url string | repository_url string | labels_url string | comments_url string | events_url string | html_url string | id int64 | node_id string | number int64 | title string | user dict | labels list | state string | locked bool | assignee dict | assignees list | milestone null | comments list | created_at timestamp[ms] | updated_at timestamp[ms] | closed_at timestamp[ms] | author_association string | type dict | active_lock_reason null | draft bool | pull_request dict | body string | closed_by dict | reactions dict | timeline_url string | performed_via_github_app null | state_reason string | sub_issues_summary dict | issue_dependencies_summary dict | is_pull_request bool | is_closed bool |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
https://api.github.com/repos/huggingface/transformers/issues/38833 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38833/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38833/comments | https://api.github.com/repos/huggingface/transformers/issues/38833/events | https://github.com/huggingface/transformers/pull/38833 | 3,147,579,361 | PR_kwDOCUB6oc6al5aY | 38,833 | Granite speech - minor fixes to support training with the HF trainer | {
"login": "avihu111",
"id": 39214195,
"node_id": "MDQ6VXNlcjM5MjE0MTk1",
"avatar_url": "https://avatars.githubusercontent.com/u/39214195?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/avihu111",
"html_url": "https://github.com/avihu111",
"followers_url": "https://api.github.com/users/avihu111/followers",
"following_url": "https://api.github.com/users/avihu111/following{/other_user}",
"gists_url": "https://api.github.com/users/avihu111/gists{/gist_id}",
"starred_url": "https://api.github.com/users/avihu111/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/avihu111/subscriptions",
"organizations_url": "https://api.github.com/users/avihu111/orgs",
"repos_url": "https://api.github.com/users/avihu111/repos",
"events_url": "https://api.github.com/users/avihu111/events{/privacy}",
"received_events_url": "https://api.github.com/users/avihu111/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-15T14:00:37 | 2025-06-24T16:11:43 | 2025-06-24T15:06:52 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38833",
"html_url": "https://github.com/huggingface/transformers/pull/38833",
"diff_url": "https://github.com/huggingface/transformers/pull/38833.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38833.patch",
"merged_at": "2025-06-24T15:06:52"
} | # What does this PR do?
Minor updates to `granite_speech` to enable finetuning it with HF trainers.
- avoids a crash when trainers pass `padding=True` to the processor
- ensure all trainable parameters have gradients (bugfix) - remove `.data` from a forward call
- rename `melspec` to `mel_filters` to leverage [this](https://github.com/huggingface/transformers/blob/d5d007a1a0f0c11a726a54c8f00bd71825f84d02/src/transformers/feature_extraction_utils.py#L595), which avoids a crash on `save_pretrained`
## 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#create-a-pull-request),
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.
CC: @ArthurZucker @eustlb @alex-jw-brooks @avishaiElmakies @gsaon
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38833/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/38833/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38832 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38832/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38832/comments | https://api.github.com/repos/huggingface/transformers/issues/38832/events | https://github.com/huggingface/transformers/issues/38832 | 3,147,461,439 | I_kwDOCUB6oc67mnM_ | 38,832 | ImportError: cannot import name 'BitsAndBytesConfig' from 'bitsandbytes' | {
"login": "tanuj-rai",
"id": 84439872,
"node_id": "MDQ6VXNlcjg0NDM5ODcy",
"avatar_url": "https://avatars.githubusercontent.com/u/84439872?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tanuj-rai",
"html_url": "https://github.com/tanuj-rai",
"followers_url": "https://api.github.com/users/tanuj-rai/followers",
"following_url": "https://api.github.com/users/tanuj-rai/following{/other_user}",
"gists_url": "https://api.github.com/users/tanuj-rai/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tanuj-rai/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tanuj-rai/subscriptions",
"organizations_url": "https://api.github.com/users/tanuj-rai/orgs",
"repos_url": "https://api.github.com/users/tanuj-rai/repos",
"events_url": "https://api.github.com/users/tanuj-rai/events{/privacy}",
"received_events_url": "https://api.github.com/users/tanuj-rai/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-15T12:26:40 | 2025-06-15T13:51:31 | 2025-06-15T13:51:31 | CONTRIBUTOR | null | null | null | null | ### System Info
```
Torch version: 2.7.1+cu126
Transformers version: 4.41.2
Datasets version: 2.20.0
PEFT version: 0.11.1
TRL version: 0.8.6
Bitsandbytes version: 0.46.0
Accelerate version: 0.31.0
```
```!pip install -q transformers datasets accelerate peft bitsandbytes einops```
```
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, TrainingArguments, Trainer
from datasets import load_dataset,
from peft import LoraConfig, get_peft_model,
from trl import SFTTrainer,
from bitsandbytes import BitsAndBytesConfig,
import os
```
On running the code I am getting following error.
