": "https://api.github.com/users/dcy0577/followers", "following_url": "https://api.github.com/users/dcy0577/following{/other_user}", "gists_url": "https://api.github.com/users/dcy0577/gists{/gist_id}", "starred_url": "https://api.github.com/users/dcy0577/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/dcy0577/subscriptions", "organizations_url": "https://api.github.com/users/dcy0577/orgs", "repos_url": "https://api.github.com/users/dcy0577/repos", "events_url": "https://api.github.com/users/dcy0577/events{/privacy}", "received_events_url": "https://api.github.com/users/dcy0577/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @sgugger ", "Would you like to open a PR for this?", "I would like to, but I don't know how to implement it yet. Do you have any suggestions?", "Should be added with the PR mentioned above :-)" ]

1,687

NONE

null

### Feature request Hello, I would like to add for loop support in https://github.com/huggingface/transformers/blame/c2393cad085e3875ee2206d917d46d15e50602a3/src/transformers/tools/python_interpreter.py Any idea about how to implement this? ### Motivation For loop is quite common in generated code. Usually it will not cause infinite loop. ### Your contribution An additional 'elif' for 'ast.For' before the 'else' will be nice: https://github.com/huggingface/transformers/blame/c2393cad085e3875ee2206d917d46d15e50602a3/src/transformers/tools/python_interpreter.py#L132

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24362/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24362/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24361

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24361/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24361/comments

https://api.github.com/repos/huggingface/transformers/issues/24361/events

https://github.com/huggingface/transformers/issues/24361

1,763,790,540

I_kwDOCUB6oc5pIU7M

24,361

Accelerate preprocessing crashing due to non-tensor input

{ "login": "ElleLeonne", "id": 87243032, "node_id": "MDQ6VXNlcjg3MjQzMDMy", "avatar_url": "https://avatars.githubusercontent.com/u/87243032?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ElleLeonne", "html_url": "https://github.com/ElleLeonne", "followers_url": "https://api.github.com/users/ElleLeonne/followers", "following_url": "https://api.github.com/users/ElleLeonne/following{/other_user}", "gists_url": "https://api.github.com/users/ElleLeonne/gists{/gist_id}", "starred_url": "https://api.github.com/users/ElleLeonne/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ElleLeonne/subscriptions", "organizations_url": "https://api.github.com/users/ElleLeonne/orgs", "repos_url": "https://api.github.com/users/ElleLeonne/repos", "events_url": "https://api.github.com/users/ElleLeonne/events{/privacy}", "received_events_url": "https://api.github.com/users/ElleLeonne/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @ElleLeonne, \r\n\r\nSo that we can best help, could you share a minimal code snippet so that we can reproduce the error and information about the running environment: run `transformers-cli env` in the terminal and copy-paste the output?", "Out of town today, but I'll have something shortly.", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "I'm also getting this error when trying to use data that isn't a tensor. @ElleLeonne did you ever find a solution to this?", "@irowberryFS If you could provide a reproducer for this issue, the full error traceback and information about the running environment then we'll be able to look into it. ", "@amyeroberts I couldn't reproduce it using a small training script. For the small script I am using PyTorch Geometric datasets, and their DataLoaders and Accelerator worked fine. Unfortunately all the other threads about this issue don't contain any snippets of training code. I believe it has to do with the `IterableDataset` as described in this issue #26548 . I have temporarily bypassed the issue by not wrapping the DataLoader in the `accelerator.prepare()` call. Here is my dataset processing code. \r\n\r\n```\r\ndataset = load_dataset(\"parquet\", data_dir=\"../../mobTrain/\", streaming=True)['train']\r\ndataset = dataset.shuffle(seed=42)\r\n\r\ndef convert_mobs(row):\r\n row['mob1'] = process_mob(row['mob1'])\r\n row['mob2'] = process_mob(row['mob2'])\r\n return row\r\n\r\ndataset = dataset.map(convert_mobs, remove_columns=['...', '...'])\r\ndl = DataLoader(dataset, batch_size)\r\n```\r\nHere is the traceback\r\n```\r\nTraceback (most recent call last):\r\n File \"/home/username/projects/GraphBlocking/train_model.py\", line 121, in <module>\r\n main()\r\n File \"/home/username/projects/GraphBlocking/train_model.py\", line 118, in main\r\n train()\r\n File \"/home/username/projects/GraphBlocking/train_model.py\", line 93, in train\r\n for batch in tepoch:\r\n File \"/home/username/projects/venv/lib/python3.11/site-packages/tqdm/std.py\", line 1178, in __iter__\r\n for obj in iterable:\r\n File \"/home/username/projects/venv/lib/python3.11/site-packages/accelerate/data_loader.py\", line 639, in __iter__\r\n next_batch, next_batch_info = self._fetch_batches(main_iterator)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/home/username/projects/venv/lib/python3.11/site-packages/accelerate/data_loader.py\", line 602, in _fetch_batches\r\n batch = concatenate(batches, dim=0)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/home/username/projects/venv/lib/python3.11/site-packages/accelerate/utils/operations.py\", line 530, in concatenate\r\n return type(data[0])({k: concatenate([d[k] for d in data], dim=dim) for k in data[0].keys()})\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/home/username/projects/venv/lib/python3.11/site-packages/accelerate/utils/operations.py\", line 530, in <dictcomp>\r\n return type(data[0])({k: concatenate([d[k] for d in data], dim=dim) for k in data[0].keys()})\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/home/username/projects/venv/lib/python3.11/site-packages/accelerate/utils/operations.py\", line 532, in concatenate\r\n raise TypeError(f\"Can only concatenate tensors but got {type(data[0])}\")\r\nTypeError: Can only concatenate tensors but got <class 'torch_geometric.data.batch.HeteroDataBatch'>\r\n```\r\nThe `DataLoader` is a PyG wrapper of the PyTorch `DataLoader`. I access the data in the typical way of `for batch in dataloader:` Again, I believe the issue is using Accelerate with an `IterableDataset`", "@irowberryFS Have you tried any the suggestions in [this comment](https://github.com/huggingface/transformers/issues/26548#issuecomment-1885798533) or the rest of the thread you linked to? ", "@amyeroberts yes. I set `dispatch_batches=False` and it fixed the issue. ", "@irowberryFS Thanks for reporting back! " ]

1,687

1,705

1,690

NONE

null

### System Info I believe a recent update has caused Accelerate to try and concatenate all tensor data in the input dictionary. This is a problem, because my inputs contain non-tensor data, due to the fact that such data is intermittent and not always provided in the batch. Rather than skipping over this information, it instead tries to torch.cat the data, which results in crashes. ``` Traceback (most recent call last): File "/home/lily/Desktop/Project/finetune_dynamic.py", line 304, in <module> fire.Fire(train) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/fire/core.py", line 141, in Fire component_trace = _Fire(component, args, parsed_flag_args, context, name) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/fire/core.py", line 475, in _Fire component, remaining_args = _CallAndUpdateTrace( File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/fire/core.py", line 691, in _CallAndUpdateTrace component = fn(*varargs, **kwargs) File "/home/lily/Desktop/Emme (copy)/finetune_dynamic.py", line 295, in train trainer.train(resume_from_checkpoint=resume_from_checkpoint) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/transformers/trainer.py", line 1539, in train return inner_training_loop( File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/transformers/trainer.py", line 1779, in _inner_training_loop for step, inputs in enumerate(epoch_iterator): File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/data_loader.py", line 553, in __iter__ next_batch, next_batch_info = self._fetch_batches(main_iterator) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/data_loader.py", line 521, in _fetch_batches batch = concatenate(batches, dim=0) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 413, in concatenate return type(data[0])({k: concatenate([d[k] for d in data], dim=dim) for k in data[0].keys()}) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 413, in <dictcomp> return type(data[0])({k: concatenate([d[k] for d in data], dim=dim) for k in data[0].keys()}) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 411, in concatenate return honor_type(data[0], (concatenate([d[i] for d in data], dim=dim) for i in range(len(data[0])))) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 84, in honor_type return type(obj)(generator) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 411, in <genexpr> return honor_type(data[0], (concatenate([d[i] for d in data], dim=dim) for i in range(len(data[0])))) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 411, in concatenate return honor_type(data[0], (concatenate([d[i] for d in data], dim=dim) for i in range(len(data[0])))) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 84, in honor_type return type(obj)(generator) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 411, in <genexpr> return honor_type(data[0], (concatenate([d[i] for d in data], dim=dim) for i in range(len(data[0])))) File "/home/lily/anaconda3/envs/emme/lib/python3.10/site-packages/accelerate/utils/operations.py", line 415, in concatenate raise TypeError(f"Can only concatenate tensors but got {type(data[0])}") TypeError: Can only concatenate tensors but got <class 'str'> ``` I'd like to know if there's a way to turn off this feature in accelerate. I can (and have been, up until now) handle batching my own data. It's only now that it's become a problem. Issue raised on accelerate as well: https://github.com/huggingface/accelerate/issues/1611 ### Who can help? _No response_ ### Information - [ ] The official example scripts - [ ] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction Reproduce: 1: Create model that accepts non-tensor inputs as a nested list 2: Feed input to model via huggingface trainer 3: Observe crash ### Expected behavior Accelerate should ignore the data that it can't process, and pass it to the model as normal.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24361/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24361/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24360

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24360/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24360/comments

https://api.github.com/repos/huggingface/transformers/issues/24360/events

https://github.com/huggingface/transformers/pull/24360

1,763,767,703

PR_kwDOCUB6oc5TWnxS

24,360

TensorFlow CI fixes

{ "login": "Rocketknight1", "id": 12866554, "node_id": "MDQ6VXNlcjEyODY2NTU0", "avatar_url": "https://avatars.githubusercontent.com/u/12866554?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Rocketknight1", "html_url": "https://github.com/Rocketknight1", "followers_url": "https://api.github.com/users/Rocketknight1/followers", "following_url": "https://api.github.com/users/Rocketknight1/following{/other_user}", "gists_url": "https://api.github.com/users/Rocketknight1/gists{/gist_id}", "starred_url": "https://api.github.com/users/Rocketknight1/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Rocketknight1/subscriptions", "organizations_url": "https://api.github.com/users/Rocketknight1/orgs", "repos_url": "https://api.github.com/users/Rocketknight1/repos", "events_url": "https://api.github.com/users/Rocketknight1/events{/privacy}", "received_events_url": "https://api.github.com/users/Rocketknight1/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

MEMBER

null

I made a lot of changes to the TF tests, and this exposed a few issues. This PR fixes all the exposed issues, so hopefully after this the only remaining CI issues should be related to generation or the `SharedEmbeddings` refactor.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24360/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24360/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24360", "html_url": "https://github.com/huggingface/transformers/pull/24360", "diff_url": "https://github.com/huggingface/transformers/pull/24360.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24360.patch", "merged_at": 1687262362000 }

https://api.github.com/repos/huggingface/transformers/issues/24359

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24359/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24359/comments

https://api.github.com/repos/huggingface/transformers/issues/24359/events

https://github.com/huggingface/transformers/issues/24359

1,763,753,112

I_kwDOCUB6oc5pILyY

24,359

ValueError: Found `optimizer` configured in the DeepSpeed config, but no `scheduler`. Please configure a scheduler in the DeepSpeed config.

{ "login": "luohao123", "id": 49749220, "node_id": "MDQ6VXNlcjQ5NzQ5MjIw", "avatar_url": "https://avatars.githubusercontent.com/u/49749220?v=4", "gravatar_id": "", "url": "https://api.github.com/users/luohao123", "html_url": "https://github.com/luohao123", "followers_url": "https://api.github.com/users/luohao123/followers", "following_url": "https://api.github.com/users/luohao123/following{/other_user}", "gists_url": "https://api.github.com/users/luohao123/gists{/gist_id}", "starred_url": "https://api.github.com/users/luohao123/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/luohao123/subscriptions", "organizations_url": "https://api.github.com/users/luohao123/orgs", "repos_url": "https://api.github.com/users/luohao123/repos", "events_url": "https://api.github.com/users/luohao123/events{/privacy}", "received_events_url": "https://api.github.com/users/luohao123/received_events", "type": "User", "site_admin": false }

[ { "id": 5616426447, "node_id": "LA_kwDOCUB6oc8AAAABTsPdzw", "url": "https://api.github.com/repos/huggingface/transformers/labels/solved", "name": "solved", "color": "B1D6DC", "default": false, "description": "" } ]

closed

false

null

[]