```
ImportError Traceback (most recent call last)
[<ipython-input-1-3270306226>](https://localhost:8080/#) in <cell line: 0>()
4 from peft import LoraConfig, get_peft_model
5 from trl import SFTTrainer
----> 6 from bitsandbytes import BitsAndBytesConfig
7 import os
ImportError: cannot import name 'BitsAndBytesConfig' from 'bitsandbytes' (/usr/local/lib/python3.11/dist-packages/bitsandbytes/__init__.py)
---------------------------------------------------------------------------
NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.
To view examples of installing some common dependencies, click the
```
### Who can help?
@Rocketknight1 @ArthurZucker
### 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
Run the code provided in the issue.
### Expected behavior
BitsAndBytesConfig will import without any error. | {
"login": "tanuj-rai",
"id": 84439872,
"node_id": "MDQ6VXNlcjg0NDM5ODcy",
"avatar_url": "https://avatars.githubusercontent.com/u/84439872?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tanuj-rai",
"html_url": "https://github.com/tanuj-rai",
"followers_url": "https://api.github.com/users/tanuj-rai/followers",
"following_url": "https://api.github.com/users/tanuj-rai/following{/other_user}",
"gists_url": "https://api.github.com/users/tanuj-rai/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tanuj-rai/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tanuj-rai/subscriptions",
"organizations_url": "https://api.github.com/users/tanuj-rai/orgs",
"repos_url": "https://api.github.com/users/tanuj-rai/repos",
"events_url": "https://api.github.com/users/tanuj-rai/events{/privacy}",
"received_events_url": "https://api.github.com/users/tanuj-rai/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38832/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/38832/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38831 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38831/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38831/comments | https://api.github.com/repos/huggingface/transformers/issues/38831/events | https://github.com/huggingface/transformers/pull/38831 | 3,147,332,468 | PR_kwDOCUB6oc6alEhd | 38,831 | shark test | {
"login": "SeohyunNam",
"id": 110328943,
"node_id": "U_kgDOBpN8bw",
"avatar_url": "https://avatars.githubusercontent.com/u/110328943?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SeohyunNam",
"html_url": "https://github.com/SeohyunNam",
"followers_url": "https://api.github.com/users/SeohyunNam/followers",
"following_url": "https://api.github.com/users/SeohyunNam/following{/other_user}",
"gists_url": "https://api.github.com/users/SeohyunNam/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SeohyunNam/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SeohyunNam/subscriptions",
"organizations_url": "https://api.github.com/users/SeohyunNam/orgs",
"repos_url": "https://api.github.com/users/SeohyunNam/repos",
"events_url": "https://api.github.com/users/SeohyunNam/events{/privacy}",
"received_events_url": "https://api.github.com/users/SeohyunNam/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-15T10:29:50 | 2025-06-16T13:12:03 | 2025-06-16T13:12:02 | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38831",
"html_url": "https://github.com/huggingface/transformers/pull/38831",
"diff_url": "https://github.com/huggingface/transformers/pull/38831.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38831.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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).
- [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38831/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/38831/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38830 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38830/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38830/comments | https://api.github.com/repos/huggingface/transformers/issues/38830/events | https://github.com/huggingface/transformers/pull/38830 | 3,147,194,477 | PR_kwDOCUB6oc6akrjM | 38,830 | Update phi4_multimodal.md | {
"login": "tanuj-rai",
"id": 84439872,
"node_id": "MDQ6VXNlcjg0NDM5ODcy",
"avatar_url": "https://avatars.githubusercontent.com/u/84439872?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tanuj-rai",
"html_url": "https://github.com/tanuj-rai",
"followers_url": "https://api.github.com/users/tanuj-rai/followers",
"following_url": "https://api.github.com/users/tanuj-rai/following{/other_user}",
"gists_url": "https://api.github.com/users/tanuj-rai/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tanuj-rai/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tanuj-rai/subscriptions",
"organizations_url": "https://api.github.com/users/tanuj-rai/orgs",
"repos_url": "https://api.github.com/users/tanuj-rai/repos",
"events_url": "https://api.github.com/users/tanuj-rai/events{/privacy}",
"received_events_url": "https://api.github.com/users/tanuj-rai/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-15T07:10:19 | 2025-07-15T10:40:34 | 2025-07-14T17:35:17 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38830",
"html_url": "https://github.com/huggingface/transformers/pull/38830",
"diff_url": "https://github.com/huggingface/transformers/pull/38830.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38830.patch",
"merged_at": "2025-07-14T17:35:17"
} | # What does this PR do?
This PR updates the Phi-4 Multimodal card documentation as described in #36979, in an attempt to standardize all model-cards.
#36979
## 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#create-a-pull-request),
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.
- [ ] 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?
@stevhliu
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38830/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/38830/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38829 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38829/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38829/comments | https://api.github.com/repos/huggingface/transformers/issues/38829/events | https://github.com/huggingface/transformers/issues/38829 | 3,147,180,915 | I_kwDOCUB6oc67litz | 38,829 | Support Asynchronous Evaluation on Separate GPU in `Trainer` | {
"login": "AmitMY",
"id": 5757359,
"node_id": "MDQ6VXNlcjU3NTczNTk=",
"avatar_url": "https://avatars.githubusercontent.com/u/5757359?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/AmitMY",
"html_url": "https://github.com/AmitMY",
"followers_url": "https://api.github.com/users/AmitMY/followers",
"following_url": "https://api.github.com/users/AmitMY/following{/other_user}",
"gists_url": "https://api.github.com/users/AmitMY/gists{/gist_id}",
"starred_url": "https://api.github.com/users/AmitMY/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/AmitMY/subscriptions",
"organizations_url": "https://api.github.com/users/AmitMY/orgs",
"repos_url": "https://api.github.com/users/AmitMY/repos",
"events_url": "https://api.github.com/users/AmitMY/events{/privacy}",
"received_events_url": "https://api.github.com/users/AmitMY/received_events",
"type": "User",
"user_view_type": "public",
"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 | [] | null | [] | 2025-06-15T06:52:31 | 2025-06-30T14:20:59 | null | NONE | null | null | null | null | ### Feature request
Add support for asynchronous evaluation in `transformers.Trainer`, ideally enabling evaluation to run in parallel to training — potentially on a separate GPU — without blocking the training loop.
Even ideally, to really utilize the hardware fully, one GPU should be training, the other one should be evaluating, constantly. That means, that it might take a few minutes to evaluate, and maybe save the checkpoint to disk, but afterwards, a new eval should start immediately with the latest model weights, in an infinite loop.
### Motivation
In my training scenarios, especially with:
**slow checkpointing (e.g. network-mounted disk, NFS, HDD)** and **multi-GPU machines with underutilized resources**
the current blocking evaluation step can create a significant bottleneck. It halts the training loop and prevents full utilization of the hardware.
<img width="562" alt="Image" src="https://github.com/user-attachments/assets/426629b0-3cbe-446c-95bd-26324d2620dc" />
I do not want to try to patch this on top:
* Forking Trainer manually and running evaluation with a subprocess
* Spawning external watchdog scripts for async eval
* Deep copying the model to avoid disk writes
This seems like a natural fit for `Trainer`, similar to how `save_strategy` and `log_strategy` are already modular.