[ "Hi @luohao123, \r\n\r\nSo that we can help you, could you follow the issue template and provide a minimal code snippet to reproduce the error and the running environment: run `transformers-cli env` in the terminal and copy-paste the output? \r\n\r\ncc @pacman100 ", "**TLDR;** if you're in a rush, downgrading to version `<4.30` (4.29.2) worked for me\r\n\r\n**I've had the same issue 👇**\r\nI believe the previous behaviour allowed you to not include any DeepSpeed configuration `scheduler` key and the one specified in your `TrainerArguments` would be used. Now it seems you have to include the corresponding scheduler between DeepSpeed and Hugging Face `Trainer`.\r\n\r\ni.e.\r\n\r\n| DeepSpeed scheduler | Trainer scheduler | Resulting scheduler |\r\n| ----------- | ----------- | ----------- |\r\n| WarmupLR | constant_with_warmup | constant_with_warmup |\r\n| WarmupDecayLR | linear | linear |\r\n\r\nwhereas before you could just ignore the first column and leave it blank to get the same result\r\n\r\n| DeepSpeed scheduler | Trainer scheduler | Resulting scheduler |\r\n| ----------- | ----------- | ----------- |\r\n| | constant_with_warmup | constant_with_warmup |\r\n| | linear | linear |\r\n\r\npersonally, I found it handier before where I only had to specify the scheduler in one place rather than tracking this over a DeepSpeed config and a Trainer config which are generally separate objects.", "Hello, the supported combinations now are:\r\n1. Trainer optimizer + Trainer scheduler - Don't specify these in the DS config and use trainer args\r\n2. DeepSpeed optimizer + DeeepSpeed Scheduler - Specify both in DeepSpeed config and no need to use/specify them via Trainer args (@jackapbutler, please note this as you happen to be doing both)\r\n3. Trainer optimizer + DeepSpeed Scheduler - Don't specify optimizer in DS config; only set the scheduler there. Don't specify the scheduler via Trainer args.\r\n\r\n@luohao123, the case you want is DeepSpeed Optimizer + Trainer Scheduler which isn't supported now. The suggested approach in your case would be to use `Trainer optimizer + Trainer scheduler` (Settting 1. above). \r\n\r\nHope this helps. \r\n\r\n", "@pacman100 I actually got some errors when specifci via trainingargs with cosine scheduler while not specific in deepspeed config:\r\n\r\n```\r\n│ ❱ 485 │ │ self.initialize_optimizer_states() │\r\n│ 486 │ │ see_memory_usage(\"After initializing optimizer states\", force=True) │\r\n│ 487 │ │ │\r\n│ 488 │ │ if dist.get_rank() == 0: │\r\n│ │\r\n│ /root/anaconda3/lib/python3.10/site-packages/deepspeed/runtime/zero/stage_1_and_2.py:620 in │\r\n│ initialize_optimizer_states │\r\n│ │\r\n│ 617 │ │ if isinstance(self.optimizer, torch.optim.Adagrad): │\r\n│ 618 │ │ │ self.optimizer = torch.optim.Adagrad(self.single_partition_of_fp32_groups, * │\r\n│ 619 │ │ else: │\r\n│ ❱ 620 │ │ │ self.optimizer.step() │\r\n│ 621 │ │ │\r\n│ 622 │ │ if not self.cpu_offload: │\r\n│ 623 │ │ │ for group in self.single_partition_of_fp32_groups: │\r\n│ │\r\n│ /root/anaconda3/lib/python3.10/site-packages/torch/optim/lr_scheduler.py:69 in wrapper │\r\n│ │\r\n│ 66 │ │ │ │ instance = instance_ref() │\r\n│ 67 │ │ │ │ instance._step_count += 1 │\r\n│ 68 │ │ │ │ wrapped = func.__get__(instance, cls) │\r\n│ ❱ 69 │ │ │ │ return wrapped(*args, **kwargs) │\r\n│ 70 │ │ │ │\r\n│ 71 │ │ │ # Note that the returned function here is no longer a bound method, │\r\n│ 72 │ │ │ # so attributes like `__func__` and `__self__` no longer exist. │\r\n│ │\r\n│ /root/anaconda3/lib/python3.10/site-packages/torch/optim/optimizer.py:280 in wrapper │\r\n│ │\r\n│ 277 │ │ │ │ │ │ │ raise RuntimeError(f\"{func} must return None or a tuple of ( │\r\n│ 278 │ │ │ │ │ │ │ │ │ │ │ f\"but got {result}.\") │\r\n│ 279 │ │ │ │ │\r\n│ ❱ 280 │ │ │ │ out = func(*args, **kwargs) │\r\n│ 281 │ │ │ │ self._optimizer_step_code() │\r\n│ 282 │ │ │ │ │\r\n│ 283 │ │ │ │ # call optimizer step post hooks │\r\n│ │\r\n│ /root/anaconda3/lib/python3.10/site-packages/torch/optim/optimizer.py:33 in _use_grad │\r\n│ │\r\n│ 30 │ │ prev_grad = torch.is_grad_enabled() │\r\n│ 31 │ │ try: │\r\n│ 32 │ │ │ torch.set_grad_enabled(self.defaults['differentiable']) │\r\n│ ❱ 33 │ │ │ ret = func(self, *args, **kwargs) │\r\n│ 34 │ │ finally: │\r\n│ 35 │ │ │ torch.set_grad_enabled(prev_grad) │\r\n│ 36 │ │ return ret │\r\n│ │\r\n│ /root/anaconda3/lib/python3.10/site-packages/torch/optim/adamw.py:171 in step │\r\n│ │\r\n│ 168 │ │ │ │ state_steps, │\r\n│ 169 │ │ │ ) │\r\n│ 170 │ │ │ │\r\n│ ❱ 171 │ │ │ adamw( │\r\n│ 172 │ │ │ │ params_with_grad, │\r\n│ 173 │ │ │ │ grads, │\r\n│ 174 │ │ │ │ exp_avgs, │\r\n│ │\r\n│ /root/anaconda3/lib/python3.10/site-packages/torch/optim/adamw.py:321 in adamw │\r\n│ │\r\n│ 318 │ else: │\r\n│ 319 │ │ func = _single_tensor_adamw │\r\n│ 320 │ │\r\n│ ❱ 321 │ func( │\r\n│ 322 │ │ params, │\r\n│ 323 │ │ grads, │\r\n│ 324 │ │ exp_avgs, │\r\n│ │\r\n│ /root/anaconda3/lib/python3.10/site-packages/torch/optim/adamw.py:564 in _multi_tensor_adamw │\r\n│ │\r\n│ 561 │ │ │ │ torch._foreach_div_(max_exp_avg_sq_sqrt, bias_correction2_sqrt) │\r\n│ 562 │ │ │ │ denom = torch._foreach_add(max_exp_avg_sq_sqrt, eps) │\r\n│ 563 │ │ │ else: │\r\n│ ❱ 564 │ │ │ │ exp_avg_sq_sqrt = torch._foreach_sqrt(device_exp_avg_sqs) │\r\n│ 565 │ │ │ │ torch._foreach_div_(exp_avg_sq_sqrt, bias_correction2_sqrt) │\r\n│ 566 │ │ │ │ denom = torch._foreach_add(exp_avg_sq_sqrt, eps) │\r\n│ 567 │\r\n╰──────────────────────────────────────────────────────────────────────────────────────────────────╯\r\nRuntimeError: CUDA error: an illegal memory access was encountered\r\nCUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.\r\nFor debugging consider passing CUDA_LAUNCH_BLOCKING=1.\r\nCompile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.\r\n```\r\n\r\nWhich is not right, A100, can u take a llok?\r\n\r\nthis is my ds config:\r\n\r\n```\r\n{\r\n \"zero_allow_untested_optimizer\": true,\r\n \"fp16\": {\r\n \"enabled\": \"auto\",\r\n \"opt_level\": \"O2\",\r\n \"initial_scale_power\": 16,\r\n \"loss_scale_window\": 1000,\r\n \"hysteresis\": 2,\r\n \"min_loss_scale\": 1,\r\n \"loss_scale\": 0\r\n },\r\n \"zero_optimization\": {\r\n \"stage\": 2,\r\n \"allgather_partitions\": true,\r\n \"allgather_bucket_size\": 5e8,\r\n \"overlap_comm\": false,\r\n \"reduce_scatter\": true,\r\n \"reduce_bucket_size\": 5e8,\r\n \"contiguous_gradients\": true\r\n },\r\n \"train_micro_batch_size_per_gpu\": \"auto\",\r\n \"gradient_accumulation_steps\": \"auto\"\r\n}\r\n\r\n```\r\n\r\nthis is my training args:\r\n\r\n```\r\nCUDA_VISIBLE_DEVICES=2,3 deepspeed --master_port 61000 train_full.py \\\r\n --data_path ./data/train_data.json \\\r\n --model_name_or_path ./checkpoints/baichuan-7B/ \\\r\n --per_device_train_batch_size 4 --output_dir out/bc_full \\\r\n --bf16 --num_train_epochs 3 \\\r\n --per_device_eval_batch_size 4 \\\r\n --gradient_accumulation_steps 16 \\\r\n --learning_rate 2e-5 --weight_decay 0. \\\r\n --warmup_ratio 0.03 --lr_scheduler_type \"cosine\" \\\r\n --model_max_length 1024 \\\r\n --logging_steps 50 \\\r\n --lazy_preprocess True \\\r\n --deepspeed configs/ds_s2_fschat.json\r\n```\r\n\r\nwhat did wrong???", "Hello @luohao123, please provide minimal reproducible example for further deep dive. Things work fine for me with official example:\r\n\r\nds config:\r\n\r\n```\r\n{\r\n \"fp16\": {\r\n \"enabled\": \"auto\",\r\n \"loss_scale\": 0,\r\n \"loss_scale_window\": 1000,\r\n \"initial_scale_power\": 16,\r\n \"hysteresis\": 2,\r\n \"min_loss_scale\": 1\r\n },\r\n\r\n \"bf16\": {\r\n \"enabled\": \"auto\"\r\n },\r\n\r\n \"zero_optimization\": {\r\n \"stage\": 2,\r\n \"offload_optimizer\": {\r\n \"device\": \"cpu\",\r\n \"pin_memory\": true\r\n },\r\n \"allgather_partitions\": true,\r\n \"allgather_bucket_size\": 2e8,\r\n \"overlap_comm\": true,\r\n \"reduce_scatter\": true,\r\n \"reduce_bucket_size\": 2e8,\r\n \"contiguous_gradients\": true\r\n },\r\n \"gradient_accumulation_steps\": \"auto\",\r\n \"gradient_clipping\": \"auto\",\r\n \"steps_per_print\": 2000,\r\n \"train_batch_size\": \"auto\",\r\n \"train_micro_batch_size_per_gpu\": \"auto\",\r\n \"wall_clock_breakdown\": false\r\n}\r\n```\r\n\r\nCommand:\r\n```\r\ncd transformers\r\nexport TASK_NAME=mrpc\r\nCUDA_VISIBLE_DEVICES=2,3 deepspeed ./examples/pytorch/text-classification/run_glue.py --model_name_or_path bert-base-cased --task_name $TASK_NAME --do_train --do_eval --max_seq_length 128 --per_device_train_batch_size 16 --learning_rate 5e-5 --num_train_epochs 3 --output_dir /tmp/$TASK_NAME/ --overwrite_output_dir --deepspeed ds_config_zero2.json --lr_scheduler_type \"cosine\"\r\n```\r\n\r\noutput logs:\r\n```\r\n[2023-06-22 09:47:48,765] [INFO] [config.py:964:print] zero_enabled ................. True\r\n[2023-06-22 09:47:48,765] [INFO] [config.py:964:print] zero_force_ds_cpu_optimizer .. True\r\n[2023-06-22 09:47:48,765] [INFO] [config.py:964:print] zero_optimization_stage ...... 2\r\n[2023-06-22 09:47:48,765] [INFO] [config.py:950:print_user_config] json = {\r\n \"fp16\": {\r\n \"enabled\": false, \r\n \"loss_scale\": 0, \r\n \"loss_scale_window\": 1000, \r\n \"initial_scale_power\": 16, \r\n \"hysteresis\": 2, \r\n \"min_loss_scale\": 1\r\n }, \r\n \"bf16\": {\r\n \"enabled\": false\r\n }, \r\n \"zero_optimization\": {\r\n \"stage\": 2, \r\n \"offload_optimizer\": {\r\n \"device\": \"cpu\", \r\n \"pin_memory\": true\r\n }, \r\n \"allgather_partitions\": true, \r\n \"allgather_bucket_size\": 2.000000e+08, \r\n \"overlap_comm\": true, \r\n \"reduce_scatter\": true, \r\n \"reduce_bucket_size\": 2.000000e+08, \r\n \"contiguous_gradients\": true\r\n }, \r\n \"gradient_accumulation_steps\": 1, \r\n \"gradient_clipping\": 1.0, \r\n \"steps_per_print\": inf, \r\n \"train_batch_size\": 32, \r\n \"train_micro_batch_size_per_gpu\": 16, \r\n \"wall_clock_breakdown\": false, \r\n \"zero_allow_untested_optimizer\": true\r\n}\r\nUsing /raid/sourab/.cache/huggingface/torch_extensions/py311_cu118 as PyTorch extensions root...\r\nNo modifications detected for re-loaded extension module utils, skipping build step...\r\nLoading extension module utils...\r\nTime to load utils op: 0.00022840499877929688 seconds\r\n[INFO|trainer.py:1680] 2023-06-22 09:47:48,766 >> ***** Running training *****\r\n[INFO|trainer.py:1681] 2023-06-22 09:47:48,766 >> Num examples = 3,668\r\n[INFO|trainer.py:1682] 2023-06-22 09:47:48,766 >> Num Epochs = 3\r\n[INFO|trainer.py:1683] 2023-06-22 09:47:48,766 >> Instantaneous batch size per device = 16\r\n[INFO|trainer.py:1684] 2023-06-22 09:47:48,766 >> Total train batch size (w. parallel, distributed & accumulation) = 32\r\n[INFO|trainer.py:1685] 2023-06-22 09:47:48,766 >> Gradient Accumulation steps = 1\r\n[INFO|trainer.py:1686] 2023-06-22 09:47:48,766 >> Total optimization steps = 345\r\n[INFO|trainer.py:1687] 2023-06-22 09:47:48,766 >> Number of trainable parameters = 108,311,810\r\n[INFO|integrations.py:727] 2023-06-22 09:47:48,767 >> Automatic Weights & Biases logging enabled, to disable set os.environ[\"WANDB_DISABLED\"] = \"true\"\r\nwandb: Currently logged in as: smangrul. Use `wandb login --relogin` to force relogin\r\nwandb: Tracking run with wandb version 0.15.4\r\nwandb: Run data is saved locally in /home/sourab/transformers/wandb/run-20230622_094749-h2mion2e\r\nwandb: Run `wandb offline` to turn off syncing.\r\nwandb: Syncing run rose-vortex-320\r\nwandb: ⭐️ View project at https://wandb.ai/smangrul/huggingface\r\nwandb: 🚀 View run at https://wandb.ai/smangrul/huggingface/runs/h2mion2e\r\n 0%| | 0/345 [00:00<?, ?it/s]/home/sourab/miniconda3/envs/ml/lib/python3.11/site-packages/deepspeed/runtime/zero/stage_1_and_2.py:1829: UserWarning: The torch.cuda.*DtypeTensor constructors are no longer recommended. It's best to use methods such as torch.tensor(data, dtype=*, device='cuda') to create tensors. (Triggered internally at /opt/conda/conda-bld/pytorch_1687280020902/work/torch/csrc/tensor/python_tensor.cpp:83.)\r\n overflow_gpu = get_accelerator().ByteTensor([overflow])\r\n/home/sourab/miniconda3/envs/ml/lib/python3.11/site-packages/deepspeed/runtime/zero/stage_1_and_2.py:1829: UserWarning: The torch.cuda.*DtypeTensor constructors are no longer recommended. It's best to use methods such as torch.tensor(data, dtype=*, device='cuda') to create tensors. (Triggered internally at /opt/conda/conda-bld/pytorch_1687280020902/work/torch/csrc/tensor/python_tensor.cpp:83.)\r\n overflow_gpu = get_accelerator().ByteTensor([overflow])\r\n100%|████████████████████████████████████████████████████████████████████████████████████████| 345/345 [00:57<00:00, 6.13it/s][INFO|trainer.py:1924] 2023-06-22 09:48:49,820 >> \r\n\r\nTraining completed. Do not forget to share your model on huggingface.co/models =)\r\n\r\n\r\n{'train_runtime': 61.0539, 'train_samples_per_second': 180.234, 'train_steps_per_second': 5.651, 'train_loss': 0.4465487715126812, 'epoch': 3.0}\r\n100%|████████████████████████████████████████████████████████████████████████████████████████| 345/345 [00:57<00:00, 6.03it/s]\r\n[INFO|trainer.py:2832] 2023-06-22 09:48:49,823 >> Saving model checkpoint to /tmp/mrpc/\r\n[INFO|configuration_utils.py:458] 2023-06-22 09:48:49,824 >> Configuration saved in /tmp/mrpc/config.json\r\n[INFO|modeling_utils.py:1845] 2023-06-22 09:48:50,616 >> Model weights saved in /tmp/mrpc/pytorch_model.bin\r\n[INFO|tokenization_utils_base.py:2215] 2023-06-22 09:48:50,617 >> tokenizer config file saved in /tmp/mrpc/tokenizer_config.json\r\n[INFO|tokenization_utils_base.py:2222] 2023-06-22 09:48:50,617 >> Special tokens file saved in /tmp/mrpc/special_tokens_map.json\r\n***** train metrics *****\r\n epoch = 3.0\r\n train_loss = 0.4465\r\n train_runtime = 0:01:01.05\r\n train_samples = 3668\r\n train_samples_per_second = 180.234\r\n train_steps_per_second = 5.651\r\n06/22/2023 09:48:50 - INFO - __main__ - *** Evaluate ***\r\n[INFO|trainer.py:769] 2023-06-22 09:48:50,645 >> The following columns in the evaluation set don't have a corresponding argument in `BertForSequenceClassification.forward` and have been ignored: sentence1, sentence2, idx. If sentence1, sentence2, idx are not expected by `BertForSequenceClassification.forward`, you can safely ignore this message.\r\n[INFO|trainer.py:3106] 2023-06-22 09:48:50,646 >> ***** Running Evaluation *****\r\n[INFO|trainer.py:3108] 2023-06-22 09:48:50,646 >> Num examples = 408\r\n[INFO|trainer.py:3111] 2023-06-22 09:48:50,646 >> Batch size = 8\r\n100%|██████████████████████████████████████████████████████████████████████████████████████████| 26/26 [00:00<00:00, 52.94it/s]\r\n***** eval metrics *****\r\n epoch = 3.0\r\n eval_accuracy = 0.8431\r\n eval_combined_score = 0.8664\r\n eval_f1 = 0.8897\r\n eval_loss = 0.3868\r\n eval_runtime = 0:00:00.51\r\n eval_samples = 408\r\n eval_samples_per_second = 797.59\r\n eval_steps_per_second = 50.827\r\nwandb: Waiting for W&B process to finish... (success).\r\n[2023-06-22 09:48:52,926] [INFO] [launch.py:347:main] Process 3002010 exits successfully.\r\nwandb: \r\nwandb: Run history:\r\nwandb: eval/accuracy ▁\r\nwandb: eval/combined_score ▁\r\nwandb: eval/f1 ▁\r\nwandb: eval/loss ▁\r\nwandb: eval/runtime ▁\r\nwandb: eval/samples_per_second ▁\r\nwandb: eval/steps_per_second ▁\r\nwandb: train/epoch ▁▁\r\nwandb: train/global_step ▁▁\r\nwandb: train/total_flos ▁\r\nwandb: train/train_loss ▁\r\nwandb: train/train_runtime ▁\r\nwandb: train/train_samples_per_second ▁\r\nwandb: train/train_steps_per_second ▁\r\nwandb: \r\nwandb: Run summary:\r\nwandb: eval/accuracy 0.84314\r\nwandb: eval/combined_score 0.8664\r\nwandb: eval/f1 0.88966\r\nwandb: eval/loss 0.38684\r\nwandb: eval/runtime 0.5115\r\nwandb: eval/samples_per_second 797.59\r\nwandb: eval/steps_per_second 50.827\r\nwandb: train/epoch 3.0\r\nwandb: train/global_step 345\r\nwandb: train/total_flos 726186493739008.0\r\nwandb: train/train_loss 0.44655\r\nwandb: train/train_runtime 61.0539\r\nwandb: train/train_samples_per_second 180.234\r\nwandb: train/train_steps_per_second 5.651\r\nwandb: \r\nwandb: 🚀 View run rose-vortex-320 at: https://wandb.ai/smangrul/huggingface/runs/h2mion2e\r\nwandb: Synced 6 W&B file(s), 0 media file(s), 2 artifact file(s) and 0 other file(s)\r\nwandb: Find logs at: ./wandb/run-20230622_094749-h2mion2e/logs\r\n[2023-06-22 09:49:01,927] [INFO] [launch.py:347:main] Process 3002009 exits successfully.\r\n```", "@pacman100 thank u, let me try your config and have a test again, I notice your config are not exactly as mine.", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "> Hello, the supported combinations now are:\r\n> \r\n> 1. Trainer optimizer + Trainer scheduler - Don't specify these in the DS config and use trainer args\r\n> 2. DeepSpeed optimizer + DeeepSpeed Scheduler - Specify both in DeepSpeed config and no need to use/specify them via Trainer args (@jackapbutler, please note this as you happen to be doing both)\r\n> 3. Trainer optimizer + DeepSpeed Scheduler - Don't specify optimizer in DS config; only set the scheduler there. Don't specify the scheduler via Trainer args.\r\n> \r\n> @luohao123, the case you want is DeepSpeed Optimizer + Trainer Scheduler which isn't supported now. The suggested approach in your case would be to use `Trainer optimizer + Trainer scheduler` (Settting 1. above).\r\n> \r\n> Hope this helps.\r\n\r\nHi, I want to know if I use setting 1, will the optimizer utilize DeepSpeed's cpuAdam? ", "> Hi, I want to know if I use setting 1, will the optimizer utilize DeepSpeed's cpuAdam?\r\n\r\nYes, by default `zero_force_ds_cpu_optimizer` is set to True if not explicitly specified in the ds_config. As such, it will leverage the DeepSpeed's cpuAdam when offloading as it is strongly recommended by DeepSpeed team", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored.", "I'm trying to use DeepSpeed optimizer + Trainer scheduler because DeepSpeed has the most best optimizer (fused Adam) and Trainer has the best scheduler for my use case (cosine). DeepSpeed does not support cosine. Why was `DeepSpeed optimizer + Trainer scheduler` deprecate without any warning? I think this is a mistake and that you should reconsider @pacman100.", "Hello @michaelroyzen, the PRs https://github.com/huggingface/transformers/pull/25863 and https://github.com/huggingface/accelerate/pull/1909 should bring back the support for `DeepSpeed optimizer + Trainer scheduler`. Could you try it out and let us know.", "Seems to work well so far @pacman100. Thanks!", "Hi @pacman100 ,\r\nIs PR #25863 part of the latest transformers version?\r\n\r\nI still observe the following error, despite using the `ds_config_z3_ds_optim_hf_scheduler.json`.\r\n ValueError: Found `optimizer` configured in the DeepSpeed config, but no `scheduler`. Please configure a scheduler in the DeepSpeed config. ", "Hello @awasthiabhijeet, it should be part of the latest release, could you recheck it?", "Thanks, @pacman100 :)\r\nYes, hf scheduler + ds optimizer combination is working well with the latest release!\r\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,698

NONE

null

ValueError: Found `optimizer` configured in the DeepSpeed config, but no `scheduler`. Please configure a scheduler in the DeepSpeed config. Am using `--warmup_ratio 0.03 --lr_scheduler_type "cosine" \` Here, and I didn't found a properly shceduler in deepspeed ssame as cosine, what should to set?

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24359/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24359/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24358

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24358/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24358/comments

https://api.github.com/repos/huggingface/transformers/issues/24358/events

https://github.com/huggingface/transformers/pull/24358

1,763,612,699

PR_kwDOCUB6oc5TWGJQ

24,358

Fix the order in `GPTNeo`'s docstring

{ "login": "qgallouedec", "id": 45557362, "node_id": "MDQ6VXNlcjQ1NTU3MzYy", "avatar_url": "https://avatars.githubusercontent.com/u/45557362?v=4", "gravatar_id": "", "url": "https://api.github.com/users/qgallouedec", "html_url": "https://github.com/qgallouedec", "followers_url": "https://api.github.com/users/qgallouedec/followers", "following_url": "https://api.github.com/users/qgallouedec/following{/other_user}", "gists_url": "https://api.github.com/users/qgallouedec/gists{/gist_id}", "starred_url": "https://api.github.com/users/qgallouedec/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/qgallouedec/subscriptions", "organizations_url": "https://api.github.com/users/qgallouedec/orgs", "repos_url": "https://api.github.com/users/qgallouedec/repos", "events_url": "https://api.github.com/users/qgallouedec/events{/privacy}", "received_events_url": "https://api.github.com/users/qgallouedec/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Et voilà :)", "_The documentation is not available anymore as the PR was closed or merged._", "@qgallouedec To get the setup_and_quality tests passing, you'll need to run `make style` at the top level of the repo and push any changes to this branch. " ]

1,687

1,701

1,687

CONTRIBUTOR

null

# What does this PR do?   Fixes # (issue) ## Before submitting - [x] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24358/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24358/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24358", "html_url": "https://github.com/huggingface/transformers/pull/24358", "diff_url": "https://github.com/huggingface/transformers/pull/24358.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24358.patch", "merged_at": 1687197576000 }

https://api.github.com/repos/huggingface/transformers/issues/24357

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24357/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24357/comments

https://api.github.com/repos/huggingface/transformers/issues/24357/events

https://github.com/huggingface/transformers/pull/24357

1,763,568,939

PR_kwDOCUB6oc5TV8lw

24,357

Make `AutoFormer` work with previous torch version

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? Without `import torch.utils.checkpoint` (which we have in other files, like `Bart`), with torch 1.13, we got an error (running `RUN_SLOW=1 python3 -m pytest -v tests/models/autoformer/test_modeling_autoformer.py::AutoformerModelTest::test_training_gradient_checkpointing`) ```bash > layer_outputs = torch.utils.checkpoint.checkpoint( create_custom_forward(encoder_layer), hidden_states, attention_mask, (head_mask[idx] if head_mask is not None else None), ) E AttributeError: module 'torch.utils' has no attribute 'checkpoint' ``` Let's make it work with previous torch version(s) ❤️ .

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24357/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24357/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24357", "html_url": "https://github.com/huggingface/transformers/pull/24357", "diff_url": "https://github.com/huggingface/transformers/pull/24357.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24357.patch", "merged_at": 1687183326000 }

https://api.github.com/repos/huggingface/transformers/issues/24356

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24356/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24356/comments

https://api.github.com/repos/huggingface/transformers/issues/24356/events

https://github.com/huggingface/transformers/issues/24356

1,763,555,047

I_kwDOCUB6oc5pHbbn

24,356

Deepspeed OOM When training 7B model on V100 16GB (2)

{ "login": "shahules786", "id": 25312635, "node_id": "MDQ6VXNlcjI1MzEyNjM1", "avatar_url": "https://avatars.githubusercontent.com/u/25312635?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shahules786", "html_url": "https://github.com/shahules786", "followers_url": "https://api.github.com/users/shahules786/followers", "following_url": "https://api.github.com/users/shahules786/following{/other_user}", "gists_url": "https://api.github.com/users/shahules786/gists{/gist_id}", "starred_url": "https://api.github.com/users/shahules786/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/shahules786/subscriptions", "organizations_url": "https://api.github.com/users/shahules786/orgs", "repos_url": "https://api.github.com/users/shahules786/repos", "events_url": "https://api.github.com/users/shahules786/events{/privacy}", "received_events_url": "https://api.github.com/users/shahules786/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Did you solved? " ]

1,687

1,697

1,687

NONE

null

### System Info - Python version 3.8 - transformers - installed from source latest **Describe the bug** OOM When training 7B model on V100 16GB (2) with zero stage 2 and CPU offloading even when memory estimation showed far less per GPU mem requirement. ``` -- memory estimation-- DEVICES ['Tesla V100-PCIE-16GB', 'Tesla V100-PCIE-16GB'] -------------ZERO 2------------ Estimated memory needed for params, optim states and gradients for a: HW: Setup with 1 node, 2 GPUs per node. SW: Model with 6650M total params. per CPU | per GPU | Options 148.66GB | 12.39GB | offload_optimizer=cpu 74.33GB | 74.33GB | offload_optimizer=none ``` **Screenshots** nvidia-smi during run <img width="648" alt="Screenshot 2023-06-19 at 5 30 41 PM" src="https://github.com/microsoft/DeepSpeed/assets/25312635/47a512c6-b509-49c2-b111-8f7e9dac8532"> RAM usage <img width="723" alt="Screenshot 2023-06-19 at 5 55 08 PM" src="https://github.com/microsoft/DeepSpeed/assets/25312635/e7e97e07-7203-42a7-8b89-468c2de35546"> Can see free RAM available. **System info (please complete the following information):** - OS: CentOS Linux - GPU count and types : V100 16B X 2 single node ### Who can help? @pacman100 ### Information - [ ] The official example scripts - [ ] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction Steps to reproduce the behavior: My code is [here](https://github.com/explodinggradients/Funtuner/blob/main/funtuner/trainer.py) Run deepspeed funtuner/trainer.py export PYTHONPATH="${PYTHONPATH}:/your-path/Funtuner" please change the log_dir to your folder [here](https://github.com/explodinggradients/Funtuner/blob/c4e66209d5ee276a7eb8caf582435f1eaafbf18f/funtuner/config/config.yaml#L4) also you might want to set log_wandb=False `dev-train` branch ### Expected behavior Run w/o OOM error.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24356/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24356/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24354

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24354/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24354/comments

https://api.github.com/repos/huggingface/transformers/issues/24354/events

https://github.com/huggingface/transformers/issues/24354

1,763,502,622

I_kwDOCUB6oc5pHOoe

24,354

PEFT Models are not resuming from checkpoint as expected.

{ "login": "techthiyanes", "id": 25921035, "node_id": "MDQ6VXNlcjI1OTIxMDM1", "avatar_url": "https://avatars.githubusercontent.com/u/25921035?v=4", "gravatar_id": "", "url": "https://api.github.com/users/techthiyanes", "html_url": "https://github.com/techthiyanes", "followers_url": "https://api.github.com/users/techthiyanes/followers", "following_url": "https://api.github.com/users/techthiyanes/following{/other_user}", "gists_url": "https://api.github.com/users/techthiyanes/gists{/gist_id}", "starred_url": "https://api.github.com/users/techthiyanes/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/techthiyanes/subscriptions", "organizations_url": "https://api.github.com/users/techthiyanes/orgs", "repos_url": "https://api.github.com/users/techthiyanes/repos", "events_url": "https://api.github.com/users/techthiyanes/events{/privacy}", "received_events_url": "https://api.github.com/users/techthiyanes/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @techthiyanes \r\nThank you very much for double checking, here are the snippets that I have ran and they work fine on my end using the branh you have mentioned:\r\n\r\n<details><summary>Without `resume_from_checkpoint`</summary>\r\n\r\n```python\r\n\r\nimport os\r\nfrom transformers import TrainingArguments\r\nfrom datasets import load_dataset\r\nfrom trl import SFTTrainer\r\nfrom peft import LoraConfig\r\n\r\ndataset = load_dataset(\"imdb\", split=\"train\")\r\noutput_dir = \"test\"\r\n\r\ntraining_args = TrainingArguments(\r\n output_dir=output_dir,\r\n per_device_train_batch_size=1,\r\n per_device_eval_batch_size=1,\r\n max_steps=5,\r\n save_steps=1,\r\n save_strategy='steps'\r\n)\r\n\r\npeft_config = LoraConfig(\r\n r=16,\r\n lora_alpha=32,\r\n lora_dropout=0.05,\r\n bias=\"none\",\r\n task_type=\"CAUSAL_LM\",\r\n)\r\n\r\ntrainer = SFTTrainer(\r\n \"EleutherAI/gpt-neo-125m\",\r\n train_dataset=dataset,\r\n args=training_args,\r\n dataset_text_field=\"text\",\r\n peft_config=peft_config\r\n)\r\ntrainer.train()\r\ntrainer.save_model(os.path.join(output_dir, \"checkpoint-1\"))\r\ntrainer.train(resume_from_checkpoint=True)\r\n```\r\n\r\n</details>\r\n\r\n<details><summary>With `resume_from_checkpoint`</summary>\r\n\r\n```python\r\n\r\nimport os\r\nfrom transformers import TrainingArguments\r\nfrom datasets import load_dataset\r\nfrom trl import SFTTrainer\r\nfrom peft import LoraConfig\r\n\r\ndataset = load_dataset(\"imdb\", split=\"train\")\r\noutput_dir = \"test\"\r\n\r\ntraining_args = TrainingArguments(\r\n output_dir=output_dir,\r\n per_device_train_batch_size=1,\r\n per_device_eval_batch_size=1,\r\n max_steps=5,\r\n save_steps=1,\r\n save_strategy='steps'\r\n)\r\n\r\npeft_config = LoraConfig(\r\n r=16,\r\n lora_alpha=32,\r\n lora_dropout=0.05,\r\n bias=\"none\",\r\n task_type=\"CAUSAL_LM\",\r\n)\r\n\r\ntrainer = SFTTrainer(\r\n \"EleutherAI/gpt-neo-125m\",\r\n train_dataset=dataset,\r\n args=training_args,\r\n dataset_text_field=\"text\",\r\n peft_config=peft_config\r\n)\r\ntrainer.train()\r\ntrainer.save_model(os.path.join(output_dir, \"checkpoint-1\"))\r\ntrainer.train()\r\n```\r\n\r\n</details>\r\n\r\nCan you elaborate more on:\r\n\r\n> For resuming from checkpoint i have updated num of epochs much higher than previous one.\r\nwhile passing as trainer.train(resume from checkpoint=True) then it is showing as can't find a valid checkpoint.\r\nAlso while passing as trainer.train(resume from checkpoint = path of saved model)then it is showing as can't find a valid checkpoint.\r\n\r\nThanks! ", "> ```python\r\n> ```python\r\n> trainer.train(resume_from_checkpoint=True)\r\n> ```\r\n> \r\n> \r\n> \r\n> \r\n> \r\n> \r\n> \r\n> \r\n> \r\n> \r\n> \r\n> ```\r\n\r\nSo far I'm able to replicate the issue.\r\n\r\nSteps I have followed:\r\n\r\nLibaries Installed:\r\n! pip install datasets peft evaluate\r\n!pip install git+https://github.com/huggingface/transformers\r\n\r\nClone PEFT resume from chekpoint branch:\r\n!git clone https://github.com/llohann-speranca/transformers.git -b fix-resume-checkpoint-for-peftmodel\r\n\r\nReplace this folder where the transformers library installed:\r\n!cp -r /content/transformers /usr/local/lib/python3.10/dist-packages/transformers\r\n\r\nRestart the run time.\r\n\r\nThen below code snippet:\r\n\r\nimport os\r\nfrom transformers import TrainingArguments\r\nfrom datasets import load_dataset\r\nfrom trl import SFTTrainer\r\nfrom peft import LoraConfig\r\n\r\ndataset = load_dataset(\"imdb\", split=\"train\")\r\noutput_dir = \"test\"\r\n\r\ntraining_args = TrainingArguments(\r\n output_dir=output_dir,\r\n per_device_train_batch_size=1,\r\n per_device_eval_batch_size=1,\r\n max_steps=5,\r\n save_steps=1,\r\n save_strategy='steps'\r\n)\r\n\r\npeft_config = LoraConfig(\r\n r=16,\r\n lora_alpha=32,\r\n lora_dropout=0.05,\r\n bias=\"none\",\r\n task_type=\"CAUSAL_LM\",\r\n)\r\n\r\ntrainer = SFTTrainer(\r\n \"EleutherAI/gpt-neo-125m\",\r\n train_dataset=dataset,\r\n args=training_args,\r\n dataset_text_field=\"text\",\r\n peft_config=peft_config\r\n)\r\ntrainer.train()\r\ntrainer.save_model(os.path.join(output_dir, \"checkpoint-1\"))\r\ntrainer.train(resume_from_checkpoint=True)\r\n\r\n![image](https://github.com/huggingface/transformers/assets/25921035/8b2230d6-14a7-43f6-b8d3-d6df9d94704d)\r\n\r\n@younesbelkada @@llohann-speranca\r\n\r\nI guess you would have run the snippet via already from modified trainer code that resides internally.\r\n\r\nCould you please try running the code that is downloaded from git on specific branch?\r\n\r\nThanks a lot on your effort on validating this.\r\n", "Hi @techthiyanes \r\nCan you try to install `transformers` with the following command ?\r\n```bash\r\npip install git+https://github.com/llohann-speranca/transformers.git@fix-resume-checkpoint-for-peftmodel\r\n```\r\nThe line 1991 of your traceback doesn't match with the line 1991 of the fork: https://github.com/llohann-speranca/transformers/blob/e01a4aa77073b847b9451c92c2df718a67960df1/src/transformers/trainer.py#L1991 so I believe you did not installed correctly transformers from that branch", "> ```shell\r\n> pip install git+https://github.com/llohann-speranca/transformers.git@fix-resume-checkpoint-for-peftmodel\r\n> ```\r\n\r\nThanks a lot on finding and fixing to help this issue.\r\nNow I am able to resume from checkpoint. It's working for classification and seq2seq models as well." ]