### Your contribution
Some ideas (not prescriptive):
* `eval_async=True`: runs evaluation in a forked process/thread
* `eval_device='cuda:1'`: specify a device for async eval
* `eval_strategy="async_steps"`: triggers parallel eval on step intervals
* Provide a callback hook or scheduler for async checkpoint evaluation
Implementation Ideas
* Use `torch.multiprocessing` or `concurrent.futures.ProcessPoolExecutor` to fork evaluation subprocess
* Snapshot the model state (via `state_dict()` or full checkpoint) and transfer it (memory, pipe, or fast serialization)
* Respect user-specified `eval_steps`, `metric_for_best_model`, and `load_best_model_at_end` behavior | null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38829/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/38829/timeline | null | null | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | false |
https://api.github.com/repos/huggingface/transformers/issues/38828 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38828/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38828/comments | https://api.github.com/repos/huggingface/transformers/issues/38828/events | https://github.com/huggingface/transformers/pull/38828 | 3,147,024,756 | PR_kwDOCUB6oc6akPE2 | 38,828 | Fix broken tag in Longformer model card | {
"login": "dross20",
"id": 73395516,
"node_id": "MDQ6VXNlcjczMzk1NTE2",
"avatar_url": "https://avatars.githubusercontent.com/u/73395516?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dross20",
"html_url": "https://github.com/dross20",
"followers_url": "https://api.github.com/users/dross20/followers",
"following_url": "https://api.github.com/users/dross20/following{/other_user}",
"gists_url": "https://api.github.com/users/dross20/gists{/gist_id}",
"starred_url": "https://api.github.com/users/dross20/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/dross20/subscriptions",
"organizations_url": "https://api.github.com/users/dross20/orgs",
"repos_url": "https://api.github.com/users/dross20/repos",
"events_url": "https://api.github.com/users/dross20/events{/privacy}",
"received_events_url": "https://api.github.com/users/dross20/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-15T03:04:57 | 2025-06-16T14:44:41 | 2025-06-16T14:44:40 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38828",
"html_url": "https://github.com/huggingface/transformers/pull/38828",
"diff_url": "https://github.com/huggingface/transformers/pull/38828.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38828.patch",
"merged_at": "2025-06-16T14:44:40"
} | # What does this PR do?
Fixes a broken closing tag in the Longformer model card. The broken tag caused the code examples to display incorrectly on the [model card page](https://huggingface.co/docs/transformers/main/en/model_doc/longformer).
## Before submitting
- [x] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
## Who can review?
@stevhliu | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38828/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/38828/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38827 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38827/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38827/comments | https://api.github.com/repos/huggingface/transformers/issues/38827/events | https://github.com/huggingface/transformers/pull/38827 | 3,146,975,902 | PR_kwDOCUB6oc6akHWZ | 38,827 | Allow customization of sdpa in executorch.py | {
"login": "kimishpatel",
"id": 17488981,
"node_id": "MDQ6VXNlcjE3NDg4OTgx",
"avatar_url": "https://avatars.githubusercontent.com/u/17488981?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/kimishpatel",
"html_url": "https://github.com/kimishpatel",
"followers_url": "https://api.github.com/users/kimishpatel/followers",
"following_url": "https://api.github.com/users/kimishpatel/following{/other_user}",
"gists_url": "https://api.github.com/users/kimishpatel/gists{/gist_id}",
"starred_url": "https://api.github.com/users/kimishpatel/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/kimishpatel/subscriptions",
"organizations_url": "https://api.github.com/users/kimishpatel/orgs",
"repos_url": "https://api.github.com/users/kimishpatel/repos",
"events_url": "https://api.github.com/users/kimishpatel/events{/privacy}",
"received_events_url": "https://api.github.com/users/kimishpatel/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-15T01:53:03 | 2025-06-17T08:38:21 | 2025-06-17T08:38:20 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38827",
"html_url": "https://github.com/huggingface/transformers/pull/38827",
"diff_url": "https://github.com/huggingface/transformers/pull/38827.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38827.patch",
"merged_at": "2025-06-17T08:38:20"
} | # What does this PR do?
Earlier PR put executorch specific sdpa and mask function in the export function. This prevent any customization that can be done to sdpa, prior to export. By moving this to __init__, we still keep the original behavior but allow users like optimum-executorch to override sdpa by setting model.config._attn_implementation.
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
Fixes # (issue)
optimum-executorch to export with custom sdpa. Mainly this `optimum-cli export executorch --model "HuggingFaceTB/SmolLM2-135M" --task "text-generation" --recipe "xnnpack" --output_dir="smollm" --use_custom_sdpa` in optimum-executorch should work (although it needs this pr + https://github.com/huggingface/optimum-executorch/pull/82
## 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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38827/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/38827/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38826 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38826/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38826/comments | https://api.github.com/repos/huggingface/transformers/issues/38826/events | https://github.com/huggingface/transformers/pull/38826 | 3,145,927,693 | PR_kwDOCUB6oc6ahJIH | 38,826 | Fix redundant code in Janus | {
"login": "yaswanth19",
"id": 82788246,
"node_id": "MDQ6VXNlcjgyNzg4MjQ2",
"avatar_url": "https://avatars.githubusercontent.com/u/82788246?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/yaswanth19",
"html_url": "https://github.com/yaswanth19",
"followers_url": "https://api.github.com/users/yaswanth19/followers",
"following_url": "https://api.github.com/users/yaswanth19/following{/other_user}",
"gists_url": "https://api.github.com/users/yaswanth19/gists{/gist_id}",
"starred_url": "https://api.github.com/users/yaswanth19/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/yaswanth19/subscriptions",
"organizations_url": "https://api.github.com/users/yaswanth19/orgs",
"repos_url": "https://api.github.com/users/yaswanth19/repos",
"events_url": "https://api.github.com/users/yaswanth19/events{/privacy}",
"received_events_url": "https://api.github.com/users/yaswanth19/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-14T09:50:47 | 2025-06-16T06:54:28 | 2025-06-16T06:53:59 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38826",
"html_url": "https://github.com/huggingface/transformers/pull/38826",
"diff_url": "https://github.com/huggingface/transformers/pull/38826.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38826.patch",
"merged_at": "2025-06-16T06:53:59"
} | Just fixes redundant code in modular file which has no affect on modelling file and the return statements felt a bit odd :sweat_smile: . | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38826/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/38826/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38825 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38825/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38825/comments | https://api.github.com/repos/huggingface/transformers/issues/38825/events | https://github.com/huggingface/transformers/pull/38825 | 3,145,452,033 | PR_kwDOCUB6oc6affU2 | 38,825 | asdfBranch name | {
"login": "gspeter-max",
"id": 193389584,
"node_id": "U_kgDOC4bkEA",
"avatar_url": "https://avatars.githubusercontent.com/u/193389584?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/gspeter-max",
"html_url": "https://github.com/gspeter-max",
"followers_url": "https://api.github.com/users/gspeter-max/followers",
"following_url": "https://api.github.com/users/gspeter-max/following{/other_user}",
"gists_url": "https://api.github.com/users/gspeter-max/gists{/gist_id}",
"starred_url": "https://api.github.com/users/gspeter-max/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/gspeter-max/subscriptions",
"organizations_url": "https://api.github.com/users/gspeter-max/orgs",
"repos_url": "https://api.github.com/users/gspeter-max/repos",
"events_url": "https://api.github.com/users/gspeter-max/events{/privacy}",
"received_events_url": "https://api.github.com/users/gspeter-max/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-14T05:30:12 | 2025-06-16T16:22:43 | 2025-06-16T13:11:29 | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38825",
"html_url": "https://github.com/huggingface/transformers/pull/38825",
"diff_url": "https://github.com/huggingface/transformers/pull/38825.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38825.patch",
"merged_at": null
} | # What does this PR do?