1,687

NONE

null

### System Info transformers : 4.30 ### Who can help? @llohann-speranca @younesbelkada ### Information - [ ] The official example scripts - [ ] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction Please try below code snippet as per example: ```python import os from transformers import TrainingArguments from datasets import load_dataset from trl import SFTTrainer from peft import LoraConfig dataset = load_dataset("imdb", split="train") output_dir = "test" training_args = TrainingArguments( output_dir=output_dir, per_device_train_batch_size=1, per_device_eval_batch_size=1, max_steps=5, save_steps=1, save_strategy='steps' ) peft_config = LoraConfig( r=16, lora_alpha=32, lora_dropout=0.05, bias="none", task_type="CAUSAL_LM", ) trainer = SFTTrainer( "EleutherAI/gpt-neo-125m", train_dataset=dataset, args=training_args, dataset_text_field="text", peft_config=peft_config ) trainer.train() trainer.save_model(os.path.join(output_dir, "checkpoint-1")) trainer.train() ``` For the above code snippet I have pulled @llohann-speranca's resume from checkpoint repo then replaced the installed transformers repo. Inital version of trainer.train() is working without any issues. As mentioned that I have overridden the model by using trainer.save_model(path of saved model). For resuming from checkpoint i have updated num of epochs much higher than previous one. while passing as trainer.train(resume from checkpoint=True) then it is showing as can't find a valid checkpoint. Also while passing as trainer.train(resume from checkpoint = path of saved model)then it is showing as can't find a valid checkpoint. The same issue persists in the transformers source installed version as well. ### Expected behavior The model should be resumed from checkpoint.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24354/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24354/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24353

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24353/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24353/comments

https://api.github.com/repos/huggingface/transformers/issues/24353/events

https://github.com/huggingface/transformers/pull/24353

1,763,486,363

PR_kwDOCUB6oc5TVqWA

24,353

Fix ImageGPT doctest

{ "login": "amyeroberts", "id": 22614925, "node_id": "MDQ6VXNlcjIyNjE0OTI1", "avatar_url": "https://avatars.githubusercontent.com/u/22614925?v=4", "gravatar_id": "", "url": "https://api.github.com/users/amyeroberts", "html_url": "https://github.com/amyeroberts", "followers_url": "https://api.github.com/users/amyeroberts/followers", "following_url": "https://api.github.com/users/amyeroberts/following{/other_user}", "gists_url": "https://api.github.com/users/amyeroberts/gists{/gist_id}", "starred_url": "https://api.github.com/users/amyeroberts/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/amyeroberts/subscriptions", "organizations_url": "https://api.github.com/users/amyeroberts/orgs", "repos_url": "https://api.github.com/users/amyeroberts/repos", "events_url": "https://api.github.com/users/amyeroberts/events{/privacy}", "received_events_url": "https://api.github.com/users/amyeroberts/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "@ydshieh Running the doctests (properly this time :) ), the tests pass with the ignore statement on the for loop, and fail without (in the same way as on the CI). " ]

1,687

COLLABORATOR

null

# What does this PR do? #24317 Resolved the ImageGPT doc test failing issue, as `clusters` in the image processor were not stored as numpy arrays as expected. This was tested by running the code directly, but I didn't run using ` pytest --doctest-modules src/transformers/models/imagegpt/modeling_imagegpt.py ` 🙃 The tests were failing because some code produces an output e.g. model architecture when caling `model.to`, but no "expected" output is provided. We don't want to check these outputs, so this PR adds controls to ignore. ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests?

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24353/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24353/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24353", "html_url": "https://github.com/huggingface/transformers/pull/24353", "diff_url": "https://github.com/huggingface/transformers/pull/24353.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24353.patch", "merged_at": 1687184610000 }

https://api.github.com/repos/huggingface/transformers/issues/24352

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24352/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24352/comments

https://api.github.com/repos/huggingface/transformers/issues/24352/events

https://github.com/huggingface/transformers/pull/24352

1,763,429,594

PR_kwDOCUB6oc5TVd4f

24,352

Fix device issue in `SwitchTransformers`

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Merge now. Don't hesitate to leave comments if any @ArthurZucker ." ]

1,687

COLLABORATOR

null

# What does this PR do? Need a tiny fix after #24300. Currently, we have a failure ```bash self = <tests.models.switch_transformers.test_modeling_switch_transformers.SwitchTransformersEncoderOnlyModelTest testMethod=test_multi_gpu_data_parallel_forward> @staticmethod def forward(ctx, target_device, dim, *inputs): > assert all(i.device.type != 'cpu' for i in inputs), ( 'Gather function not implemented for CPU tensors' ) E AssertionError: Gather function not implemented for CPU tensors /usr/local/lib/python3.8/dist-packages/torch/nn/parallel/_functions.py:56: AssertionError ```

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24352/reactions", "total_count": 1, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 1, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24352/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24352", "html_url": "https://github.com/huggingface/transformers/pull/24352", "diff_url": "https://github.com/huggingface/transformers/pull/24352.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24352.patch", "merged_at": 1687179965000 }

https://api.github.com/repos/huggingface/transformers/issues/24351

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24351/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24351/comments

https://api.github.com/repos/huggingface/transformers/issues/24351/events

https://github.com/huggingface/transformers/pull/24351

1,763,169,502

PR_kwDOCUB6oc5TUki9

24,351

pin `apex` to a speicifc commit (for DeepSpeed CI docker image)

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? The docker image build for DeepSpeed job in CI fails since ~ one week due to this [apex issue](https://github.com/NVIDIA/apex/issues/1679). Let's pin to the previous commit until the above mentioned issue is resolved on `apex` side. Currently, the DeepSpeed job fails as the above failure prevents to use newer images that include some fixes on `accelerate` side.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24351/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24351/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24351", "html_url": "https://github.com/huggingface/transformers/pull/24351", "diff_url": "https://github.com/huggingface/transformers/pull/24351.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24351.patch", "merged_at": 1687171733000 }

https://api.github.com/repos/huggingface/transformers/issues/24350

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24350/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24350/comments

https://api.github.com/repos/huggingface/transformers/issues/24350/events

https://github.com/huggingface/transformers/pull/24350

1,763,141,993

PR_kwDOCUB6oc5TUeqy

24,350

byebye Hub connection timeout

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? No more timeout for connection to Hub in CI, and everyone is happy with ✅

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24350/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24350/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24350", "html_url": "https://github.com/huggingface/transformers/pull/24350", "diff_url": "https://github.com/huggingface/transformers/pull/24350.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24350.patch", "merged_at": 1687171820000 }

https://api.github.com/repos/huggingface/transformers/issues/24349

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24349/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24349/comments

https://api.github.com/repos/huggingface/transformers/issues/24349/events

https://github.com/huggingface/transformers/pull/24349

1,763,128,388

PR_kwDOCUB6oc5TUbu_

24,349

[GPTNeoX] Nit in config

{ "login": "ArthurZucker", "id": 48595927, "node_id": "MDQ6VXNlcjQ4NTk1OTI3", "avatar_url": "https://avatars.githubusercontent.com/u/48595927?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ArthurZucker", "html_url": "https://github.com/ArthurZucker", "followers_url": "https://api.github.com/users/ArthurZucker/followers", "following_url": "https://api.github.com/users/ArthurZucker/following{/other_user}", "gists_url": "https://api.github.com/users/ArthurZucker/gists{/gist_id}", "starred_url": "https://api.github.com/users/ArthurZucker/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ArthurZucker/subscriptions", "organizations_url": "https://api.github.com/users/ArthurZucker/orgs", "repos_url": "https://api.github.com/users/ArthurZucker/repos", "events_url": "https://api.github.com/users/ArthurZucker/events{/privacy}", "received_events_url": "https://api.github.com/users/ArthurZucker/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? Fixes #23081: when the number of heads is not a divisor of the hidden size, the attention will not work. This is most probably from the design of GPTNeoX's attention.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24349/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24349/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24349", "html_url": "https://github.com/huggingface/transformers/pull/24349", "diff_url": "https://github.com/huggingface/transformers/pull/24349.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24349.patch", "merged_at": 1687281559000 }

https://api.github.com/repos/huggingface/transformers/issues/24348

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24348/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24348/comments

https://api.github.com/repos/huggingface/transformers/issues/24348/events

https://github.com/huggingface/transformers/pull/24348

1,763,101,314

PR_kwDOCUB6oc5TUV3x

24,348

Add mul choice train script

{ "login": "HDThang", "id": 132823983, "node_id": "U_kgDOB-q7rw", "avatar_url": "https://avatars.githubusercontent.com/u/132823983?v=4", "gravatar_id": "", "url": "https://api.github.com/users/HDThang", "html_url": "https://github.com/HDThang", "followers_url": "https://api.github.com/users/HDThang/followers", "following_url": "https://api.github.com/users/HDThang/following{/other_user}", "gists_url": "https://api.github.com/users/HDThang/gists{/gist_id}", "starred_url": "https://api.github.com/users/HDThang/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/HDThang/subscriptions", "organizations_url": "https://api.github.com/users/HDThang/orgs", "repos_url": "https://api.github.com/users/HDThang/repos", "events_url": "https://api.github.com/users/HDThang/events{/privacy}", "received_events_url": "https://api.github.com/users/HDThang/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

1,687

NONE

null

- Modify train script all done tasks - Add common libraries for environments in env.yaml

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24348/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24348/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24348", "html_url": "https://github.com/huggingface/transformers/pull/24348", "diff_url": "https://github.com/huggingface/transformers/pull/24348.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24348.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24346

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24346/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24346/comments

https://api.github.com/repos/huggingface/transformers/issues/24346/events

https://github.com/huggingface/transformers/pull/24346

1,762,923,238

PR_kwDOCUB6oc5TTvWs

24,346

Clean up disk sapce during docker image build for `transformers-pytorch-gpu`

{ "login": "ydshieh", "id": 2521628, "node_id": "MDQ6VXNlcjI1MjE2Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2521628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ydshieh", "html_url": "https://github.com/ydshieh", "followers_url": "https://api.github.com/users/ydshieh/followers", "following_url": "https://api.github.com/users/ydshieh/following{/other_user}", "gists_url": "https://api.github.com/users/ydshieh/gists{/gist_id}", "starred_url": "https://api.github.com/users/ydshieh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ydshieh/subscriptions", "organizations_url": "https://api.github.com/users/ydshieh/orgs", "repos_url": "https://api.github.com/users/ydshieh/repos", "events_url": "https://api.github.com/users/ydshieh/events{/privacy}", "received_events_url": "https://api.github.com/users/ydshieh/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

COLLABORATOR

null

# What does this PR do? PyTorch pipeline CI job start to fail due to ```bash ImportError: accelerate>=0.20.3 is required for a normal functioning of this module, but found accelerate==0.20.2. Try: pip install transformers -U or pip install -e '.[dev]' if you're working with git main ``` **The root cause is: docker image for this job failed to build due to disk space issue** ```bash RROR: Could not install packages due to an OSError: [Errno 28] No space left on device ``` As usual, let's us save Space!

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24346/reactions", "total_count": 1, "+1": 1, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24346/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24346", "html_url": "https://github.com/huggingface/transformers/pull/24346", "diff_url": "https://github.com/huggingface/transformers/pull/24346.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24346.patch", "merged_at": 1687172043000 }

https://api.github.com/repos/huggingface/transformers/issues/24345

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24345/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24345/comments

https://api.github.com/repos/huggingface/transformers/issues/24345/events

https://github.com/huggingface/transformers/issues/24345

1,762,688,184

I_kwDOCUB6oc5pEHy4

24,345

Trainer reports batch size different from argument on multiple GPUs with DP

{ "login": "cgbahk", "id": 34672141, "node_id": "MDQ6VXNlcjM0NjcyMTQx", "avatar_url": "https://avatars.githubusercontent.com/u/34672141?v=4", "gravatar_id": "", "url": "https://api.github.com/users/cgbahk", "html_url": "https://github.com/cgbahk", "followers_url": "https://api.github.com/users/cgbahk/followers", "following_url": "https://api.github.com/users/cgbahk/following{/other_user}", "gists_url": "https://api.github.com/users/cgbahk/gists{/gist_id}", "starred_url": "https://api.github.com/users/cgbahk/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/cgbahk/subscriptions", "organizations_url": "https://api.github.com/users/cgbahk/orgs", "repos_url": "https://api.github.com/users/cgbahk/repos", "events_url": "https://api.github.com/users/cgbahk/events{/privacy}", "received_events_url": "https://api.github.com/users/cgbahk/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "How are you launching your training script? If it's just with python (no distributed), the `Trainer` will use `DataParallel` which requires your batch size to be mulitiplied by the number of GPUs to work properly. I'm guessing that's why you see the \"instanteneous batch size\" at 4x what you put.\r\n\r\nThis is the only case it will happen (if you launch in distributed mode, the batch size per device will show up correctly) and is a good mean to track whether you are using Distributed training properly (you shouldn't use DataParallel as per PyTorch documentation) so you should launch your script with `torchrun` or `accelerate launch`.", "I just begin to try training with multiple GPUs :smile: And everybody gives warning on using DP, and recommends to use DDP over DP. Okay I'll try.\r\n\r\nBut that is out of this issue topic. So let's not talk about it anymore here.\r\n\r\n---\r\n\r\n> How are you launching your training script? If it's just with python (no distributed), the `Trainer` will use `DataParallel`\r\n\r\nYes this is the case I meant. This issue is about DP not DDP.\r\n\r\nI think in this communication, it is extremely important to use same terms for same concepts, especially about several 'batch' concepts.\r\n\r\nLet me use term\r\n- `batch size per update`: count of input that used for one model parameter update\r\n- `device`: in this case, let's fix this to GPU\r\n - And I think that is what term 'device' mean in training arg `per_device_train_batch_size` and log `Instanteneous batch size per device`.\r\n https://github.com/huggingface/transformers/blob/66fd3a8d626a32989f4569260db32785c6cbf42a/src/transformers/training_args.py#L193-L194\r\n- `batch size per device`: count of input source for each device(i.e. GPU) for one model parameter update iteration\r\n - Depending on documentations and communications, ambiguous terms used, like \"mini-batch\"([:link:](https://github.com/huggingface/transformers/blob/v4.30.2/docs/source/en/perf_train_gpu_many.mdx#L89) [:link:](https://github.com/huggingface/transformers/blob/v4.30.2/docs/source/en/perf_train_gpu_many.mdx#L95)) or \"sub mini-batch\" [:link:](https://www.telesens.co/2019/04/04/distributed-data-parallel-training-using-pytorch-on-aws/). So let's fix to this term for this issue communication.\r\n - I expect this is the same concept with training arg `per_device_train_batch_size` and log `Instanteneous batch size per device`\r\n\r\n\r\n> `DataParallel` which requires your batch size to be mulitiplied by the number of GPUs to work properly. I'm guessing that's why you see the \"instanteneous batch size\" at 4x what you put.\r\n\r\nIn your comment, 'batch size' seems to mean `batch size per update`. And yes, that is true, it should be 'GPU count' x `batch size per device`.\r\n\r\nBut the log `Instantaneous batch size per device` means `batch size per device`, not `batch size per update`. That is what I'm pointing out as a bug, which can lead user to misunderstanding.\r\n", "(I will only use DDP, so this issue is not anymore important for me. But if I'm someone who cares about the project, like maintainer, I would leave this open before the bug fixed. Any maintainers can close this issue if they want so :smile:)", "I made the PR linked above to clarify the logging a bit more. Let me know if it's better!" ]

1,687

NONE

null

### System Info - `transformers` version: 4.30.1 - Platform: Linux-4.18.0-240.el8.x86_64-x86_64-with-glibc2.28 - Python version: 3.10.11 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): 1.12.1 (False) - Tensorflow version (GPU?): not installed (NA) - Flax version (CPU?/GPU?/TPU?): not installed (NA) - Jax version: not installed - JaxLib version: not installed - Using GPU in script?: <fill in> - Using distributed or parallel set-up in script?: <fill in> Not sure why "PyTorch version (GPU?)" is False. I think this is because GPU not connected when I run this on report time. On actual training, GPU connected. I'm pretty sure my pytorch environment is with GPU support, like I can use same conda environment to train normal, single GPU training exploiting GPU resource. ``` $ conda list | grep pytorch pytorch 1.12.1 py3.10_cuda11.3_cudnn8.3.2_0 pytorch ``` ### Who can help? @sgugger ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction I've observed `Instantaneous batch size per device` in trainer log reported as `per_device_train_batch_size` x GPU count, reproducible in multiple cases. I can't give full reproduction detail, but pretty sure that scenario below can give idea of the situation. For example, I tried to train with 2 GPUs in DP sense(DP as described in [:link:](https://github.com/huggingface/transformers/blob/v4.30.1/docs/source/en/perf_train_gpu_many.mdx#data-parallelism)), with following TrainingArgument: ```py TrainingArgument( auto_find_batch_size=False, per_device_train_batch_size=1, ... ) ``` Then training log looks like this. Note on `Instantaneous batch size per device` value. I expected 1 from `per_device_train_batch_size` ``` ***** Running training ***** Num examples = ... Num Epochs = ... Instantaneous batch size per device = 2 Total train batch size (w. parallel, distributed & accumulation) = ... Gradient Accumulation steps = ... Total optimization steps = ... Number of trainable parameters = ... ... ``` (I've experienced some other logging bug, like `Total train batch size` especially when with `auto_find_batch_size=True` but let's only focus on batch size mismatch in this issue) I could check `Instantaneous batch size per device` reported as `per_device_train_batch_size` x GPU count happens again in other cases, like - 4 GPUs / `per_device_train_batch_size=128` -> `Instantaneous batch size per device = 512` This maybe - correct actual behavior but logging is not correct, or - actual bug, or - I may have misunderstanding about DP, in this case please blame me :smile: ### Expected behavior I expected - `Instantaneous batch size per device` reported as `per_device_train_batch_size` not - `Instantaneous batch size per device` reported as `per_device_train_batch_size` x GPU count

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24345/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24345/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24344

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24344/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24344/comments

https://api.github.com/repos/huggingface/transformers/issues/24344/events

https://github.com/huggingface/transformers/pull/24344

1,762,551,554

PR_kwDOCUB6oc5TSe6k

24,344

docs: add BentoML to awesome-transformers

{ "login": "aarnphm", "id": 29749331, "node_id": "MDQ6VXNlcjI5NzQ5MzMx", "avatar_url": "https://avatars.githubusercontent.com/u/29749331?v=4", "gravatar_id": "", "url": "https://api.github.com/users/aarnphm", "html_url": "https://github.com/aarnphm", "followers_url": "https://api.github.com/users/aarnphm/followers", "following_url": "https://api.github.com/users/aarnphm/following{/other_user}", "gists_url": "https://api.github.com/users/aarnphm/gists{/gist_id}", "starred_url": "https://api.github.com/users/aarnphm/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/aarnphm/subscriptions", "organizations_url": "https://api.github.com/users/aarnphm/orgs", "repos_url": "https://api.github.com/users/aarnphm/repos", "events_url": "https://api.github.com/users/aarnphm/events{/privacy}", "received_events_url": "https://api.github.com/users/aarnphm/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "cc @LysandreJik ", "I have updated the docs to the bottom of the page.", "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24344). All of your documentation changes will be reflected on that endpoint." ]

1,687

CONTRIBUTOR

null

# What does this PR do?   Kindly ask to add BentoML to the list of awesome projects that has transformers support cc @parano Signed-off-by: Aaron <[email protected]> ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? cc @sgugger Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24344/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24344/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24344", "html_url": "https://github.com/huggingface/transformers/pull/24344", "diff_url": "https://github.com/huggingface/transformers/pull/24344.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24344.patch", "merged_at": 1687191450000 }

https://api.github.com/repos/huggingface/transformers/issues/24343

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24343/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24343/comments

https://api.github.com/repos/huggingface/transformers/issues/24343/events

https://github.com/huggingface/transformers/issues/24343

1,762,545,472

I_kwDOCUB6oc5pDk9A

24,343

Enable non-causal mask (to enable MLM) for VisionEncoderDecoder models

{ "login": "metemadi", "id": 4220153, "node_id": "MDQ6VXNlcjQyMjAxNTM=", "avatar_url": "https://avatars.githubusercontent.com/u/4220153?v=4", "gravatar_id": "", "url": "https://api.github.com/users/metemadi", "html_url": "https://github.com/metemadi", "followers_url": "https://api.github.com/users/metemadi/followers", "following_url": "https://api.github.com/users/metemadi/following{/other_user}", "gists_url": "https://api.github.com/users/metemadi/gists{/gist_id}", "starred_url": "https://api.github.com/users/metemadi/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/metemadi/subscriptions", "organizations_url": "https://api.github.com/users/metemadi/orgs", "repos_url": "https://api.github.com/users/metemadi/repos", "events_url": "https://api.github.com/users/metemadi/events{/privacy}", "received_events_url": "https://api.github.com/users/metemadi/received_events", "type": "User", "site_admin": false }

[ { "id": 2648621985, "node_id": "MDU6TGFiZWwyNjQ4NjIxOTg1", "url": "https://api.github.com/repos/huggingface/transformers/labels/Feature%20request", "name": "Feature request", "color": "FBCA04", "default": false, "description": "Request for a new feature" } ]

closed

false

null

[]

[ "Hi @metemadi, thanks for opening this issue! \r\n\r\nThis sounds like an interesting project! I believe there's a few places that would need to be adapted in order to enable this properly, such as not forcing `add_cross_attention` to the decoder config and not shifting tokens (cc @ydshieh). The VisionEncoderDecoder model is not intended to be compatible with all encoder-decoder pairs or use cases. This isn't something we'll add to the library at the moment, but feel free to share a fork branch with an implementation here if you'd like!", "Thank you for the insanely fast reply - HuggingFace is amazing as always! This all makes sense. Thanks again.", "Sorry, I forgot to reply:\r\n\r\nThere is however `class VisionTextDualEncoderModel`. One checkpoint on the Hub is [clip-italian](https://huggingface.co/clip-italian/clip-italian). If you look the config file, it uses `BertForMaskedLM` and `clip_vision_model`.\r\n\r\nIt might be helpful, but some slight modification might be necessary if the goal is to do what have been done in the paper you mentioned.\r\n" ]

1,687

NONE

null

### Feature request Hello! The current (amazing!) VisionEncoderDecoder library supports text generation via a standard causal LM. Some recent work (linked [here](https://arxiv.org/abs/2306.07915)) has shown promise in having the text decoder be a MLM instead of a causal LM. I believe this is doable with the current VisionEncoderDecoder library by passing in [MASK] tokens for the decoder_input_ids and passing in the labels as usual, but this would still result in a causal mask. The code comment is as follows which makes me think this: ``` decoder_attention_mask (`torch.BoolTensor` of shape `(batch_size, target_sequence_length)`, *optional*): Default behavior: generate a tensor that ignores pad tokens in `decoder_input_ids`. Causal mask will also be used by default. ``` Is there a way to turn off causal masking to predict multiple text tokens at once using a VisionEncoderDecoder model? ### Motivation Masked language modeling on top of a Vision encoder appears to be a promising new approach for image captioning and pre-training of vision models according to [this recent work](https://arxiv.org/abs/2306.07915). ### Your contribution Thank you!