<!--asdf
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38825/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/38825/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38824 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38824/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38824/comments | https://api.github.com/repos/huggingface/transformers/issues/38824/events | https://github.com/huggingface/transformers/pull/38824 | 3,145,422,321 | PR_kwDOCUB6oc6afZfL | 38,824 | Add `SepCache` [An efficient and easy-to-use Cache from the SepLLM paper - ICML 2025 (https://arxiv.org/abs/2412.12094) ] to the `cache_utils.py` and `__init__.py` | {
"login": "GaussonTschen",
"id": 191336691,
"node_id": "U_kgDOC2eQ8w",
"avatar_url": "https://avatars.githubusercontent.com/u/191336691?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/GaussonTschen",
"html_url": "https://github.com/GaussonTschen",
"followers_url": "https://api.github.com/users/GaussonTschen/followers",
"following_url": "https://api.github.com/users/GaussonTschen/following{/other_user}",
"gists_url": "https://api.github.com/users/GaussonTschen/gists{/gist_id}",
"starred_url": "https://api.github.com/users/GaussonTschen/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/GaussonTschen/subscriptions",
"organizations_url": "https://api.github.com/users/GaussonTschen/orgs",
"repos_url": "https://api.github.com/users/GaussonTschen/repos",
"events_url": "https://api.github.com/users/GaussonTschen/events{/privacy}",
"received_events_url": "https://api.github.com/users/GaussonTschen/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-14T04:53:32 | 2025-08-05T22:51:51 | 2025-08-05T22:51:51 | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38824",
"html_url": "https://github.com/huggingface/transformers/pull/38824",
"diff_url": "https://github.com/huggingface/transformers/pull/38824.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38824.patch",
"merged_at": null
} | Add `SepCache` [An efficient and easy-to-use Cache from the SepLLM paper - ICML 2025 (https://arxiv.org/abs/2412.12094) ] to the `cache_utils.py` and `__init__.py`
A cache as described in the [SepLLM paper - ICML 2025](https://arxiv.org/abs/2412.12094). In the training phase,
SepLLM condenses the segment information into the KV of the separator that divides the segment. In the inference phase, the
corresponding SepCache only needs to store the KVs of initial tokens, separator tokens, and recent tokens for generation.
It stores the Key and Value states as lists of tensors, two lists for each layer. The expected shape for each tensor is
`[batch_size, num_heads, seq_len, head_dim]`.
Frequently-Used Parameters:
`init_cache_size: Union[int, List]`:
The maximum number of KVs to be stored for initial tokens.
`sep_cache_size: Union[int, List]`:
The maximum number of KVs to be stored for separator tokens.
`local_size: Union[int, List]`:
The maximum number of KVs to be stored for local tokens (i.e., sliding window).
`cache_size: Union[int, List]`:
The maximum number of KVs to be stored for all the tokens, i.e., the size for the whole KV cache.
Concerning these four parameters above:
When a list is passed (its length must be `layer_num`), it represents different values for each layer.
When an integer is passed, it means the setting is the same for all layers.
`separator_token_ids: List[int]`:
The token ids of the separator tokens for the current model's tokenizer.
We have some examples, such as the Llama-3 series models, where setting `model_type='llama'` allows you
to skip setting `separator_token_ids` and `PADDING_ID` (SepCache will auto-fill them).
`PADDING_ID: int`:
The token id of the padding token. You can just set `PADDING_ID` to the id of "<|endoftext|>" token of the tokenizer for the pretrained model.
Example:
```python
>>> from transformers import AutoTokenizer, AutoModelForCausalLM, SepCache
>>> import torch
>>> from huggingface_hub import login
>>> login("hf_xxxXXXxxx")
>>> def to_cuda(a_dict: dict) -> dict:
>>> new_dict = {}
>>> for k,v in a_dict.items():
>>> if isinstance(v, torch.Tensor):
>>> new_dict[k] = v.cuda()
>>> else:
>>> new_dict[k] = v
>>> return new_dict
>>> model = AutoModelForCausalLM.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct", attn_implementation="flash_attention_2", device_map="cuda:0")
>>> model.bfloat16().cuda()
>>> tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct")
>>> inputs = tokenizer(text="My name is Llama 3", return_tensors="pt")
>>> inputs = to_cuda(inputs)
>>> # Prepare a cache and pass it to model's forward; `layer_num` is the number of layers for the pretrained model.
>>> past_key_values = SepCache(init_cache_size=4, sep_cache_size=64, local_size=256, cache_size=512, layer_num=32, USE_MAX_SEP_CACHE=True, model_type='llama')
>>> # `separator_token_ids` and `PADDING_ID` must also be provided if you are not using `model_type='llama'` like this demo.
>>> outputs = model(**inputs, past_key_values=past_key_values, use_cache=True)
>>> outputs.past_key_values # access SepCache filled with keys/values
SepCache()
```
```python
>>> ## When using the `update` function of SepCache to update the keys/values and the past token ids (necessary in SepCache), the current `input_ids` must also be provided.