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24343/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24343/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24342

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24342/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24342/comments

https://api.github.com/repos/huggingface/transformers/issues/24342/events

https://github.com/huggingface/transformers/issues/24342

1,762,232,559

I_kwDOCUB6oc5pCYjv

24,342

Wrong pre-trained Whisper's BOS token?

{ "login": "tonywu71", "id": 28306721, "node_id": "MDQ6VXNlcjI4MzA2NzIx", "avatar_url": "https://avatars.githubusercontent.com/u/28306721?v=4", "gravatar_id": "", "url": "https://api.github.com/users/tonywu71", "html_url": "https://github.com/tonywu71", "followers_url": "https://api.github.com/users/tonywu71/followers", "following_url": "https://api.github.com/users/tonywu71/following{/other_user}", "gists_url": "https://api.github.com/users/tonywu71/gists{/gist_id}", "starred_url": "https://api.github.com/users/tonywu71/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/tonywu71/subscriptions", "organizations_url": "https://api.github.com/users/tonywu71/orgs", "repos_url": "https://api.github.com/users/tonywu71/repos", "events_url": "https://api.github.com/users/tonywu71/events{/privacy}", "received_events_url": "https://api.github.com/users/tonywu71/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi!\r\n\r\nThe first issue seems to be a feature of the whisper model. It has `<|endoftext|>` as token text for `bos`, `eos`, `pad` and `unk`. I see there are no dedicated tokens for `unk` and `pad`, so I think this is a feature of the model, and not a bug. If you look at the [original code](https://github.com/openai/whisper/blob/main/whisper/tokenizer.py), you can see that there is no dedicated token for `eos`, `bos`, `pad` or `unk`. This seems to indicate that these tokens are simply not used by the model.\r\n\r\nThe second issue is due to `add_special_tokens` being set to `True` by default. So this is not unexpected behavior.\r\n\r\n```python\r\ntokenizer.encode([\"<|startoftranscript|>\"], add_special_tokens=False)\r\n>>> [50258]\r\ntokenizer.decode([50258])\r\n>>> '<|startoftranscript|>'\r\n```\r\n\r\n\r\n", "cc @ArthurZucker ", "Hey, not entirely sure which part of the documentation you are referring to, but this is expected. The `bos_token` is not used to start a transcript. More details [here](https://huggingface.co/openai/whisper-base) about the starting tokens, and why we don't use this `bos`.", "> Hi!\r\n> \r\n> The first issue seems to be a feature of the whisper model. It has `<|endoftext|>` as token text for `bos`, `eos`, `pad` and `unk`. I see there are no dedicated tokens for `unk` and `pad`, so I think this is a feature of the model, and not a bug. If you look at the [original code](https://github.com/openai/whisper/blob/main/whisper/tokenizer.py), you can see that there is no dedicated token for `eos`, `bos`, `pad` or `unk`. This seems to indicate that these tokens are simply not used by the model.\r\n> \r\n> The second issue is due to `add_special_tokens` being set to `True` by default. So this is not unexpected behavior.\r\n> \r\n> ```python\r\n> tokenizer.encode([\"<|startoftranscript|>\"], add_special_tokens=False)\r\n> >>> [50258]\r\n> tokenizer.decode([50258])\r\n> >>> '<|startoftranscript|>'\r\n> ```\r\n\r\nThanks, makes sense!", "> Hey, not entirely sure which part of the documentation you are referring to, but this is expected. The `bos_token` is not used to start a transcript. More details [here](https://huggingface.co/openai/whisper-base) about the starting tokens, and why we don't use this `bos`.\r\n\r\nAccording to this [part](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperTokenizer.bos_token), we have:\r\n\r\n> bos_token (str, optional, defaults to \"<|startoftranscript|>\") — The beginning of sequence token.\r\n\r\nWhich in my opinion is a bit confusing.\r\n\r\nBut I do understand your point and how I should handle the `<|startoftranscript|>` now. Thanks for the help!", "I'll update the documentation to make it less confusing. The token used to store the ` \"<|startoftranscript|>\"` token is `decoder_start_token_id`. The `bos_token` is pretty much unused, which is why it was set to the same as `eos_token`. " ]

1,687

NONE

null

### System Info - `transformers` version: 4.30.2 - Platform: macOS-13.4-arm64-arm-64bit - Python version: 3.10.11 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): 2.0.0 (False) - Tensorflow version (GPU?): not installed (NA) - Flax version (CPU?/GPU?/TPU?): not installed (NA) - Jax version: not installed - JaxLib version: not installed - Using GPU in script?: <fill in> - Using distributed or parallel set-up in script?: <fill in> ### Who can help? @sanchit ### Information - [X] The official example scripts - [ ] My own modified scripts ### Tasks - [X] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction ```python from transformers import WhisperTokenizer WhisperTokenizer.from_pretrained("openai/whisper-tiny").bos_token >> '<|endoftext|>' ``` ### Expected behavior Dear Gandhi, From the [documentation](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperTokenizer) and from what I expect from the Whisper tokenizer, `processor.tokenizer.bos_token` should be equal to `"<|startoftranscript|>"` when using one of the official vanilla Whisper model. Currently, it is equal to `"<|endoftext|>"`. Is it an intended behavior? What do you think? On a different note, there is another weird behavior when encoding/decoding: ```python tokenizer.encode(["<|startoftranscript|>"]) >> [50258, 50363, 50258, 50257] processor.tokenizer.decode([50258, 50363, 50258, 50257]) >> '<|startoftranscript|><|notimestamps|><|startoftranscript|><|endoftext|>' ``` while I was expecting the last line to return `'<|startoftranscript|>'` only. Yours sincerely, Tony

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24342/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24342/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24341

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24341/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24341/comments

https://api.github.com/repos/huggingface/transformers/issues/24341/events

https://github.com/huggingface/transformers/issues/24341

1,762,223,297

I_kwDOCUB6oc5pCWTB

24,341

Colab Translation notebook link not found

{ "login": "SoyGema", "id": 24204714, "node_id": "MDQ6VXNlcjI0MjA0NzE0", "avatar_url": "https://avatars.githubusercontent.com/u/24204714?v=4", "gravatar_id": "", "url": "https://api.github.com/users/SoyGema", "html_url": "https://github.com/SoyGema", "followers_url": "https://api.github.com/users/SoyGema/followers", "following_url": "https://api.github.com/users/SoyGema/following{/other_user}", "gists_url": "https://api.github.com/users/SoyGema/gists{/gist_id}", "starred_url": "https://api.github.com/users/SoyGema/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/SoyGema/subscriptions", "organizations_url": "https://api.github.com/users/SoyGema/orgs", "repos_url": "https://api.github.com/users/SoyGema/repos", "events_url": "https://api.github.com/users/SoyGema/events{/privacy}", "received_events_url": "https://api.github.com/users/SoyGema/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "cc @Rocketknight1 ", "I opened a PR to the notebooks repo here to fix this: https://github.com/huggingface/notebooks/pull/398\r\n\r\nThanks for warning us about the issue - we appreciate the help to keep our docs up to date!" ]

1,687

CONTRIBUTOR

null

### System Info Hello There! First and foremost, congrats for Transformers Translation [tutorial](https://huggingface.co/docs/transformers/tasks/translation). 👍 It serves as a Spark for building english-to-many translation languages models! I´m following it along with TF mostly reproducing it in a jupyter Notebook with TF for mac with GPU enabled At the end of the [Train](https://huggingface.co/docs/transformers/tasks/translation) section , it is showed _For a more in-depth example of how to finetune a model for translation, take a look at the corresponding PyTorch notebook or [TensorFlow notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/translation-tf.ipynb)._ Inside the notebook, at cell [4] , there shows a message _**You** can find a script version of this notebook to fine-tune your model in a distributed fashion using multiple GPUs or TPUs [here](https://github.com/huggingface/transformers/tree/master/examples/seq2seq)._ The link is broken . ## Potential fix. Maybe it could point to Transformer [performance docs](https://huggingface.co/docs/transformers/performance) if you want to go for a more general overview or some specific part of [run_translation.py](https://github.com/huggingface/transformers/blob/main/examples/tensorflow/translation/run_translation.py) script facilitated by team member [here](https://github.com/huggingface/transformers/issues/24254#issuecomment-1594830054) during #24254 help? Please , don't hesitate to share the link as there could be a benefit in implementing it Thanks so much for the time dedicated to this Keep up the amazing work in the Open! ### Who can help? @Rocket ### Information - [X] The official example scripts - [ ] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction 1. Follow tutorial in docs . Go to Notebook at the end of Train Section 2. Go to Tensorflow Notebook 3. Click link in cell [4] . It seems to go to /seq2seq examples ### Expected behavior The link should point at a fine-tune script version of the notebook, or at least to docs

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24341/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24341/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24340

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24340/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24340/comments

https://api.github.com/repos/huggingface/transformers/issues/24340/events

https://github.com/huggingface/transformers/pull/24340

1,762,137,831

PR_kwDOCUB6oc5TRK2c

24,340

Fix TypeError: Object of type int64 is not JSON serializable

{ "login": "xiaoli", "id": 458922, "node_id": "MDQ6VXNlcjQ1ODkyMg==", "avatar_url": "https://avatars.githubusercontent.com/u/458922?v=4", "gravatar_id": "", "url": "https://api.github.com/users/xiaoli", "html_url": "https://github.com/xiaoli", "followers_url": "https://api.github.com/users/xiaoli/followers", "following_url": "https://api.github.com/users/xiaoli/following{/other_user}", "gists_url": "https://api.github.com/users/xiaoli/gists{/gist_id}", "starred_url": "https://api.github.com/users/xiaoli/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/xiaoli/subscriptions", "organizations_url": "https://api.github.com/users/xiaoli/orgs", "repos_url": "https://api.github.com/users/xiaoli/repos", "events_url": "https://api.github.com/users/xiaoli/events{/privacy}", "received_events_url": "https://api.github.com/users/xiaoli/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @xiaoli, thanks for opening this PR. \r\n\r\nCould you provide some more information about when the error occurs? Does this happen when running with the values from [the example readme](https://github.com/huggingface/transformers/tree/main/examples/pytorch/token-classification#pytorch-version-no-trainer)?", "Hi @amyeroberts, it happened on executing [./run_no_trainer.sh](https://github.com/huggingface/transformers/blob/main/examples/pytorch/token-classification/run_no_trainer.sh), and everything works smoothly but the last step of that saving results into JSON file.\r\n\r\nI got this error:\r\n`TypeError: Object of type int64 is not JSON serializable`, so this commit is trying to fix that.\r\n\r\nThis was happened on my Ubuntu 22.04 workstation.", "```sh\r\n(transformers) ➜ token-classification git:(main) ./run_no_trainer.sh && echo $(date +%d.%m.%y-%H:%M:%S)\r\nThe following values were not passed to `accelerate launch` and had defaults used instead:\r\n\t`--num_processes` was set to a value of `0`\r\n\t`--num_machines` was set to a value of `1`\r\n\t`--mixed_precision` was set to a value of `'no'`\r\n\t`--dynamo_backend` was set to a value of `'no'`\r\nTo avoid this warning pass in values for each of the problematic parameters or run `accelerate config`.\r\n06/20/2023 10:54:40 - INFO - __main__ - Distributed environment: DistributedType.NO\r\nNum processes: 1\r\nProcess index: 0\r\nLocal process index: 0\r\nDevice: mps\r\n\r\nMixed precision type: no\r\n\r\nDownloading builder script: 100%|████████████████████████████████████████████| 9.57k/9.57k [00:00<00:00, 8.80MB/s]\r\nDownloading metadata: 100%|██████████████████████████████████████████████████| 3.73k/3.73k [00:00<00:00, 9.41MB/s]\r\nDownloading readme: 100%|████████████████████████████████████████████████████| 12.3k/12.3k [00:00<00:00, 16.9MB/s]\r\nDownloading and preparing dataset conll2003/conll2003 to /Users/xiaoliwang/.cache/huggingface/datasets/conll2003/conll2003/1.0.0/9a4d16a94f8674ba3466315300359b0acd891b68b6c8743ddf60b9c702adce98...\r\nDownloading data: 100%|████████████████████████████████████████████████████████| 983k/983k [00:00<00:00, 3.57MB/s]\r\nGenerating train split: 0%| | 0/14041 [00:00<?, ? examples/s]06/20/2023 10:54:47 - INFO - datasets_modules.datasets.conll2003.9a4d16a94f8674ba3466315300359b0acd891b68b6c8743ddf60b9c702adce98.conll2003 - ⏳ Generating examples from = /Users/xiaoliwang/.cache/huggingface/datasets/downloads/extracted/31a52031f62b2a9281d3b6c2723006e2fa05b33157a4249729067b79f7aa068a/train.txt\r\nGenerating validation split: 0%| | 0/3250 [00:00<?, ? examples/s]06/20/2023 10:54:48 - INFO - datasets_modules.datasets.conll2003.9a4d16a94f8674ba3466315300359b0acd891b68b6c8743ddf60b9c702adce98.conll2003 - ⏳ Generating examples from = /Users/xiaoliwang/.cache/huggingface/datasets/downloads/extracted/31a52031f62b2a9281d3b6c2723006e2fa05b33157a4249729067b79f7aa068a/valid.txt\r\nGenerating test split: 0%| | 0/3453 [00:00<?, ? examples/s]06/20/2023 10:54:48 - INFO - datasets_modules.datasets.conll2003.9a4d16a94f8674ba3466315300359b0acd891b68b6c8743ddf60b9c702adce98.conll2003 - ⏳ Generating examples from = /Users/xiaoliwang/.cache/huggingface/datasets/downloads/extracted/31a52031f62b2a9281d3b6c2723006e2fa05b33157a4249729067b79f7aa068a/test.txt\r\nDataset conll2003 downloaded and prepared to /Users/xiaoliwang/.cache/huggingface/datasets/conll2003/conll2003/1.0.0/9a4d16a94f8674ba3466315300359b0acd891b68b6c8743ddf60b9c702adce98. Subsequent calls will reuse this data.\r\n100%|█████████████████████████████████████████████████████████████████████████████| 3/3 [00:00<00:00, 1282.14it/s]\r\nloading configuration file config.json from cache at /Users/xiaoliwang/.cache/huggingface/hub/models--bert-base-uncased/snapshots/a265f773a47193eed794233aa2a0f0bb6d3eaa63/config.json\r\nModel config BertConfig {\r\n \"_name_or_path\": \"bert-base-uncased\",\r\n \"architectures\": [\r\n \"BertForMaskedLM\"\r\n ],\r\n \"attention_probs_dropout_prob\": 0.1,\r\n \"classifier_dropout\": null,\r\n \"gradient_checkpointing\": false,\r\n \"hidden_act\": \"gelu\",\r\n \"hidden_dropout_prob\": 0.1,\r\n \"hidden_size\": 768,\r\n \"id2label\": {\r\n \"0\": \"LABEL_0\",\r\n \"1\": \"LABEL_1\",\r\n \"2\": \"LABEL_2\",\r\n \"3\": \"LABEL_3\",\r\n \"4\": \"LABEL_4\",\r\n \"5\": \"LABEL_5\",\r\n \"6\": \"LABEL_6\",\r\n \"7\": \"LABEL_7\",\r\n \"8\": \"LABEL_8\"\r\n },\r\n \"initializer_range\": 0.02,\r\n \"intermediate_size\": 3072,\r\n \"label2id\": {\r\n \"LABEL_0\": 0,\r\n \"LABEL_1\": 1,\r\n \"LABEL_2\": 2,\r\n \"LABEL_3\": 3,\r\n \"LABEL_4\": 4,\r\n \"LABEL_5\": 5,\r\n \"LABEL_6\": 6,\r\n \"LABEL_7\": 7,\r\n \"LABEL_8\": 8\r\n },\r\n \"layer_norm_eps\": 1e-12,\r\n \"max_position_embeddings\": 512,\r\n \"model_type\": \"bert\",\r\n \"num_attention_heads\": 12,\r\n \"num_hidden_layers\": 12,\r\n \"pad_token_id\": 0,\r\n \"position_embedding_type\": \"absolute\",\r\n \"transformers_version\": \"4.31.0.dev0\",\r\n \"type_vocab_size\": 2,\r\n \"use_cache\": true,\r\n \"vocab_size\": 30522\r\n}\r\n\r\nloading configuration file config.json from cache at /Users/xiaoliwang/.cache/huggingface/hub/models--bert-base-uncased/snapshots/a265f773a47193eed794233aa2a0f0bb6d3eaa63/config.json\r\nModel config BertConfig {\r\n \"_name_or_path\": \"bert-base-uncased\",\r\n \"architectures\": [\r\n \"BertForMaskedLM\"\r\n ],\r\n \"attention_probs_dropout_prob\": 0.1,\r\n \"classifier_dropout\": null,\r\n \"gradient_checkpointing\": false,\r\n \"hidden_act\": \"gelu\",\r\n \"hidden_dropout_prob\": 0.1,\r\n \"hidden_size\": 768,\r\n \"initializer_range\": 0.02,\r\n \"intermediate_size\": 3072,\r\n \"layer_norm_eps\": 1e-12,\r\n \"max_position_embeddings\": 512,\r\n \"model_type\": \"bert\",\r\n \"num_attention_heads\": 12,\r\n \"num_hidden_layers\": 12,\r\n \"pad_token_id\": 0,\r\n \"position_embedding_type\": \"absolute\",\r\n \"transformers_version\": \"4.31.0.dev0\",\r\n \"type_vocab_size\": 2,\r\n \"use_cache\": true,\r\n \"vocab_size\": 30522\r\n}\r\n\r\nloading file vocab.txt from cache at /Users/xiaoliwang/.cache/huggingface/hub/models--bert-base-uncased/snapshots/a265f773a47193eed794233aa2a0f0bb6d3eaa63/vocab.txt\r\nloading file tokenizer.json from cache at /Users/xiaoliwang/.cache/huggingface/hub/models--bert-base-uncased/snapshots/a265f773a47193eed794233aa2a0f0bb6d3eaa63/tokenizer.json\r\nloading file added_tokens.json from cache at None\r\nloading file special_tokens_map.json from cache at None\r\nloading file tokenizer_config.json from cache at /Users/xiaoliwang/.cache/huggingface/hub/models--bert-base-uncased/snapshots/a265f773a47193eed794233aa2a0f0bb6d3eaa63/tokenizer_config.json\r\nloading configuration file config.json from cache at /Users/xiaoliwang/.cache/huggingface/hub/models--bert-base-uncased/snapshots/a265f773a47193eed794233aa2a0f0bb6d3eaa63/config.json\r\nModel config BertConfig {\r\n \"_name_or_path\": \"bert-base-uncased\",\r\n \"architectures\": [\r\n \"BertForMaskedLM\"\r\n ],\r\n \"attention_probs_dropout_prob\": 0.1,\r\n \"classifier_dropout\": null,\r\n \"gradient_checkpointing\": false,\r\n \"hidden_act\": \"gelu\",\r\n \"hidden_dropout_prob\": 0.1,\r\n \"hidden_size\": 768,\r\n \"initializer_range\": 0.02,\r\n \"intermediate_size\": 3072,\r\n \"layer_norm_eps\": 1e-12,\r\n \"max_position_embeddings\": 512,\r\n \"model_type\": \"bert\",\r\n \"num_attention_heads\": 12,\r\n \"num_hidden_layers\": 12,\r\n \"pad_token_id\": 0,\r\n \"position_embedding_type\": \"absolute\",\r\n \"transformers_version\": \"4.31.0.dev0\",\r\n \"type_vocab_size\": 2,\r\n \"use_cache\": true,\r\n \"vocab_size\": 30522\r\n}\r\n\r\nDownloading model.safetensors: 100%|███████████████████████████████████████████| 440M/440M [00:22<00:00, 19.8MB/s]\r\nloading weights file model.safetensors from cache at /Users/xiaoliwang/.cache/huggingface/hub/models--bert-base-uncased/snapshots/a265f773a47193eed794233aa2a0f0bb6d3eaa63/model.safetensors\r\nSome weights of the model checkpoint at bert-base-uncased were not used when initializing BertForTokenClassification: ['cls.predictions.transform.LayerNorm.bias', 'cls.seq_relationship.weight', 'cls.seq_relationship.bias', 'cls.predictions.bias', 'cls.predictions.transform.dense.weight', 'cls.predictions.transform.dense.bias', 'cls.predictions.transform.LayerNorm.weight']\r\n- This IS expected if you are initializing BertForTokenClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\r\n- This IS NOT expected if you are initializing BertForTokenClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\r\nSome weights of BertForTokenClassification were not initialized from the model checkpoint at bert-base-uncased and are newly initialized: ['classifier.weight', 'classifier.bias']\r\nYou should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\r\n06/20/2023 10:55:15 - INFO - __main__ - Sample 622 of the training set: {'input_ids': [101, 2522, 6657, 15222, 6962, 1015, 19739, 20486, 2072, 1014, 102, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'labels': [-100, 3, -100, -100, -100, 0, 3, -100, -100, 0, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100]}.\r\n06/20/2023 10:55:15 - INFO - __main__ - Sample 12142 of the training set: {'input_ids': [101, 2019, 26354, 4861, 2056, 2008, 9779, 9048, 2015, 1010, 2007, 2095, 1011, 2203, 2727, 7045, 1997, 2149, 1002, 2184, 1012, 1023, 2454, 1998, 10067, 1997, 1002, 2184, 1012, 1019, 2454, 1010, 2052, 2022, 3205, 2006, 1996, 5548, 4518, 3863, 1010, 2021, 2106, 2025, 2360, 2043, 1012, 102, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'labels': [-100, 0, 3, 0, 0, 0, 3, -100, -100, 0, 0, 0, -100, -100, 0, 0, 0, 7, -100, 0, -100, -100, 0, 0, 0, 0, 0, 0, -100, -100, 0, 0, 0, 0, 0, 0, 0, 5, 0, 0, 0, 0, 0, 0, 0, 0, 0, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100]}.\r\n06/20/2023 10:55:15 - INFO - __main__ - Sample 4570 of the training set: {'input_ids': [101, 2117, 2679, 102, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'attention_mask': [1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'labels': [-100, 0, 0, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100]}.\r\nDownloading builder script: 100%|████████████████████████████████████████████| 6.34k/6.34k [00:00<00:00, 9.02MB/s]\r\n06/20/2023 10:55:18 - INFO - __main__ - ***** Running training *****\r\n06/20/2023 10:55:18 - INFO - __main__ - Num examples = 14041\r\n06/20/2023 10:55:18 - INFO - __main__ - Num Epochs = 3\r\n06/20/2023 10:55:18 - INFO - __main__ - Instantaneous batch size per device = 8\r\n06/20/2023 10:55:18 - INFO - __main__ - Total train batch size (w. parallel, distributed & accumulation) = 8\r\n06/20/2023 10:55:18 - INFO - __main__ - Gradient Accumulation steps = 1\r\n06/20/2023 10:55:18 - INFO - __main__ - Total optimization steps = 5268\r\n 33%|███████████████████████▋ | 1756/5268 [24:08<1:29:30, 1.53s/it]epoch 0: {'LOC_precision': 0.9499192245557351, 'LOC_recall': 0.9602612955906369, 'LOC_f1': 0.9550622631293991, 'LOC_number': 1837, 'MISC_precision': 0.8572972972972973, 'MISC_recall': 0.8600867678958786, 'MISC_f1': 0.858689767190038, 'MISC_number': 922, 'ORG_precision': 0.8539482879105521, 'ORG_recall': 0.9112602535421327, 'ORG_f1': 0.8816738816738816, 'ORG_number': 1341, 'PER_precision': 0.9776810016330975, 'PER_recall': 0.9766177270255574, 'PER_f1': 0.9771490750816105, 'PER_number': 1839, 'overall_precision': 0.9214876033057852, 'overall_recall': 0.9387102205758545, 'overall_f1': 0.9300191842522312, 'overall_accuracy': 0.9868336482091035}\r\n 67%|████████████████████████████████████████████████▋ | 3512/5268 [50:27<18:04, 1.62it/s]epoch 1: {'LOC_precision': 0.9637760702524698, 'LOC_recall': 0.9559063690800218, 'LOC_f1': 0.9598250888220825, 'LOC_number': 1837, 'MISC_precision': 0.8524251805985552, 'MISC_recall': 0.89587852494577, 'MISC_f1': 0.8736118455843469, 'MISC_number': 922, 'ORG_precision': 0.892675852066715, 'ORG_recall': 0.9179716629381058, 'ORG_f1': 0.9051470588235293, 'ORG_number': 1341, 'PER_precision': 0.9721925133689839, 'PER_recall': 0.9885807504078303, 'PER_f1': 0.9803181450525748, 'PER_number': 1839, 'overall_precision': 0.9322847682119205, 'overall_recall': 0.9481394174103385, 'overall_f1': 0.940145254194841, 'overall_accuracy': 0.9880217361665661}\r\n100%|███████████████████████████████████████████████████████████████████████| 5268/5268 [1:15:39<00:00, 1.44it/s]epoch 2: {'LOC_precision': 0.9538378958668814, 'LOC_recall': 0.9673380511703865, 'LOC_f1': 0.9605405405405405, 'LOC_number': 1837, 'MISC_precision': 0.8783351120597652, 'MISC_recall': 0.8926247288503254, 'MISC_f1': 0.8854222700376547, 'MISC_number': 922, 'ORG_precision': 0.9074759437453738, 'ORG_recall': 0.9142431021625652, 'ORG_f1': 0.9108469539375927, 'ORG_number': 1341, 'PER_precision': 0.9751619870410367, 'PER_recall': 0.9820554649265906, 'PER_f1': 0.978596586290978, 'PER_number': 1839, 'overall_precision': 0.9381975678827253, 'overall_recall': 0.94830779592524, 'overall_f1': 0.9432255903533747, 'overall_accuracy': 0.9891513935687436}\r\nConfiguration saved in /tmp/test-ner/config.json\r\nModel weights saved in /tmp/test-ner/pytorch_model.bin\r\ntokenizer config file saved in /tmp/test-ner/tokenizer_config.json\r\nSpecial tokens file saved in /tmp/test-ner/special_tokens_map.json\r\nTraceback (most recent call last):\r\n File \"/Users/xiaoliwang/repo/research/huggingface/transformers/examples/pytorch/token-classification/run_ner_no_trainer.py\", line 784, in <module>\r\n main()\r\n File \"/Users/xiaoliwang/repo/research/huggingface/transformers/examples/pytorch/token-classification/run_ner_no_trainer.py\", line 780, in main\r\n json.dump(all_results, f)\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/lib/python3.11/json/__init__.py\", line 179, in dump\r\n for chunk in iterable:\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/lib/python3.11/json/encoder.py\", line 432, in _iterencode\r\n yield from _iterencode_dict(o, _current_indent_level)\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/lib/python3.11/json/encoder.py\", line 406, in _iterencode_dict\r\n yield from chunks\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/lib/python3.11/json/encoder.py\", line 439, in _iterencode\r\n o = _default(o)\r\n ^^^^^^^^^^^\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/lib/python3.11/json/encoder.py\", line 180, in default\r\n raise TypeError(f'Object of type {o.__class__.__name__} '\r\nTypeError: Object of type int64 is not JSON serializable\r\n100%|███████████████████████████████████████████████████████████████████████| 5268/5268 [1:17:11<00:00, 1.14it/s]\r\nTraceback (most recent call last):\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/bin/accelerate\", line 8, in <module>\r\n sys.exit(main())\r\n ^^^^^^\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/lib/python3.11/site-packages/accelerate/commands/accelerate_cli.py\", line 45, in main\r\n args.func(args)\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/lib/python3.11/site-packages/accelerate/commands/launch.py\", line 969, in launch_command\r\n simple_launcher(args)\r\n File \"/Users/xiaoliwang/development/miniforge3/envs/transformers/lib/python3.11/site-packages/accelerate/commands/launch.py\", line 625, in simple_launcher\r\n raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)\r\nsubprocess.CalledProcessError: Command '['/Users/xiaoliwang/development/miniforge3/envs/transformers/bin/python3.11', 'run_ner_no_trainer.py', '--model_name_or_path', 'bert-base-uncased', '--dataset_name', 'conll2003', '--output_dir', '/tmp/test-ner', '--pad_to_max_length', '--task_name', 'ner', '--return_entity_level_metrics']' returned non-zero exit status 1.\r\n```\r\n\r\nI have reproduced this on my Macbook Air M1 with mps accleration enabled. The full error messages have been posted above here, same as on my Ubuntu workstation.", "@amyeroberts Thanks for your comments!\r\n\r\nI think your idea is good, and I understand that your intention is obviously to avoid that `int` convertment of everything. \r\n\r\nBut according to this page https://docs.python.org/3/library/json.html\r\n```\r\nIf specified, default should be a function that gets called for objects that can’t otherwise be serialized. \r\nIt should return a JSON encodable version of the object or raise a [TypeError](https://docs.python.org/3/library/exceptions.html#TypeError). \r\nIf not specified, [TypeError](https://docs.python.org/3/library/exceptions.html#TypeError) is raised.\r\n```\r\n\r\nFrom my understanding, this `default` parameter is just likely giving a new converter function, and in this case that function is a concise `int()`, yes, that's it. I think we don't need to write a new handler function to handling all different object types here, because we only cannot handle/serialize the `np.int64` here.\r\nSo in the future if we have something more than that, I could definitely to write a new hanlder to take good care of them, hence for the time being, I think `default=int` is a good enough solution :)", "Hi @amyeroberts, I have changed that a little bit as you mentioned before :)", "_The documentation is not available anymore as the PR was closed or merged._", "@xiaoli For the quality CI checks, you'll need to run `make style` at the top level of the repo and push any changes that are applied. Once this is done, CI should all be green and branch good to merge in 👍 ", "> @xiaoli For the quality CI checks, you'll need to run `make style` at the top level of the repo and push any changes that are applied. Once this is done, CI should all be green and branch good to merge in 👍\r\n\r\n@amyeroberts Thanks for intructions, but I am afraid that so many files being changed after `make style` execution:\r\n\r\n```\r\n(transformers) ➜ transformers git:(main) ✗ git status\r\nOn branch main\r\nYour branch is ahead of 'origin/main' by 8 commits.\r\n (use \"git push\" to publish your local commits)\r\n\r\nChanges not staged for commit:\r\n (use \"git add <file>...\" to update what will be committed)\r\n (use \"git restore <file>...\" to discard changes in working directory)\r\n\tmodified: examples/research_projects/codeparrot/scripts/human_eval.py\r\n\tmodified: examples/research_projects/fsner/src/fsner/tokenizer_utils.py\r\n\tmodified: examples/research_projects/jax-projects/big_bird/prepare_natural_questions.py\r\n\tmodified: examples/research_projects/luke/run_luke_ner_no_trainer.py\r\n\tmodified: examples/research_projects/lxmert/modeling_frcnn.py\r\n\tmodified: examples/research_projects/visual_bert/modeling_frcnn.py\r\n\tmodified: src/transformers/generation/logits_process.py\r\n\tmodified: src/transformers/generation/tf_logits_process.py\r\n\tmodified: src/transformers/generation/tf_utils.py\r\n\tmodified: src/transformers/keras_callbacks.py\r\n\tmodified: src/transformers/models/bert/convert_bert_pytorch_checkpoint_to_original_tf.py\r\n\tmodified: src/transformers/models/bigbird_pegasus/convert_bigbird_pegasus_tf_to_pytorch.py\r\n\tmodified: src/transformers/models/deta/modeling_deta.py\r\n\tmodified: src/transformers/models/dpr/tokenization_dpr.py\r\n\tmodified: src/transformers/models/dpr/tokenization_dpr_fast.py\r\n\tmodified: src/transformers/models/pegasus/convert_pegasus_tf_to_pytorch.py\r\n\tmodified: src/transformers/models/sam/processing_sam.py\r\n\tmodified: tests/generation/test_framework_agnostic.py\r\n\tmodified: tests/models/codegen/test_modeling_codegen.py\r\n\tmodified: tests/models/data2vec/test_modeling_data2vec_audio.py\r\n\tmodified: tests/models/encodec/test_modeling_encodec.py\r\n\tmodified: tests/models/gpt2/test_modeling_gpt2.py\r\n\tmodified: tests/models/gptj/test_modeling_gptj.py\r\n\tmodified: tests/models/hubert/test_modeling_hubert.py\r\n\tmodified: tests/models/mctct/test_modeling_mctct.py\r\n\tmodified: tests/models/rwkv/test_modeling_rwkv.py\r\n\tmodified: tests/models/sew/test_modeling_sew.py\r\n\tmodified: tests/models/sew_d/test_modeling_sew_d.py\r\n\tmodified: tests/models/speecht5/test_modeling_speecht5.py\r\n\tmodified: tests/models/unispeech/test_modeling_unispeech.py\r\n\tmodified: tests/models/unispeech_sat/test_modeling_unispeech_sat.py\r\n\tmodified: tests/models/wav2vec2/test_modeling_flax_wav2vec2.py\r\n\tmodified: tests/models/wav2vec2/test_modeling_wav2vec2.py\r\n\tmodified: tests/models/wav2vec2_conformer/test_modeling_wav2vec2_conformer.py\r\n\tmodified: tests/models/wavlm/test_modeling_wavlm.py\r\n\tmodified: tests/models/whisper/test_modeling_whisper.py\r\n\tmodified: tests/onnx/test_onnx.py\r\n\tmodified: tests/test_modeling_tf_common.py\r\n\tmodified: tests/test_tokenization_common.py\r\n\tmodified: tests/trainer/test_trainer_seq2seq.py\r\n\tmodified: utils/check_copies.py\r\n\tmodified: utils/create_dummy_models.py\r\n\tmodified: utils/tests_fetcher.py\r\n\r\nno changes added to commit (use \"git add\" and/or \"git commit -a\")\r\n```", "@amyeroberts `make style` changes are committed, thank you 😁" ]

1,687

CONTRIBUTOR

null

# What does this PR do? Fixed that "TypeError: Object of type int64 is not JSON serializable" ## Who can review? Anyone in the community is free to review the PR once the tests have passed.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24340/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24340/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24340", "html_url": "https://github.com/huggingface/transformers/pull/24340", "diff_url": "https://github.com/huggingface/transformers/pull/24340.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24340.patch", "merged_at": 1687864549000 }

https://api.github.com/repos/huggingface/transformers/issues/24339

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24339/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24339/comments

https://api.github.com/repos/huggingface/transformers/issues/24339/events

https://github.com/huggingface/transformers/issues/24339

1,762,099,274

I_kwDOCUB6oc5pB4BK

24,339

feat: `agent.run(return_agent_types=True)`

{ "login": "aarnphm", "id": 29749331, "node_id": "MDQ6VXNlcjI5NzQ5MzMx", "avatar_url": "https://avatars.githubusercontent.com/u/29749331?v=4", "gravatar_id": "", "url": "https://api.github.com/users/aarnphm", "html_url": "https://github.com/aarnphm", "followers_url": "https://api.github.com/users/aarnphm/followers", "following_url": "https://api.github.com/users/aarnphm/following{/other_user}", "gists_url": "https://api.github.com/users/aarnphm/gists{/gist_id}", "starred_url": "https://api.github.com/users/aarnphm/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/aarnphm/subscriptions", "organizations_url": "https://api.github.com/users/aarnphm/orgs", "repos_url": "https://api.github.com/users/aarnphm/repos", "events_url": "https://api.github.com/users/aarnphm/events{/privacy}", "received_events_url": "https://api.github.com/users/aarnphm/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hey @aarnphm, could you provide a code sample with the return you'd like to receive so that I can play with it and see if it makes sense to implement it? Thanks!", "For example, I'm currently building [OpenLLM](https://github.com/bentoml/OpenLLM) and came across a use case where one can define an agent to generate an image and then caption it via a pipeline using [BentoML Runner](https://docs.bentoml.org/en/latest/concepts/runner.html#what-is-runner)\r\n\r\nOpenLLM also provides support for HuggingFace Agent where users will can switch between The inference endpoint or hosting their own starcoder.\r\n\r\nGiven the following segment to save a `captioning` pipeline\r\n\r\n```python\r\nimport bentoml, transformers\r\n\r\nbentoml.transformers.save_model(\"captioning\", pipeline('image-captioning'))\r\n```\r\n\r\nRunner is distributed by nature, and it can be defined in a service.py like so:\r\n\r\n```python\r\nimport bentoml\r\nimport transformers\r\n\r\ncaptioning_runner = bentoml.transformers.get(\"captioning\").to_runner()\r\n\r\nagent = transformers.HfAgent(\"http://283.23.22.1:3000/hf/agent\") # `openllm start starcoder`\r\n\r\nservice = bentoml.Service(\"agent-with-runners\", runners=[captioning_runner])\r\n\r\ndef preprocess(input_tensor: torch.Tensor) -> torch.Tensor:\r\n\t... \r\n\r\[email protected](input=bentoml.io.Text(), output=bentoml.io.Text())\r\nasync def transcribe_audio_to_french(prompt: str):\r\n\timage_output: ImageAgentType = agent.run(prompt, ..., return_agent_types=True)\r\n\t# then I do some preprocess with this tensor\r\n\tinput_for_pipeline = preprocess(image_output.to_raw())\r\n\treturn await async captioning_runner.async_run(input_for_pipeline)\r\n```\r\n\r\nYou can run this with `bentoml serve service.py:svc`\r\n\r\nThis can be one use case of the AgentType that can be useful here, where one can access the tensor directly without having to convert from PIL.Image output (which is currently what the agent.run returns if it returns an image if I understand it correctly)\r\n\r\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

CONTRIBUTOR

null

### Feature request Currently, `agent.run` on main will run materializer from `AgentType` to return its corresponding type. I think it would be a great addition to just return this `AgentType` directly for external libraries to build on top of! ```python agent = transformers.HfAgent("inference-api-endpoint") res: AgentType = agent.run(..., return_agent_types=True) ``` ### Motivation I'm currently playing around with the new agent API, and found that in cases where I don't want to return the decoded outputs immediately, it would be nice to get `AgentType` and manage the materialize myself. ### Your contribution I can help to create PR, but I know that the Agent API are still very experimental and unstable cc @LysandreJik on this

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24339/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24339/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24338

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24338/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24338/comments

https://api.github.com/repos/huggingface/transformers/issues/24338/events

https://github.com/huggingface/transformers/pull/24338

1,762,016,390

PR_kwDOCUB6oc5TQx9A

24,338

Add SophiaG.

{ "login": "guilt", "id": 195178, "node_id": "MDQ6VXNlcjE5NTE3OA==", "avatar_url": "https://avatars.githubusercontent.com/u/195178?v=4", "gravatar_id": "", "url": "https://api.github.com/users/guilt", "html_url": "https://github.com/guilt", "followers_url": "https://api.github.com/users/guilt/followers", "following_url": "https://api.github.com/users/guilt/following{/other_user}", "gists_url": "https://api.github.com/users/guilt/gists{/gist_id}", "starred_url": "https://api.github.com/users/guilt/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/guilt/subscriptions", "organizations_url": "https://api.github.com/users/guilt/orgs", "repos_url": "https://api.github.com/users/guilt/repos", "events_url": "https://api.github.com/users/guilt/events{/privacy}", "received_events_url": "https://api.github.com/users/guilt/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "[Semi-related] also linking to Paper page on HF: https://huggingface.co/papers/2305.14342", "Thank you all for encouraging this. \r\n\r\nAs a first cut, I am working with the authors to see if this can be a PyPi package if authors agree. License is MIT, so hopefully we can get this out soon. [Sophia on PyPi](https://github.com/Liuhong99/Sophia/issues/29)", "As Younes said before, we won't merge this PR but can leave it for anyone who want to try this out: Transformers is a library of models, not optimizers (the optimizers inside the library are actually deprecated). Once there is a package supporting this optimizer we can add support for the Trainer like we did for the `bitsandbytes` optimizers.", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

# What does this PR do? This is a scratch PR showing how to test Sophia with Transformers. This is no way production ready, and certainly needs to look at licensing. But, this is helpful if someone needs to try this right away. I'm re-using **AdamW**'s `beta` values. Plus if you see carefully, there's an ugly hack where I'm using `eps` as `rho`. This is code directly copy-pasta-ed from: @Liuhong99 's [Sophia](https://github.com/Liuhong99/Sophia); I am putting it here so people can experiment with it and see how it compares to **AdamW**. If there is sufficient interest in adding this and it can be licensed, would be happy to work on it here. Anyone is free to take this and turn it into something of value. Please close this as necessary too. ## Before submitting This PR does none of the above. It is too early to do this, but if there is sufficient interest would be happy to go through this process. - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Models: - text models: @ArthurZucker and @younesbelkada Common: - trainer: @sgugger

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24338/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24338/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24338", "html_url": "https://github.com/huggingface/transformers/pull/24338", "diff_url": "https://github.com/huggingface/transformers/pull/24338.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24338.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24337

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24337/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24337/comments

https://api.github.com/repos/huggingface/transformers/issues/24337/events

https://github.com/huggingface/transformers/issues/24337

1,762,014,491

I_kwDOCUB6oc5pBjUb

24,337

past_key_values is not working as expected for falcon-7b

{ "login": "orgadhadas", "id": 27919205, "node_id": "MDQ6VXNlcjI3OTE5MjA1", "avatar_url": "https://avatars.githubusercontent.com/u/27919205?v=4", "gravatar_id": "", "url": "https://api.github.com/users/orgadhadas", "html_url": "https://github.com/orgadhadas", "followers_url": "https://api.github.com/users/orgadhadas/followers", "following_url": "https://api.github.com/users/orgadhadas/following{/other_user}", "gists_url": "https://api.github.com/users/orgadhadas/gists{/gist_id}", "starred_url": "https://api.github.com/users/orgadhadas/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/orgadhadas/subscriptions", "organizations_url": "https://api.github.com/users/orgadhadas/orgs", "repos_url": "https://api.github.com/users/orgadhadas/repos", "events_url": "https://api.github.com/users/orgadhadas/events{/privacy}", "received_events_url": "https://api.github.com/users/orgadhadas/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @orgadhadas \r\nThanks for the issue, I think the canonical way to use past key values is to set `use_cache=True` when calling `model.generate`. I think the remote code supports that argument as you can see here: https://huggingface.co/tiiuae/falcon-7b-instruct/blob/main/modelling_RW.py#L699 Can you share with us why you want to define a custom past key value mechanism?", "I need to use the model.forward method, to get access to the hidden states computed during inference (for research). I hope this answers the question.", "You have the `output_hidden_states` argument which should output the hidden states no? ", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,687

1,690

NONE

null

### System Info Hello, I've been trying to use past_key_values to speed up text generation, but it doesn't seem to work: instead of generating coherent text like is done when I'm not using past_key_values, it seems to generate the same token over and over again. I've been trying to search the web for usage guidelines and it seemed to me like I'm doing everything correctly, but maybe I'm missing something. Thank you! ### Who can help? @ArthurZucker @younesbelkada - I think you're the relevant people for this. ### Information - [ ] The official example scripts - [x] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [x] My own task or dataset (give details below) ### Reproduction ``` from transformers import AutoTokenizer, AutoModelForCausalLM import transformers import torch import pandas as pd import pickle from transformers import pipeline device = "cuda" if torch.cuda.is_available() else "cpu" model = "tiiuae/falcon-7b-instruct" tokenizer = AutoTokenizer.from_pretrained(model, padding_side="left") model = AutoModelForCausalLM.from_pretrained("tiiuae/falcon-7b-instruct", trust_remote_code=True).to(device) # WITH past_key_values def get_answer_from_model1(model, tokenizer, q, max_new_tokens=50): predicted_token_id = None prompt = q generated_text = "" n_new_tokens = 0 past_key_values=None while (predicted_token_id != tokenizer.eos_token_id) and (n_new_tokens < max_new_tokens): if predicted_token_id is not None: model_input = tokenizer(predicted_token, return_tensors='pt').to(device) else: model_input = tokenizer(prompt, return_tensors='pt').to(device) with torch.no_grad(): model_output = model(model_input['input_ids'], past_key_values=past_key_values) past_key_values = model_output['past_key_values'] logits = model_output['logits'] predicted_token_id = logits.argmax(-1)[0][-1] predicted_token = tokenizer.decode(predicted_token_id) if predicted_token_id != tokenizer.eos_token_id: prompt += predicted_token generated_text += predicted_token n_new_tokens += 1 return generated_text # WITHOUT past_key_values def get_answer_from_model2(model, tokenizer, q, max_new_tokens=50): predicted_token_id = None prompt = q generated_text = "" n_new_tokens = 0 past_key_values=None while (predicted_token_id != tokenizer.eos_token_id) and (n_new_tokens < max_new_tokens): model_input = tokenizer(prompt, return_tensors='pt').to(device) with torch.no_grad(): model_output = model(model_input['input_ids'], past_key_values=past_key_values) logits = model_output['logits'] predicted_token_id = logits.argmax(-1)[0][-1] predicted_token = tokenizer.decode(predicted_token_id) if predicted_token_id != tokenizer.eos_token_id: prompt += predicted_token generated_text += predicted_token n_new_tokens += 1 return generated_text q="hello" answer1 = get_answer_from_model1(model, tokenizer, q) print(answer1) answer2 = get_answer_from_model2(model, tokenizer, q) print(answer2) ``` ### Expected behavior answer1 and answer2 should be the same

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24337/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24337/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24336

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24336/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24336/comments

https://api.github.com/repos/huggingface/transformers/issues/24336/events

https://github.com/huggingface/transformers/pull/24336

1,761,997,298

PR_kwDOCUB6oc5TQuIN

24,336

Fix link to documentation in Install from Source

{ "login": "SoyGema", "id": 24204714, "node_id": "MDQ6VXNlcjI0MjA0NzE0", "avatar_url": "https://avatars.githubusercontent.com/u/24204714?v=4", "gravatar_id": "", "url": "https://api.github.com/users/SoyGema", "html_url": "https://github.com/SoyGema", "followers_url": "https://api.github.com/users/SoyGema/followers", "following_url": "https://api.github.com/users/SoyGema/following{/other_user}", "gists_url": "https://api.github.com/users/SoyGema/gists{/gist_id}", "starred_url": "https://api.github.com/users/SoyGema/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/SoyGema/subscriptions", "organizations_url": "https://api.github.com/users/SoyGema/orgs", "repos_url": "https://api.github.com/users/SoyGema/repos", "events_url": "https://api.github.com/users/SoyGema/events{/privacy}", "received_events_url": "https://api.github.com/users/SoyGema/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "@amyeroberts You are welcome! Thanks for creating Transformers library ! :)", "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

CONTRIBUTOR

null

Fix link to documentation _to install Transformers from Source_ .Probably the title changed at some point from 'Installing' to 'Install' and verbose message in utils broke. # What does this PR do?   Fixes link to install from source in verbose message inside _utils_ Context : found during exploration of translation tutorial script and work related to #24254 ## Before submitting - [X] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [X] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? @sgugger @stevhliu Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24336/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24336/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24336", "html_url": "https://github.com/huggingface/transformers/pull/24336", "diff_url": "https://github.com/huggingface/transformers/pull/24336.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24336.patch", "merged_at": 1687191176000 }

https://api.github.com/repos/huggingface/transformers/issues/24335

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24335/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24335/comments

https://api.github.com/repos/huggingface/transformers/issues/24335/events

https://github.com/huggingface/transformers/pull/24335

1,761,991,673

PR_kwDOCUB6oc5TQtBl

24,335

[Wav2Vec2 - MMS] Correct directly loading adapters weights

{ "login": "patrickvonplaten", "id": 23423619, "node_id": "MDQ6VXNlcjIzNDIzNjE5", "avatar_url": "https://avatars.githubusercontent.com/u/23423619?v=4", "gravatar_id": "", "url": "https://api.github.com/users/patrickvonplaten", "html_url": "https://github.com/patrickvonplaten", "followers_url": "https://api.github.com/users/patrickvonplaten/followers", "following_url": "https://api.github.com/users/patrickvonplaten/following{/other_user}", "gists_url": "https://api.github.com/users/patrickvonplaten/gists{/gist_id}", "starred_url": "https://api.github.com/users/patrickvonplaten/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/patrickvonplaten/subscriptions", "organizations_url": "https://api.github.com/users/patrickvonplaten/orgs", "repos_url": "https://api.github.com/users/patrickvonplaten/repos", "events_url": "https://api.github.com/users/patrickvonplaten/events{/privacy}", "received_events_url": "https://api.github.com/users/patrickvonplaten/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Sorry, I accidentally submitted the review without a saved comment. I realised in the `from_pretrained` call why you were using `pass`. I still think raising an exception would be good, as otherwise we can get silent behaviour. Would it be possible to reliably check if `load_adaptive_weights` should be implemented for a model? \r\n\r\np.s. ignoring the wandb diffs, as they're just from being out-of-date from main" ]

1,687

MEMBER

null

# What does this PR do? This PR corrects incorrect behavior when loading MMS with non-default adapter weights via `from_pretrained(...)`. The issue is explained well [here](https://github.com/huggingface/transformers/issues/24223#issuecomment-1595856093). In a nutshell, we cannot load specific weights in the init because these loaded weights are later overwritten again in `from_pretrained`. To solve this I propose to add a new generic ```py load_adaptive_weights() ``` call to `from_pretrained` that can be overridden by models that inherit from `PretrainedModel`. This both solves the issue #24223 and is also cleaner IMO since weights shouldn't be loaded when calling the `__init__` method of a model anyways really. It was weird before that: ```py model = Wav2Vec2ForCTC(config, target_lang="fra") ``` would try to load weights into the model. cc @sgugger @sanchit-gandhi @amyeroberts wdyt about the design? Happy to add some more tests if ok for you

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24335/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24335/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24335", "html_url": "https://github.com/huggingface/transformers/pull/24335", "diff_url": "https://github.com/huggingface/transformers/pull/24335.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24335.patch", "merged_at": 1687282792000 }

https://api.github.com/repos/huggingface/transformers/issues/24334

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24334/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24334/comments

https://api.github.com/repos/huggingface/transformers/issues/24334/events

https://github.com/huggingface/transformers/pull/24334

1,761,948,660

PR_kwDOCUB6oc5TQkd9

24,334

Generate: add SequenceBiasLogitsProcessor

{ "login": "gante", "id": 12240844, "node_id": "MDQ6VXNlcjEyMjQwODQ0", "avatar_url": "https://avatars.githubusercontent.com/u/12240844?v=4", "gravatar_id": "", "url": "https://api.github.com/users/gante", "html_url": "https://github.com/gante", "followers_url": "https://api.github.com/users/gante/followers", "following_url": "https://api.github.com/users/gante/following{/other_user}", "gists_url": "https://api.github.com/users/gante/gists{/gist_id}", "starred_url": "https://api.github.com/users/gante/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/gante/subscriptions", "organizations_url": "https://api.github.com/users/gante/orgs", "repos_url": "https://api.github.com/users/gante/repos", "events_url": "https://api.github.com/users/gante/events{/privacy}", "received_events_url": "https://api.github.com/users/gante/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "@sgugger @amyeroberts \r\n\r\nA further request for approval, as I've introduced a pattern that I'd like to repeat (assuming you're okay with it).\r\n\r\nIn the latest commit, you'll see:\r\n1. An example in the logit processor class;\r\n2. The new generation config field's docstring redirecting to the corresponding logit processor docs;\r\n\r\nThis allows me to write a clear explanation and example (or examples) for each configuration option, without creating a monster in the generation config docstring. The user gets high-level info in the generation config docstring, and details in each processor.\r\n\r\nThe examples are more useful if they are relative to common use cases, in this case different `.generate()` parameterization. However, the example is sitting in the logit processor class, and does not make **direct** reference to the class. The alternative, to create an example using the class directly, is not very desirable either -- very few people use the logit processor classes directly. \r\n\r\nLMK if you have suggestions and/or if you agree 🤗 \r\n", "@gante, the m4 eval code broke after this PR was merged:\r\n\r\n```\r\nstderr: File \"/mnt/nvme0/code/huggingface/m4-master-3/m4/evaluation/launch.py\", line 143, in <module>\r\nstderr: main(args)\r\nstderr: File \"/mnt/nvme0/code/huggingface/m4-master-3/m4/evaluation/launch.py\", line 97, in main\r\nstderr: score = evaluator(task, accelerator, model, args)\r\nstderr: File \"/mnt/nvme0/code/huggingface/m4-master-3/m4/evaluation/evaluators/in_contexter.py\", line 262, in in_contexter\r\nstderr: metric = task.add_batch_metric(metric, **kwargs)\r\nstderr: File \"/mnt/nvme0/code/huggingface/m4-master-3/m4/models/vgpt2/evaluation_open_ended_vqa_in_context_vgpt2.py\", line 338, in add_batch_metric\r\nstderr: generated_tokens = self.generate_tokens(**kwargs)\r\nstderr: File \"/mnt/nvme0/code/huggingface/m4-master-3/m4/models/vgpt2/evaluation_open_ended_vqa_in_context_vgpt2.py\", line 314, in generate_tokens\r\nstderr: generated_tokens = unwrapped_model.generate(\r\nstderr: File \"/home/stas/anaconda3/envs/py38-pt20/lib/python3.8/site-packages/torch/utils/_contextlib.py\", line 115, in decorate_context\r\nstderr: return func(*args, **kwargs)\r\nstderr: File \"/mnt/nvme0/code/huggingface/transformers-master/src/transformers/generation/utils.py\", line 1627, in generate\r\nstderr: return self.beam_search(\r\nstderr: File \"/mnt/nvme0/code/huggingface/transformers-master/src/transformers/generation/utils.py\", line 2951, in beam_search\r\nstderr: next_token_scores_processed = logits_processor(input_ids, next_token_scores)\r\nstderr: File \"/mnt/nvme0/code/huggingface/transformers-master/src/transformers/generation/logits_process.py\", line 92, in __call__\r\nstderr: scores = processor(input_ids, scores)\r\nstderr: File \"/mnt/nvme0/code/huggingface/transformers-master/src/transformers/generation/logits_process.py\", line 618, in __call__\r\nstderr: self._prepare_bias_variables(scores)\r\nstderr: File \"/mnt/nvme0/code/huggingface/transformers-master/src/transformers/generation/logits_process.py\", line 678, in _prepare_bias_variables\r\nstderr: raise ValueError(\r\nstderr: ValueError: Setting a bias on sequences that share a common token termination is not yet supported. Please open an issue if you see this error message (after checking that it doesn't already exist).\r\n```\r\n\r\nwhat should we change?", "@stas00 interesting, I thought no relevant use case would hit this issue. I will open a PR with a fix!\r\n\r\n(meanwhile, the solutions are either to a) downgrade transformers; or b) remove this exception if you're using the `bad_words_ids` generate argument, which should be fine)", "Thank you, Joao\r\n\r\nWe are going to port the m4-pretrained model into `transformers` shortly, so neither of these proposals is an option in the long run. But a PR with a fix is - it's not urgent urgent as we meanwhile can use the older transformers." ]

1,687

1,689

1,687

MEMBER

null

# What does this PR do? Closes #22168 As per [popular demand](https://github.com/huggingface/transformers/issues/22168#issuecomment-1477998997), adds a logits processor that applies a bias to certain sequences -- `SequenceBiasLogitsProcessor` This manipulation is a more general case of forbidding certain sequences --`NoBadWordsLogitsProcessor` corresponds to applying an infinite negative bias. As such, this PR makes `NoBadWordsLogitsProcessor` a subclass of the new processor. In the refactoring process, I've rewritten this class to a) be more readable (clear variable naming, comments, docstrings); and b) be faster (though some vectorization).

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24334/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24334/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24334", "html_url": "https://github.com/huggingface/transformers/pull/24334", "diff_url": "https://github.com/huggingface/transformers/pull/24334.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24334.patch", "merged_at": 1687342481000 }

https://api.github.com/repos/huggingface/transformers/issues/24333

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24333/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24333/comments

https://api.github.com/repos/huggingface/transformers/issues/24333/events

https://github.com/huggingface/transformers/pull/24333

1,761,923,401

PR_kwDOCUB6oc5TQfgb

24,333

Fix `KerasMetricCallback`: pass `generate_kwargs` even if `use_xla_generation` is False

{ "login": "Kripner", "id": 9218121, "node_id": "MDQ6VXNlcjkyMTgxMjE=", "avatar_url": "https://avatars.githubusercontent.com/u/9218121?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Kripner", "html_url": "https://github.com/Kripner", "followers_url": "https://api.github.com/users/Kripner/followers", "following_url": "https://api.github.com/users/Kripner/following{/other_user}", "gists_url": "https://api.github.com/users/Kripner/gists{/gist_id}", "starred_url": "https://api.github.com/users/Kripner/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Kripner/subscriptions", "organizations_url": "https://api.github.com/users/Kripner/orgs", "repos_url": "https://api.github.com/users/Kripner/repos", "events_url": "https://api.github.com/users/Kripner/events{/privacy}", "received_events_url": "https://api.github.com/users/Kripner/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,687

CONTRIBUTOR

null

# What does this PR do? Currently, `KerasMetricCallback` ignores the `generate_kwargs` argument if `use_xla_generation` is set to `False` (which is the default). This means that when not using XLA, the user can't pass arguments like `max_new_tokens` to the `generate` method being called in `on_epoch_end`. It's also in contradiction with the docstring for `generate_kwargs`, which states: > Keyword arguments to pass to `model.generate()` when generating. Has no effect if `predict_with_generate` is `False`. This PR fixes the issue by passing `generate_kwargs` to `model.generate()` in the branch of execution where `use_xla_generation` is `False`. ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. @Rocketknight1 @gante

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24333/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24333/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24333", "html_url": "https://github.com/huggingface/transformers/pull/24333", "diff_url": "https://github.com/huggingface/transformers/pull/24333.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24333.patch", "merged_at": 1687175485000 }

https://api.github.com/repos/huggingface/transformers/issues/24332

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24332/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24332/comments

https://api.github.com/repos/huggingface/transformers/issues/24332/events

https://github.com/huggingface/transformers/issues/24332

1,761,760,676

I_kwDOCUB6oc5pAlWk

24,332

Why it always raise error like this?

{ "login": "ElinLiu0", "id": 75596885, "node_id": "MDQ6VXNlcjc1NTk2ODg1", "avatar_url": "https://avatars.githubusercontent.com/u/75596885?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ElinLiu0", "html_url": "https://github.com/ElinLiu0", "followers_url": "https://api.github.com/users/ElinLiu0/followers", "following_url": "https://api.github.com/users/ElinLiu0/following{/other_user}", "gists_url": "https://api.github.com/users/ElinLiu0/gists{/gist_id}", "starred_url": "https://api.github.com/users/ElinLiu0/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ElinLiu0/subscriptions", "organizations_url": "https://api.github.com/users/ElinLiu0/orgs", "repos_url": "https://api.github.com/users/ElinLiu0/repos", "events_url": "https://api.github.com/users/ElinLiu0/events{/privacy}", "received_events_url": "https://api.github.com/users/ElinLiu0/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @RosterMouch, thanks for raising this issue. \r\n\r\nA few questions on our side to try and help dig into the issue: \r\n* Could you share which verion of `huggingface_hub` is beign run in your environment? \r\n* When you say \"I could promise that there is not a network connection problem at all\", could you share how this was tested?\r\n* Is this an error the consistently happens or sporadically? \r\n* Is this issue only ever seen with this checkpoint or with other checkpoints too? " ]

1,686

1,689

NONE

null

### System Info I could promise that there is not a network connection problem at all,but it true raised like this: ```bash Traceback (most recent call last): File "test.py", line 3, in <module> model = AutoModelForCausalLM.from_pretrained("nomic-ai/gpt4all-j", revision="v1.2-jazzy") File "/home/elin/anaconda3/envs/nemo/lib/python3.8/site-packages/transformers/models/auto/auto_factory.py", line 444, in from_pretrained config, kwargs = AutoConfig.from_pretrained( File "/home/elin/anaconda3/envs/nemo/lib/python3.8/site-packages/transformers/models/auto/configuration_auto.py", line 928, in from_pretrained config_dict, unused_kwargs = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs) File "/home/elin/anaconda3/envs/nemo/lib/python3.8/site-packages/transformers/configuration_utils.py", line 574, in get_config_dict config_dict, kwargs = cls._get_config_dict(pretrained_model_name_or_path, **kwargs) File "/home/elin/anaconda3/envs/nemo/lib/python3.8/site-packages/transformers/configuration_utils.py", line 629, in _get_config_dict resolved_config_file = cached_file( File "/home/elin/anaconda3/envs/nemo/lib/python3.8/site-packages/transformers/utils/hub.py", line 452, in cached_file raise EnvironmentError( OSError: We couldn't connect to 'https://huggingface.co' to load this file, couldn't find it in the cached files and it looks like nomic-ai/gpt4all-j is not the path to a directory containing a file named config.json. Checkout your internet connection or see how to run the library in offline mode at 'https://huggingface.co/docs/transformers/installation#offline-mode'. ``` And here is the SDKs version below: ``` Python 3.8.10 transformers 4.29.2 ``` ### Who can help? _No response_ ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction Not only gpt4-j,but also fallcon by using the code they provided: ```python from transformers import AutoTokenizer, AutoModelForCausalLM import transformers import torch model = "tiiuae/falcon-7b" tokenizer = AutoTokenizer.from_pretrained(model) pipeline = transformers.pipeline( "text-generation", model=model, tokenizer=tokenizer, torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto", ) sequences = pipeline( "Girafatron is obsessed with giraffes, the most glorious animal on the face of this Earth. Giraftron believes all other animals are irrelevant when compared to the glorious majesty of the giraffe.\nDaniel: Hello, Girafatron!\nGirafatron:", max_length=200, do_sample=True, top_k=10, num_return_sequences=1, eos_token_id=tokenizer.eos_token_id, ) for seq in sequences: print(f"Result: {seq['generated_text']}") ``` ### Expected behavior Fix it!

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24332/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24332/timeline

not_planned

null

https://api.github.com/repos/huggingface/transformers/issues/24331

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24331/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24331/comments

https://api.github.com/repos/huggingface/transformers/issues/24331/events

https://github.com/huggingface/transformers/pull/24331

1,761,638,661

PR_kwDOCUB6oc5TPqms

24,331

style: add BitsAndBytesConfig __repr__ function

{ "login": "aarnphm", "id": 29749331, "node_id": "MDQ6VXNlcjI5NzQ5MzMx", "avatar_url": "https://avatars.githubusercontent.com/u/29749331?v=4", "gravatar_id": "", "url": "https://api.github.com/users/aarnphm", "html_url": "https://github.com/aarnphm", "followers_url": "https://api.github.com/users/aarnphm/followers", "following_url": "https://api.github.com/users/aarnphm/following{/other_user}", "gists_url": "https://api.github.com/users/aarnphm/gists{/gist_id}", "starred_url": "https://api.github.com/users/aarnphm/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/aarnphm/subscriptions", "organizations_url": "https://api.github.com/users/aarnphm/orgs", "repos_url": "https://api.github.com/users/aarnphm/repos", "events_url": "https://api.github.com/users/aarnphm/events{/privacy}", "received_events_url": "https://api.github.com/users/aarnphm/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Hi @amyeroberts I have addressed all of the issue. PTAL when you are available. Thanks a bunch.", "Can you rebase your branch on main? This should fix the failing test. Thanks!", "done.", "Thanks a lot for working on this @aarnphm ! Nice job! " ]

1,686

1,687

CONTRIBUTOR

null

# What does this PR do? Add a `__repr__` to `transformers.BitsAndBytesConfig` to make it nice to print.   ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? cc @sgugger

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24331/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24331/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24331", "html_url": "https://github.com/huggingface/transformers/pull/24331", "diff_url": "https://github.com/huggingface/transformers/pull/24331.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24331.patch", "merged_at": 1687278368000 }

https://api.github.com/repos/huggingface/transformers/issues/24330

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24330/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24330/comments

https://api.github.com/repos/huggingface/transformers/issues/24330/events

https://github.com/huggingface/transformers/issues/24330

1,761,327,681

I_kwDOCUB6oc5o-7pB

24,330

Resuming / retraining the peft model

{ "login": "adityaaryan77", "id": 69278251, "node_id": "MDQ6VXNlcjY5Mjc4MjUx", "avatar_url": "https://avatars.githubusercontent.com/u/69278251?v=4", "gravatar_id": "", "url": "https://api.github.com/users/adityaaryan77", "html_url": "https://github.com/adityaaryan77", "followers_url": "https://api.github.com/users/adityaaryan77/followers", "following_url": "https://api.github.com/users/adityaaryan77/following{/other_user}", "gists_url": "https://api.github.com/users/adityaaryan77/gists{/gist_id}", "starred_url": "https://api.github.com/users/adityaaryan77/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/adityaaryan77/subscriptions", "organizations_url": "https://api.github.com/users/adityaaryan77/orgs", "repos_url": "https://api.github.com/users/adityaaryan77/repos", "events_url": "https://api.github.com/users/adityaaryan77/events{/privacy}", "received_events_url": "https://api.github.com/users/adityaaryan77/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @adityaaryan77, thanks for raising an issue! \r\n\r\nIt seems like this is a case of catastrophic forgetting, rather than a bug per se in the model or transformers library. As such question best placed in our [forums](https://discuss.huggingface.co/). We try to reserve the github issues for feature requests and bug reports.\r\n\r\nIf you believe this behaviour is related to a bug in the code, could you produce: \r\n* The running environment: run `transformers-cli env` in the terminal and copy-paste the output\r\n* A _minimal_ code reproducer which we could run to replicate i.e. with data ", "`transformers-cli env `\r\n- `transformers` version: 4.30.2\r\n- Platform: Linux-5.15.0-1040-azure-x86_64-with-glibc2.29\r\n- Python version: 3.8.10\r\n- Huggingface_hub version: 0.15.1\r\n- Safetensors version: 0.3.1\r\n- PyTorch version (GPU?): 2.0.1+cu117 (True)\r\n- Tensorflow version (GPU?): not installed (NA)\r\n- Flax version (CPU?/GPU?/TPU?): not installed (NA)\r\n- Jax version: not installed\r\n- JaxLib version: not installed\r\n- Using GPU in script?Yes\r\n- Using distributed or parallel set-up in script?: No\r\n\r\nSure here is the pastebin for the code snippet I am using : https://pastebin.pl/view/4e77a13d\r\n\r\nAnd for example for data set: Here is a small example\r\nFirst time fine tuning:\r\n[\r\n{\"question\":\"What is my cats name?\",\"answer\":\"Tom\"}\r\n]\r\nNow using generate with \"What is my cats name gives\" response as \"Tom\"\r\nNow saving this model and loading it with resume_from_checkpoint for further fine tuning with \r\n[\r\n{\"question\":\"What is my dogs name?\",\"answer\":\"Bob\"}\r\n]\r\nAnd asking \"What is my cats name?\" gives response as \"Bob\" or sometimes repeats the question\r\n", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,686

1,690

NONE

null

### System Info Although resume_from_checkpoint is now working after @llohann-speranca solved the issue but now finetuning again with new data using train(resume_from_checkpoint) and then testing it makes it forget the old datas i.e. wont remember the things in old dataset. Attaching the code below: import json import os import bitsandbytes as bnb import pandas as pd import torch import torch.nn as nn import transformers from datasets import load_dataset from peft import ( LoraConfig, PeftConfig, PeftModel, get_peft_model, prepare_model_for_kbit_training, ) from transformers import ( AutoConfig, AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig, ) os.environ["CUDA_VISIBLE_DEVICES"] = "0" MODEL_NAME = "tiiuae/falcon-7b" bnb_config = BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_use_double_quant=True, bnb_4bit_quant_type="nf4", bnb_4bit_compute_dtype=torch.bfloat16, ) model = AutoModelForCausalLM.from_pretrained( MODEL_NAME, device_map="auto", trust_remote_code=True, quantization_config=bnb_config ) tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME) tokenizer.pad_token = tokenizer.eos_token def print_trainable_parameters(model): """ Prints the number of trainable parameters in the model. """ trainable_params = 0 all_param = 0 for _, param in model.named_parameters(): all_param += param.numel() if param.requires_grad: trainable_params += param.numel() print( f"trainable params: {trainable_params} || all params: {all_param} || trainable%: {100 * trainable_params / all_param}" ) model.gradient_checkpointing_enable() model = prepare_model_for_kbit_training(model) config = LoraConfig( r=16, lora_alpha=32, target_modules=["query_key_value"], lora_dropout=0.05, bias="none", task_type="CAUSAL_LM" ) model = get_peft_model(model, config) print_trainable_parameters(model) data = load_dataset("json", data_files="../localGPT/output.json") def generate_prompt(data_point): return f""" : {data_point["question"]} : {data_point["answer"]} """.strip() def generate_and_tokenize_prompt(data_point): full_prompt = generate_prompt(data_point) tokenized_full_prompt = tokenizer(full_prompt, padding=True, truncation=True) return tokenized_full_prompt data = data["train"].shuffle().map(generate_and_tokenize_prompt) OUTPUT_DIR = "outputs" training_args = transformers.TrainingArguments( per_device_train_batch_size=1, gradient_accumulation_steps=4, num_train_epochs=1, warmup_ratio=0.05, max_steps=80, learning_rate=2e-4, fp16=True, logging_steps=1, save_total_limit=3, output_dir=OUTPUT_DIR, optim="paged_adamw_8bit", lr_scheduler_type="cosine", ) trainer = transformers.Trainer( model=model, train_dataset=data, args=training_args, data_collator=transformers.DataCollatorForLanguageModeling(tokenizer, mlm=False) ) model.config.use_cache = False trainer.train(resume_from_checkpoint=True) trainer.save_model(os.path.join(OUTPUT_DIR, "checkpoint-2")) PEFT_MODEL = OUTPUT_DIR+"/checkpoint-2" config = PeftConfig.from_pretrained(PEFT_MODEL) model = AutoModelForCausalLM.from_pretrained( config.base_model_name_or_path, return_dict=True, quantization_config=bnb_config, device_map="auto", trust_remote_code=True ) tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path) tokenizer.pad_token = tokenizer.eos_token model = PeftModel.from_pretrained(model, PEFT_MODEL) generation_config = model.generation_config generation_config.max_new_tokens = 20 generation_config.temperature = 0 generation_config.top_p = 0.7 generation_config.num_return_sequences = 1 generation_config.pad_token_id = tokenizer.eos_token_id generation_config.eos_token_id = tokenizer.eos_token_id DEVICE = "cuda:0" prompt = """ :What is my cat's name? : """.strip() encoding = tokenizer(prompt, return_tensors="pt").to(DEVICE) with torch.inference_mode(): outputs = model.generate( input_ids=encoding.input_ids, attention_mask=encoding.attention_mask, generation_config=generation_config, ) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ### Who can help? _No response_ ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction I trained my model for my cats name in first iteration and saved it in checkpoint-1 then retrained it for my dogs name although now it knows my dogs name it forgets my cats name ### Expected behavior To remember my cats name

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24330/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24330/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24329

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24329/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24329/comments

https://api.github.com/repos/huggingface/transformers/issues/24329/events

https://github.com/huggingface/transformers/pull/24329

1,761,220,653

PR_kwDOCUB6oc5TOPaM

24,329

[Doc Fix] Fix model name path in the transformers doc for AutoClasses

{ "login": "riteshghorse", "id": 25881114, "node_id": "MDQ6VXNlcjI1ODgxMTE0", "avatar_url": "https://avatars.githubusercontent.com/u/25881114?v=4", "gravatar_id": "", "url": "https://api.github.com/users/riteshghorse", "html_url": "https://github.com/riteshghorse", "followers_url": "https://api.github.com/users/riteshghorse/followers", "following_url": "https://api.github.com/users/riteshghorse/following{/other_user}", "gists_url": "https://api.github.com/users/riteshghorse/gists{/gist_id}", "starred_url": "https://api.github.com/users/riteshghorse/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/riteshghorse/subscriptions", "organizations_url": "https://api.github.com/users/riteshghorse/orgs", "repos_url": "https://api.github.com/users/riteshghorse/repos", "events_url": "https://api.github.com/users/riteshghorse/events{/privacy}", "received_events_url": "https://api.github.com/users/riteshghorse/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "R: @stevhliu", "_The documentation is not available anymore as the PR was closed or merged._" ]

1,686

1,687

CONTRIBUTOR

null

Fixes the model name path in the transformers doc for the AutoTokenizer step.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24329/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24329/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24329", "html_url": "https://github.com/huggingface/transformers/pull/24329", "diff_url": "https://github.com/huggingface/transformers/pull/24329.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24329.patch", "merged_at": 1687192015000 }

https://api.github.com/repos/huggingface/transformers/issues/24328

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24328/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24328/comments

https://api.github.com/repos/huggingface/transformers/issues/24328/events

https://github.com/huggingface/transformers/pull/24328

1,761,157,284

PR_kwDOCUB6oc5TOBMm

24,328

save full_osd in fsdp mode

{ "login": "amartino1", "id": 77289753, "node_id": "MDQ6VXNlcjc3Mjg5NzUz", "avatar_url": "https://avatars.githubusercontent.com/u/77289753?v=4", "gravatar_id": "", "url": "https://api.github.com/users/amartino1", "html_url": "https://github.com/amartino1", "followers_url": "https://api.github.com/users/amartino1/followers", "following_url": "https://api.github.com/users/amartino1/following{/other_user}", "gists_url": "https://api.github.com/users/amartino1/gists{/gist_id}", "starred_url": "https://api.github.com/users/amartino1/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/amartino1/subscriptions", "organizations_url": "https://api.github.com/users/amartino1/orgs", "repos_url": "https://api.github.com/users/amartino1/repos", "events_url": "https://api.github.com/users/amartino1/events{/privacy}", "received_events_url": "https://api.github.com/users/amartino1/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24328). All of your documentation changes will be reflected on that endpoint.", "Hello, PR #24446 addresses this issue. ", "This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.\n\nPlease note that issues that do not follow the [contributing guidelines](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md) are likely to be ignored." ]

1,686

1,690

NONE

null

Fixes a typo in which the variable full_osd is referenced before definition if run in fsdp mode. The fix allows model files to be saved when running in fsdp. Links to issue: https://github.com/huggingface/transformers/issues/24057 Fixes # 24057

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24328/reactions", "total_count": 1, "+1": 0, "-1": 0, "laugh": 0, "hooray": 1, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24328/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24328", "html_url": "https://github.com/huggingface/transformers/pull/24328", "diff_url": "https://github.com/huggingface/transformers/pull/24328.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24328.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24327

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24327/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24327/comments

https://api.github.com/repos/huggingface/transformers/issues/24327/events

https://github.com/huggingface/transformers/issues/24327

1,761,141,342

I_kwDOCUB6oc5o-OJe

24,327

AutoModelForSequenceClassification.from_config doesn't support LlamaConfig

{ "login": "kungfu-eric", "id": 87145506, "node_id": "MDQ6VXNlcjg3MTQ1NTA2", "avatar_url": "https://avatars.githubusercontent.com/u/87145506?v=4", "gravatar_id": "", "url": "https://api.github.com/users/kungfu-eric", "html_url": "https://github.com/kungfu-eric", "followers_url": "https://api.github.com/users/kungfu-eric/followers", "following_url": "https://api.github.com/users/kungfu-eric/following{/other_user}", "gists_url": "https://api.github.com/users/kungfu-eric/gists{/gist_id}", "starred_url": "https://api.github.com/users/kungfu-eric/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/kungfu-eric/subscriptions", "organizations_url": "https://api.github.com/users/kungfu-eric/orgs", "repos_url": "https://api.github.com/users/kungfu-eric/repos", "events_url": "https://api.github.com/users/kungfu-eric/events{/privacy}", "received_events_url": "https://api.github.com/users/kungfu-eric/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @kungfu-eric, \r\n\r\nThis issue is arising because you need to pass an instance of the config, rather than the config class i.e.:\r\n\r\n```python\r\nfrom transformers import AutoModelForSequenceClassification\r\nfrom transformers.models.llama.configuration_llama import LlamaConfig\r\n\r\nconfig = LlamaConfig()\r\nmodel = AutoModelForSequenceClassification.from_config(config)\r\n```", "What ended up fixing the issues was updating to transformers-4.30.2 from 4.28.1. @amyeroberts ah i wrote the simple example wrong. I did define the config in the actual full code. Thank you though." ]

1,686

NONE

null

### System Info Calling: ``` from transformers import AutoModelForSequenceClassification from transformers.models.llama.configuration_llama import LlamaConfig config = LlamaConfig() model = AutoModelForSequenceClassification.from_config(config) ``` gives: ``` ________________________________ Traceback (most recent call last) _________________________________ _ in <module>:1 _ _ _ _ /opt/conda/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py:413 in _ _ from_config _ _ _ _ 410 _ _ _ model_class = _get_model_class(config, cls._model_mapping) _ _ 411 _ _ _ return model_class._from_config(config, **kwargs) _ _ 412 _ _ _ _ _ 413 _ _ raise ValueError( _ _ 414 _ _ _ f"Unrecognized configuration class {config.__class__} for this kind of AutoM _ _ 415 _ _ _ f"Model type should be one of {', '.join(c.__name__ for c in cls._model_mapp _ _ 416 _ _ ) _ ____________________________________________________________________________________________________ ValueError: Unrecognized configuration class <class 'type'> for this kind of AutoModel: AutoModelForSequenceClassification. Model type should be one of AlbertConfig, BartConfig, BertConfig, BigBirdConfig, BigBirdPegasusConfig, BloomConfig, CamembertConfig, CanineConfig, ConvBertConfig, CTRLConfig, Data2VecTextConfig, DebertaConfig, DebertaV2Config, DistilBertConfig, ElectraConfig, ErnieConfig, ErnieMConfig, EsmConfig, FlaubertConfig, FNetConfig, FunnelConfig, GPT2Config, GPT2Config, GPTBigCodeConfig, GPTNeoConfig, GPTNeoXConfig, GPTJConfig, IBertConfig, LayoutLMConfig, LayoutLMv2Config, LayoutLMv3Config, LEDConfig, LiltConfig, LlamaConfig, LongformerConfig, LukeConfig, MarkupLMConfig, MBartConfig, MegaConfig, MegatronBertConfig, MobileBertConfig, MPNetConfig, MvpConfig, NezhaConfig, NystromformerConfig, OpenAIGPTConfig, OPTConfig, PerceiverConfig, PLBartConfig, QDQBertConfig, ReformerConfig, RemBertConfig, RobertaConfig, RobertaPreLayerNormConfig, RoCBertConfig, RoFormerConfig, SqueezeBertConfig, TapasConfig, TransfoXLConfig, XLMConfig, XLMRobertaConfig, XLMRobertaXLConfig, XLNetConfig, XmodConfig, YosoConfig. ``` ### Who can help? @ArthurZucker ### Expected behavior Model should be able to be loaded from config with randomly initialized weights, preferably with bfloat16 and load_8bit support.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24327/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24327/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24326

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24326/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24326/comments

https://api.github.com/repos/huggingface/transformers/issues/24326/events

https://github.com/huggingface/transformers/pull/24326

1,761,123,868

PR_kwDOCUB6oc5TN6Cr

24,326

Adding ddp_broadcast_buffers argument to Trainer

{ "login": "TevenLeScao", "id": 26709476, "node_id": "MDQ6VXNlcjI2NzA5NDc2", "avatar_url": "https://avatars.githubusercontent.com/u/26709476?v=4", "gravatar_id": "", "url": "https://api.github.com/users/TevenLeScao", "html_url": "https://github.com/TevenLeScao", "followers_url": "https://api.github.com/users/TevenLeScao/followers", "following_url": "https://api.github.com/users/TevenLeScao/following{/other_user}", "gists_url": "https://api.github.com/users/TevenLeScao/gists{/gist_id}", "starred_url": "https://api.github.com/users/TevenLeScao/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/TevenLeScao/subscriptions", "organizations_url": "https://api.github.com/users/TevenLeScao/orgs", "repos_url": "https://api.github.com/users/TevenLeScao/repos", "events_url": "https://api.github.com/users/TevenLeScao/events{/privacy}", "received_events_url": "https://api.github.com/users/TevenLeScao/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,686

CONTRIBUTOR

null

# What does this PR do? In #22482, using the Trainer with `gpt2` and other similar models failed in naive distributed mode. Passing `ddp_broadcast_buffers=False` to Pytorch's DDP wrapper fixes the issue. This PR surfaces that argument to the Trainer user.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24326/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24326/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24326", "html_url": "https://github.com/huggingface/transformers/pull/24326", "diff_url": "https://github.com/huggingface/transformers/pull/24326.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24326.patch", "merged_at": 1686942843000 }

https://api.github.com/repos/huggingface/transformers/issues/24325

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24325/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24325/comments

https://api.github.com/repos/huggingface/transformers/issues/24325/events

https://github.com/huggingface/transformers/pull/24325

1,760,968,723

PR_kwDOCUB6oc5TNZNQ

24,325

[Time-Series] Added link to the blog in Tips

{ "login": "elisim", "id": 17675462, "node_id": "MDQ6VXNlcjE3Njc1NDYy", "avatar_url": "https://avatars.githubusercontent.com/u/17675462?v=4", "gravatar_id": "", "url": "https://api.github.com/users/elisim", "html_url": "https://github.com/elisim", "followers_url": "https://api.github.com/users/elisim/followers", "following_url": "https://api.github.com/users/elisim/following{/other_user}", "gists_url": "https://api.github.com/users/elisim/gists{/gist_id}", "starred_url": "https://api.github.com/users/elisim/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/elisim/subscriptions", "organizations_url": "https://api.github.com/users/elisim/orgs", "repos_url": "https://api.github.com/users/elisim/repos", "events_url": "https://api.github.com/users/elisim/events{/privacy}", "received_events_url": "https://api.github.com/users/elisim/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/transformers/pr_24325). All of your documentation changes will be reflected on that endpoint.", "file names changed, opened a new one here\r\nhttps://github.com/huggingface/transformers/pull/24482" ]

1,686

1,687

CONTRIBUTOR

null

@kashif @NielsRogge

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24325/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24325/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24325", "html_url": "https://github.com/huggingface/transformers/pull/24325", "diff_url": "https://github.com/huggingface/transformers/pull/24325.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24325.patch", "merged_at": null }

https://api.github.com/repos/huggingface/transformers/issues/24324

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24324/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24324/comments

https://api.github.com/repos/huggingface/transformers/issues/24324/events

https://github.com/huggingface/transformers/pull/24324

1,760,961,588

PR_kwDOCUB6oc5TNXqU

24,324

Allow passing kwargs through to TFBertTokenizer

{ "login": "Rocketknight1", "id": 12866554, "node_id": "MDQ6VXNlcjEyODY2NTU0", "avatar_url": "https://avatars.githubusercontent.com/u/12866554?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Rocketknight1", "html_url": "https://github.com/Rocketknight1", "followers_url": "https://api.github.com/users/Rocketknight1/followers", "following_url": "https://api.github.com/users/Rocketknight1/following{/other_user}", "gists_url": "https://api.github.com/users/Rocketknight1/gists{/gist_id}", "starred_url": "https://api.github.com/users/Rocketknight1/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Rocketknight1/subscriptions", "organizations_url": "https://api.github.com/users/Rocketknight1/orgs", "repos_url": "https://api.github.com/users/Rocketknight1/repos", "events_url": "https://api.github.com/users/Rocketknight1/events{/privacy}", "received_events_url": "https://api.github.com/users/Rocketknight1/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Ping @amyeroberts for core maintainer review now that the extra functionality is working fine (see issue #23798)", "I was a bit wary about the kwargs thing too - `FastBertTokenizer` and `BertTokenizerLayer` actually have wildly different arguments, so depending on which one you're using the kwargs you need will be totally different. Still, I think for an advanced use case it's fine - we're just trying to enable some power user behaviours without forcing them to edit the library source, and I'd prefer something general like this over specifically exposing the options I think people need (because I didn't even realize in advance that the `preserve_unused` arg would be valuable!)\r\n\r\nAnyway, merging!" ]

1,686

1,687

MEMBER

null

There are some kwargs like `preserve_unused_tokens` in the underlying TF tokenizer layers that might be useful to expose to users. This PR exposes them by passing through any unrecognized `kwargs` in the model `__init__` to the TF tokenizer layer. Fixes #23798

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24324/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24324/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24324", "html_url": "https://github.com/huggingface/transformers/pull/24324", "diff_url": "https://github.com/huggingface/transformers/pull/24324.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24324.patch", "merged_at": 1687261746000 }

https://api.github.com/repos/huggingface/transformers/issues/24323

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24323/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24323/comments

https://api.github.com/repos/huggingface/transformers/issues/24323/events

https://github.com/huggingface/transformers/issues/24323

1,760,948,980

I_kwDOCUB6oc5o9fL0

24,323

Protobuf 4 support (again)

{ "login": "dustyketchum", "id": 3933135, "node_id": "MDQ6VXNlcjM5MzMxMzU=", "avatar_url": "https://avatars.githubusercontent.com/u/3933135?v=4", "gravatar_id": "", "url": "https://api.github.com/users/dustyketchum", "html_url": "https://github.com/dustyketchum", "followers_url": "https://api.github.com/users/dustyketchum/followers", "following_url": "https://api.github.com/users/dustyketchum/following{/other_user}", "gists_url": "https://api.github.com/users/dustyketchum/gists{/gist_id}", "starred_url": "https://api.github.com/users/dustyketchum/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/dustyketchum/subscriptions", "organizations_url": "https://api.github.com/users/dustyketchum/orgs", "repos_url": "https://api.github.com/users/dustyketchum/repos", "events_url": "https://api.github.com/users/dustyketchum/events{/privacy}", "received_events_url": "https://api.github.com/users/dustyketchum/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Hi @dustyketchum, \r\n\r\nThanks for raising this issue. Managing the different dependencies in the library can be quite complex. As noted in the linked issue, the blocker for upgrading protobuf support was third-party libraries support of protobuf 4 rather than our own. \r\n\r\nIf this is something you or someone else in the community believes is very important, please feel free to open a PR. Note that it's not just necessary for the CI to be green and protobuf 4 be supported, we must also remain backwards compatible with previous versions.\r\n\r\ncc @ydshieh ", "BTW, what's blocking when you try to use python 3.10 without protobuf 4.x ?", "I am not blocked; I should have been more precise.\r\n", "Fixed in #24599" ]

1,686

1,688

NONE

null

### Feature request Looking at https://github.com/huggingface/transformers/issues/21677#issuecomment-1435072007, I notice there are now new versions of tensorflow and tensorboard that may help with protobuf 4.x compatibility. It would be awesome to get this upgraded, thanks! ### Motivation easier upgrade path to python 3.10 and above ### Your contribution nope, sorry.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24323/reactions", "total_count": 1, "+1": 1, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24323/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24322

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24322/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24322/comments

https://api.github.com/repos/huggingface/transformers/issues/24322/events

https://github.com/huggingface/transformers/pull/24322

1,760,689,621

PR_kwDOCUB6oc5TMb7t

24,322

Respect explicitly set framework parameter in pipeline

{ "login": "denis-ismailaj", "id": 20902736, "node_id": "MDQ6VXNlcjIwOTAyNzM2", "avatar_url": "https://avatars.githubusercontent.com/u/20902736?v=4", "gravatar_id": "", "url": "https://api.github.com/users/denis-ismailaj", "html_url": "https://github.com/denis-ismailaj", "followers_url": "https://api.github.com/users/denis-ismailaj/followers", "following_url": "https://api.github.com/users/denis-ismailaj/following{/other_user}", "gists_url": "https://api.github.com/users/denis-ismailaj/gists{/gist_id}", "starred_url": "https://api.github.com/users/denis-ismailaj/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/denis-ismailaj/subscriptions", "organizations_url": "https://api.github.com/users/denis-ismailaj/orgs", "repos_url": "https://api.github.com/users/denis-ismailaj/repos", "events_url": "https://api.github.com/users/denis-ismailaj/events{/privacy}", "received_events_url": "https://api.github.com/users/denis-ismailaj/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Updated so that `infer_framework_load_model` is not called at all if `model` is already loaded and `framework` is defined.\r\n\r\nHowever, it may still be worth keeping the check inside `infer_framework_load_model` so that if `framework` is defined but the `model` is a `str` and we do need to call `infer_framework_load_model`, at least we don't need to then call `infer_framework` inside of it.", "Ok, this code breaks some tests. Can you fix them ?", "Noticed a \"typo\", it is fixed now.", "@denis-ismailaj it's ready to merge ! " ]

1,686

1,687

CONTRIBUTOR

null

# What does this PR do?   Fixes #24321 ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [X] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24322/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24322/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24322", "html_url": "https://github.com/huggingface/transformers/pull/24322", "diff_url": "https://github.com/huggingface/transformers/pull/24322.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24322.patch", "merged_at": 1687257833000 }

https://api.github.com/repos/huggingface/transformers/issues/24321

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24321/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24321/comments

https://api.github.com/repos/huggingface/transformers/issues/24321/events

https://github.com/huggingface/transformers/issues/24321

1,760,680,155

I_kwDOCUB6oc5o8djb

24,321

[pipeline] Explicitly set framework is ignored

{ "login": "denis-ismailaj", "id": 20902736, "node_id": "MDQ6VXNlcjIwOTAyNzM2", "avatar_url": "https://avatars.githubusercontent.com/u/20902736?v=4", "gravatar_id": "", "url": "https://api.github.com/users/denis-ismailaj", "html_url": "https://github.com/denis-ismailaj", "followers_url": "https://api.github.com/users/denis-ismailaj/followers", "following_url": "https://api.github.com/users/denis-ismailaj/following{/other_user}", "gists_url": "https://api.github.com/users/denis-ismailaj/gists{/gist_id}", "starred_url": "https://api.github.com/users/denis-ismailaj/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/denis-ismailaj/subscriptions", "organizations_url": "https://api.github.com/users/denis-ismailaj/orgs", "repos_url": "https://api.github.com/users/denis-ismailaj/repos", "events_url": "https://api.github.com/users/denis-ismailaj/events{/privacy}", "received_events_url": "https://api.github.com/users/denis-ismailaj/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

1,686

1,687

CONTRIBUTOR

null

### System Info - `transformers` version: 4.30.2 Omitting the rest because they aren't really relevant. Can submit later if needed. ### Who can help? @Narsil ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [X] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [ ] My own task or dataset (give details below) ### Reproduction ``` from transformers import pipeline from transformers import WhisperProcessor, WhisperForConditionalGeneration processor = WhisperProcessor.from_pretrained("openai/whisper-tiny") model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-tiny") # Using this just to make sure an error if thrown if pipeline tries to check using module instead of specified framework. class FakeWhisper: def __getattr__(self, item): return model.__getattr__(item) pipe = pipeline( "automatic-speech-recognition", model=FakeWhisper(), tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, chunk_length_s=30, device=model.device, framework="pt", ) ``` The above code raises this error: ``` TypeError: Could not infer framework from class <class '__main__.FakeWhisper'>. ``` ### Expected behavior When specifying the framework explicitely, there is no need to infer it from the module of the model class, as mentioned here: https://github.com/huggingface/transformers/blob/main/src/transformers/pipelines/__init__.py#L784-L796 But then, inside `infer_framework_load_model`, `infer_framework` is called regardless of the value of the `framework` parameter: https://github.com/huggingface/transformers/blob/main/src/transformers/pipelines/base.py#L280

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24321/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24321/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24320

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24320/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24320/comments

https://api.github.com/repos/huggingface/transformers/issues/24320/events

https://github.com/huggingface/transformers/pull/24320

1,760,661,965

PR_kwDOCUB6oc5TMV2f

24,320

Add test for proper TF input signatures

{ "login": "Rocketknight1", "id": 12866554, "node_id": "MDQ6VXNlcjEyODY2NTU0", "avatar_url": "https://avatars.githubusercontent.com/u/12866554?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Rocketknight1", "html_url": "https://github.com/Rocketknight1", "followers_url": "https://api.github.com/users/Rocketknight1/followers", "following_url": "https://api.github.com/users/Rocketknight1/following{/other_user}", "gists_url": "https://api.github.com/users/Rocketknight1/gists{/gist_id}", "starred_url": "https://api.github.com/users/Rocketknight1/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Rocketknight1/subscriptions", "organizations_url": "https://api.github.com/users/Rocketknight1/orgs", "repos_url": "https://api.github.com/users/Rocketknight1/repos", "events_url": "https://api.github.com/users/Rocketknight1/events{/privacy}", "received_events_url": "https://api.github.com/users/Rocketknight1/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,686

MEMBER

null

This is a relatively simple PR intended to check that all TF models have proper input signatures that match their inputs. I was calling the function `self._prune_signature` in a few places to verify this, which felt a bit hacky. This test should let us get rid of `self._prune_signature` by enforcing valid signatures for all models. Edit: I'm also slipping in a typo fix (fine-tine -> fine-tune) I saw while I was running the tests Double-edit: I'm also slipping in a fix to an incorrect indentation in the `test_dataset_conversion` test that was causing some unnecessary repetition

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24320/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24320/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24320", "html_url": "https://github.com/huggingface/transformers/pull/24320", "diff_url": "https://github.com/huggingface/transformers/pull/24320.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24320.patch", "merged_at": 1686931393000 }

https://api.github.com/repos/huggingface/transformers/issues/24319

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24319/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24319/comments

https://api.github.com/repos/huggingface/transformers/issues/24319/events

https://github.com/huggingface/transformers/pull/24319

1,760,611,507

PR_kwDOCUB6oc5TMKrp

24,319

Fix ner average grouping with no groups

{ "login": "Narsil", "id": 204321, "node_id": "MDQ6VXNlcjIwNDMyMQ==", "avatar_url": "https://avatars.githubusercontent.com/u/204321?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Narsil", "html_url": "https://github.com/Narsil", "followers_url": "https://api.github.com/users/Narsil/followers", "following_url": "https://api.github.com/users/Narsil/following{/other_user}", "gists_url": "https://api.github.com/users/Narsil/gists{/gist_id}", "starred_url": "https://api.github.com/users/Narsil/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Narsil/subscriptions", "organizations_url": "https://api.github.com/users/Narsil/orgs", "repos_url": "https://api.github.com/users/Narsil/repos", "events_url": "https://api.github.com/users/Narsil/events{/privacy}", "received_events_url": "https://api.github.com/users/Narsil/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "For bigger fixes I would add a test. This is small enough I think it's ok to skip. Let me know.", "_The documentation is not available anymore as the PR was closed or merged._" ]

1,686

CONTRIBUTOR

null

# What does this PR do? Fixes #https://github.com/huggingface/transformers/issues/24314   Fixes # (issue) ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [ ] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24319/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24319/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24319", "html_url": "https://github.com/huggingface/transformers/pull/24319", "diff_url": "https://github.com/huggingface/transformers/pull/24319.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24319.patch", "merged_at": 1686926599000 }

https://api.github.com/repos/huggingface/transformers/issues/24318

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24318/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24318/comments

https://api.github.com/repos/huggingface/transformers/issues/24318/events

https://github.com/huggingface/transformers/issues/24318

1,760,348,442

I_kwDOCUB6oc5o7Mka

24,318

Recursion error when creating AutoTokenizer

{ "login": "markovalexander", "id": 22663468, "node_id": "MDQ6VXNlcjIyNjYzNDY4", "avatar_url": "https://avatars.githubusercontent.com/u/22663468?v=4", "gravatar_id": "", "url": "https://api.github.com/users/markovalexander", "html_url": "https://github.com/markovalexander", "followers_url": "https://api.github.com/users/markovalexander/followers", "following_url": "https://api.github.com/users/markovalexander/following{/other_user}", "gists_url": "https://api.github.com/users/markovalexander/gists{/gist_id}", "starred_url": "https://api.github.com/users/markovalexander/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/markovalexander/subscriptions", "organizations_url": "https://api.github.com/users/markovalexander/orgs", "repos_url": "https://api.github.com/users/markovalexander/repos", "events_url": "https://api.github.com/users/markovalexander/events{/privacy}", "received_events_url": "https://api.github.com/users/markovalexander/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "Interesting, I can't even load the autotokenizer from remote, i.e., just loading `from_pretrained` using the remote identifier doesn't work and leads to a recursion error.", "Hey! There seems to be something a bit strange with the `tokenizer_config.json`, since the `unk_token`, the `bos_token` as well as the `eos_token` are `\"\"`, which means that they are empty. This is the root cause of the issue. \r\nThe following works: \r\n```python \r\n>>> from transformers import AutoTokenizer; tok = AutoTokenizer.from_pretrained(path, unk_token =\"<s>\")\r\n```\r\nWhile this does not work:\r\n```python \r\n>>> from transformers import AutoTokenizer; tok = AutoTokenizer.from_pretrained(path, unk_token =\" <s>\")\r\n```\r\n`\"\"` is not part of the vocab, thus it cannot be used as an unknown token. You should update the tokenizer to have a `tok._unk_token = None`", "There's something strange going on. I am facing the same problem and am unable to load a `LlamaTokenizer` on 4.30.2 that I have used previously (with 4.28.x). Here's my `tokenizer_config.json` in case that's relevant:\r\n\r\n```json\r\n{\r\n \"bos_token\": \"<s>\",\r\n \"clean_up_tokenization_spaces\": false,\r\n \"eos_token\": \"</s>\",\r\n \"model_max_length\": 1000000000000000019884624838656,\r\n \"tokenizer_class\": \"LlamaTokenizer\",\r\n \"unk_token\": \"<unk>\"\r\n}\r\n```", "You are giving way too little informations, no traceback and this is not related to the mentioned issue. If you still have a problem, feel free to open a new issue, add a full reproducer and make sure you have a correctly converted tokenizer", "> Hey! There seems to be something a bit strange with the `tokenizer_config.json`, since the `unk_token`, the `bos_token` as well as the `eos_token` are `\"\"`, which means that they are empty. This is the root cause of the issue. The following works:\r\n> \r\n> ```python\r\n> >>> from transformers import AutoTokenizer; tok = AutoTokenizer.from_pretrained(path, unk_token =\"<s>\")\r\n> ```\r\n> \r\n> While this does not work:\r\n> \r\n> ```python\r\n> >>> from transformers import AutoTokenizer; tok = AutoTokenizer.from_pretrained(path, unk_token =\" <s>\")\r\n> ```\r\n> \r\n> `\"\"` is not part of the vocab, thus it cannot be used as an unknown token. You should update the tokenizer to have a `tok._unk_token = None`\r\n\r\nI think it is not a solution. Why does simple save/load not work then? Maybe at least add that information to docs?", "If you load using `from transformers import AutoTokenizer; tok = AutoTokenizer.from_pretrained(\"Yhyu13/chimera-inst-chat-13b-hf\", legacy=False, use_fast = False)`, you can load and save. \r\n\r\nThe reason why it does not work is because the tokenizer was saved using the `LlamaTokenizer` class. When doing the automatic conversion to a fast tokenizer, (`AutoTokenizer` automatically converts the slow to a fast tokenizer using the [`LlamaConverter` ](https://github.com/ArthurZucker/transformers/blob/1f2434777ecfc436aed40c282b074034f7232d6f/src/transformers/convert_slow_tokenizer.py#L1123). \r\n\r\nThe issue lies with `self.update_post_processor()`, which does not check if the `bos` and `eos` tokens are defined or if `add_bos_token` and `add_eos_token` are set to ` True`. \r\n\r\nHowever the configuration files are still wrong, the `eos` and `bos` and `unk` tokens from the slow tokenizer are going to be different:\r\n\r\n```python \r\n >>> from transformers import AutoTokenizer; tok = AutoTokenizer.from_pretrained(\"Yhyu13/chimera-inst-chat-13b-hf\", legacy=True, use_fast = False)\r\n>>> tok\r\nLlamaTokenizer(name_or_path='Yhyu13/chimera-inst-chat-13b-hf', vocab_size=32000, model_max_length=2048, is_fast=False, padding_side='right', truncation_side='right', special_tokens={'bos_token': '<s>', 'eos_token': '</s>', 'unk_token': '<unk>', 'pad_token': '<pad>'}, clean_up_tokenization_spaces=False)\r\n```\r\nthe values the `tokenizer_config` were not used because the `[special_tokens_map](https://huggingface.co/Yhyu13/chimera-inst-chat-13b-hf/blob/main/special_tokens_map.json)` was saved. \r\n\r\nTLDR; the tokenizer was not properly saved, priority is given to the `tokenizer_config.json` when loading the tokenizer, which is wrong in this case. " ]

1,686

1,691

CONTRIBUTOR

null

### System Info ``` - `transformers` version: 4.30.2 - Platform: Linux-5.15.0-1033-oracle-x86_64-with-glibc2.29 - Python version: 3.8.10 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): 2.0.0+cu117 (False) - Tensorflow version (GPU?): not installed (NA) - Flax version (CPU?/GPU?/TPU?): not installed (NA) - Jax version: not installed - JaxLib version: not installed - Using GPU in script?: No - Using distributed or parallel set-up in script?: No ``` ### Who can help? @ArthurZucker ### Information - [ ] The official example scripts - [X] My own modified scripts ### Tasks - [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...) - [X] My own task or dataset (give details below) ### Reproduction 1) Save `"Yhyu13/chimera-inst-chat-13b-hf"` tokenizer as `save_pretrained` to some folder 2) Try to create auto tokenizer from that folder ```bash python -c "from transformers import AutoTokenizer; tok = AutoTokenizer.from_pretrained('/local/path/to/tokenizer'); print(tok)" ``` And see recursion error when running `.from_pretrained` (last lines in stack trace): ``` File "/home/alexander/llm_training/.venv/lib/python3.8/site-packages/transformers/tokenization_utils_fast.py", line 257, in _convert_token_to_id_with_added_voc return self.unk_token_id File "/home/alexander/llm_training/.venv/lib/python3.8/site-packages/transformers/tokenization_utils_base.py", line 1155, in unk_token_id return self.convert_tokens_to_ids(self.unk_token) File "/home/alexander/llm_training/.venv/lib/python3.8/site-packages/transformers/tokenization_utils_fast.py", line 250, in convert_tokens_to_ids return self._convert_token_to_id_with_added_voc(tokens) File "/home/alexander/llm_training/.venv/lib/python3.8/site-packages/transformers/tokenization_utils_fast.py", line 257, in _convert_token_to_id_with_added_voc return self.unk_token_id File "/home/alexander/llm_training/.venv/lib/python3.8/site-packages/transformers/tokenization_utils_base.py", line 1155, in unk_token_id return self.convert_tokens_to_ids(self.unk_token) File "/home/alexander/llm_training/.venv/lib/python3.8/site-packages/transformers/tokenization_utils_base.py", line 1035, in unk_token return str(self._unk_token) RecursionError: maximum recursion depth exceeded while getting the str of an object ``` ### Expected behavior After running `pip install transformers==4.29` everything works fine: ```bash ❯ python -c "from transformers import AutoTokenizer; tok = AutoTokenizer.from_pretrained('/local/path/to/tokenizer'); print(tok)" LlamaTokenizerFast(name_or_path='/local/path/to/tokenizer', vocab_size=32000, model_max_length=2048, is_fast=True, padding_side='right', truncation_side='right', special_tokens={'bos_token': '<s>', 'eos_token': '</s>', 'unk_token': '<unk>', 'pad_token': '<pad>'}, clean_up_tokenization_spaces=False) ``` Working transformers-cli env: ``` - `transformers` version: 4.29.0 - Platform: Linux-5.15.0-1033-oracle-x86_64-with-glibc2.29 - Python version: 3.8.10 - Huggingface_hub version: 0.15.1 - Safetensors version: 0.3.1 - PyTorch version (GPU?): 2.0.0+cu117 (False) - Tensorflow version (GPU?): not installed (NA) - Flax version (CPU?/GPU?/TPU?): not installed (NA) - Jax version: not installed - JaxLib version: not installed - Using GPU in script?: No - Using distributed or parallel set-up in script?: No ```

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24318/reactions", "total_count": 1, "+1": 1, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24318/timeline

completed

null

https://api.github.com/repos/huggingface/transformers/issues/24317

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24317/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24317/comments

https://api.github.com/repos/huggingface/transformers/issues/24317/events

https://github.com/huggingface/transformers/pull/24317

1,760,332,952

PR_kwDOCUB6oc5TLMdO

24,317

Fix ImageGPT doc example

{ "login": "amyeroberts", "id": 22614925, "node_id": "MDQ6VXNlcjIyNjE0OTI1", "avatar_url": "https://avatars.githubusercontent.com/u/22614925?v=4", "gravatar_id": "", "url": "https://api.github.com/users/amyeroberts", "html_url": "https://github.com/amyeroberts", "followers_url": "https://api.github.com/users/amyeroberts/followers", "following_url": "https://api.github.com/users/amyeroberts/following{/other_user}", "gists_url": "https://api.github.com/users/amyeroberts/gists{/gist_id}", "starred_url": "https://api.github.com/users/amyeroberts/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/amyeroberts/subscriptions", "organizations_url": "https://api.github.com/users/amyeroberts/orgs", "repos_url": "https://api.github.com/users/amyeroberts/repos", "events_url": "https://api.github.com/users/amyeroberts/events{/privacy}", "received_events_url": "https://api.github.com/users/amyeroberts/received_events", "type": "User", "site_admin": false }

[]

closed

false

null

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

1,686

COLLABORATOR

null

# What does this PR do? There was a bug in the example, where the clusters in the image processor were stored as lists, but the example assumed they were numpy arrays. At the moment, clusters are stored as a list of lists, but [converted to a numpy array during processing](https://github.com/huggingface/transformers/blob/0b7b4429c78de68acaf72224eb6dae43616d820c/src/transformers/models/imagegpt/image_processing_imagegpt.py#L230). This PR converts to a numpy array when setting the class attribute and if new `cluster` are passed into `preprocess`. This: * Maintains backwards compatibility with old configurations * Saved configs aren't changed (`clusters` is still converted to a list of lists when serializing) see this [dummy image processor](https://huggingface.co/amyeroberts/dummy_imagegpt_image_processor_np_clusters) created from this branch. * Is more efficient - we're not converting the same list of lists every batch. A potential breaking change is if users were accessing the `clusters` attribute and using it as a list of lists. As this was caught because users were using the clusters as a numpy array (according to the example) I expect this impact to be low. Fixes #24189 ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [x] Did you write any new necessary tests?

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24317/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24317/timeline

null

false

{ "url": "https://api.github.com/repos/huggingface/transformers/pulls/24317", "html_url": "https://github.com/huggingface/transformers/pull/24317", "diff_url": "https://github.com/huggingface/transformers/pull/24317.diff", "patch_url": "https://github.com/huggingface/transformers/pull/24317.patch", "merged_at": 1686931282000 }

https://api.github.com/repos/huggingface/transformers/issues/24316

https://api.github.com/repos/huggingface/transformers

https://api.github.com/repos/huggingface/transformers/issues/24316/labels{/name}

https://api.github.com/repos/huggingface/transformers/issues/24316/comments

https://api.github.com/repos/huggingface/transformers/issues/24316/events

https://github.com/huggingface/transformers/issues/24316

1,760,298,276

I_kwDOCUB6oc5o7AUk

24,316

[Tokenizer] `skip_special_tokens` not working as expected

{ "login": "ArthurZucker", "id": 48595927, "node_id": "MDQ6VXNlcjQ4NTk1OTI3", "avatar_url": "https://avatars.githubusercontent.com/u/48595927?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ArthurZucker", "html_url": "https://github.com/ArthurZucker", "followers_url": "https://api.github.com/users/ArthurZucker/followers", "following_url": "https://api.github.com/users/ArthurZucker/following{/other_user}", "gists_url": "https://api.github.com/users/ArthurZucker/gists{/gist_id}", "starred_url": "https://api.github.com/users/ArthurZucker/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ArthurZucker/subscriptions", "organizations_url": "https://api.github.com/users/ArthurZucker/orgs", "repos_url": "https://api.github.com/users/ArthurZucker/repos", "events_url": "https://api.github.com/users/ArthurZucker/events{/privacy}", "received_events_url": "https://api.github.com/users/ArthurZucker/received_events", "type": "User", "site_admin": false }

[ { "id": 1834056635, "node_id": "MDU6TGFiZWwxODM0MDU2NjM1", "url": "https://api.github.com/repos/huggingface/transformers/labels/Core:%20Tokenization", "name": "Core: Tokenization", "color": "FF4446", "default": false, "description": "Internals of the library; Tokenization." } ]

closed

false

{ "login": "ArthurZucker", "id": 48595927, "node_id": "MDQ6VXNlcjQ4NTk1OTI3", "avatar_url": "https://avatars.githubusercontent.com/u/48595927?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ArthurZucker", "html_url": "https://github.com/ArthurZucker", "followers_url": "https://api.github.com/users/ArthurZucker/followers", "following_url": "https://api.github.com/users/ArthurZucker/following{/other_user}", "gists_url": "https://api.github.com/users/ArthurZucker/gists{/gist_id}", "starred_url": "https://api.github.com/users/ArthurZucker/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ArthurZucker/subscriptions", "organizations_url": "https://api.github.com/users/ArthurZucker/orgs", "repos_url": "https://api.github.com/users/ArthurZucker/repos", "events_url": "https://api.github.com/users/ArthurZucker/events{/privacy}", "received_events_url": "https://api.github.com/users/ArthurZucker/received_events", "type": "User", "site_admin": false }

[ { "login": "ArthurZucker", "id": 48595927, "node_id": "MDQ6VXNlcjQ4NTk1OTI3", "avatar_url": "https://avatars.githubusercontent.com/u/48595927?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ArthurZucker", "html_url": "https://github.com/ArthurZucker", "followers_url": "https://api.github.com/users/ArthurZucker/followers", "following_url": "https://api.github.com/users/ArthurZucker/following{/other_user}", "gists_url": "https://api.github.com/users/ArthurZucker/gists{/gist_id}", "starred_url": "https://api.github.com/users/ArthurZucker/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ArthurZucker/subscriptions", "organizations_url": "https://api.github.com/users/ArthurZucker/orgs", "repos_url": "https://api.github.com/users/ArthurZucker/repos", "events_url": "https://api.github.com/users/ArthurZucker/events{/privacy}", "received_events_url": "https://api.github.com/users/ArthurZucker/received_events", "type": "User", "site_admin": false } ]

[ "Okay! the fix PR should be merged soon. Note that the behaviour is gonna be deprecated! " ]

1,686

1,695

COLLABORATOR

null

# Reporting a failing API design This is mostly to help me record some of the biggest issues with the current API for adding tokens. This is linked to #23909. Here is a simple snippet: ```python >>> from transformers import AutoTokenizer >>> tokenizer = AutoTokenizer.from_pretrained("t5-base", use_fast = False) >>> tokenizer = [ AddedToken("[ABC]", normalized=False), AddedToken("[DEF]", normalized=False), AddedToken("GHI IHG", normalized=False), ] >>> tokenizer.add_tokens(new_toks) >>> tokenizer.add_tokens([AddedToken("[SAMPLE]", normalized=True)], special_tokens = True) >>> print(tokenizer.added_tokens_encoder) >>> print( tokenizer.all_special_ids) ``` This will show that the newly added token (`[SAMPLE]`) is not part of the `all_special_ids`. However, `all_special_ids` is used when decoding, to check whether the token should be skipped or not: ```python for token in filtered_tokens: if skip_special_tokens and token in self.all_special_ids: continue if token in self.added_tokens_encoder: if current_sub_text: sub_texts.append(self.convert_tokens_to_string(current_sub_text)) current_sub_text = [] sub_texts.append(token) else: current_sub_text.append(token) ``` Thus ```python >>> encoded = tokenizer.encode("[ABC] [DEF][SAMPLE]", add_special_tokens=False) >>> tokenizer.decode(encoded, skip_special_tokens = True) "[ABC] [DEF][SAMPLE]" ``` However, the token is in `added_tokens_encoder` but not in `additional_special_tokens`. Now imagine you want `spaces_between_special_tokens` ? This will add spaces between all added tokens, and thus checks if a token is part of `tokenzier.added_tokens_encoder`. ```python >>> encoded = tokenizer.encode("[ABC] [DEF][SAMPLE]", add_special_tokens=False) >>> tokenizer.decode(encoded, spaces_between_special_tokens = True) "[ABC] [DEF] [SAMPLE]" >>> tokenizer.decode(encoded, spaces_between_special_tokens = False) "[ABC][DEF][SAMPLE]" ```

{ "url": "https://api.github.com/repos/huggingface/transformers/issues/24316/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

https://api.github.com/repos/huggingface/transformers/issues/24316/timeline

completed

null