>>> key_states, value_states = past_key_values.update(
key_states = key_states,
value_states = value_states,
input_ids = input_ids,
layer_idx = layer_idx,
PREFILLING_FLAG = q_len > 1, ## `q_len` is the sequence length of the current `query_states`
)
```
For detailed usage instructions, please refer to https://github.com/HKUDS/SepLLM
@ArthurZucker @gante
<!-- Remove if not applicable -->
## 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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "GaussonTschen",
"id": 191336691,
"node_id": "U_kgDOC2eQ8w",
"avatar_url": "https://avatars.githubusercontent.com/u/191336691?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/GaussonTschen",
"html_url": "https://github.com/GaussonTschen",
"followers_url": "https://api.github.com/users/GaussonTschen/followers",
"following_url": "https://api.github.com/users/GaussonTschen/following{/other_user}",
"gists_url": "https://api.github.com/users/GaussonTschen/gists{/gist_id}",
"starred_url": "https://api.github.com/users/GaussonTschen/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/GaussonTschen/subscriptions",
"organizations_url": "https://api.github.com/users/GaussonTschen/orgs",
"repos_url": "https://api.github.com/users/GaussonTschen/repos",
"events_url": "https://api.github.com/users/GaussonTschen/events{/privacy}",
"received_events_url": "https://api.github.com/users/GaussonTschen/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38824/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/38824/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38823 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38823/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38823/comments | https://api.github.com/repos/huggingface/transformers/issues/38823/events | https://github.com/huggingface/transformers/issues/38823 | 3,145,268,286 | I_kwDOCUB6oc67ePw- | 38,823 | Cannot create a 2-class Deformable DETR model from a 1-class model | {
"login": "jveitchmichaelis",
"id": 3159591,
"node_id": "MDQ6VXNlcjMxNTk1OTE=",
"avatar_url": "https://avatars.githubusercontent.com/u/3159591?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jveitchmichaelis",
"html_url": "https://github.com/jveitchmichaelis",
"followers_url": "https://api.github.com/users/jveitchmichaelis/followers",
"following_url": "https://api.github.com/users/jveitchmichaelis/following{/other_user}",
"gists_url": "https://api.github.com/users/jveitchmichaelis/gists{/gist_id}",
"starred_url": "https://api.github.com/users/jveitchmichaelis/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/jveitchmichaelis/subscriptions",
"organizations_url": "https://api.github.com/users/jveitchmichaelis/orgs",
"repos_url": "https://api.github.com/users/jveitchmichaelis/repos",
"events_url": "https://api.github.com/users/jveitchmichaelis/events{/privacy}",
"received_events_url": "https://api.github.com/users/jveitchmichaelis/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-14T01:54:15 | 2025-06-14T02:23:44 | 2025-06-14T02:23:44 | NONE | null | null | null | null |
I have a test case that instantiates a model with a number of class count options e.g. (1,2,10). The source model is trained on a single class. `ignore_mismatched_sizes` is set to `True`. If I change `num_classes` to 2, the model load fails. If I change it to anything else, it seems to be fine. The error seems to be in loading a bias term:
```
File ~/miniconda3/envs/x/lib/python3.12/site-packages/transformers/modeling_utils.py:731, in _load_parameter_into_model(model, param_name, tensor)
729 module, param_type = get_module_from_name(model, param_name)
730 # This will check potential shape mismatch if skipped before
--> 731 module.load_state_dict({param_type: tensor}, strict=False, assign=True)
File ~/miniconda3/envs/x/lib/python3.12/site-packages/torch/nn/modules/module.py:2593, in Module.load_state_dict(self, state_dict, strict, assign)
2585 error_msgs.insert(
2586 0,
2587 "Missing key(s) in state_dict: {}. ".format(
2588 ", ".join(f'"{k}"' for k in missing_keys)
2589 ),
2590 )
2592 if len(error_msgs) > 0:
-> 2593 raise RuntimeError(
2594 "Error(s) in loading state_dict for {}:\n\t{}".format(
2595 self.__class__.__name__, "\n\t".join(error_msgs)
2596 )
2597 )
2598 return _IncompatibleKeys(missing_keys, unexpected_keys)
RuntimeError: Error(s) in loading state_dict for Linear:
size mismatch for bias: copying a param with shape torch.Size([1]) from checkpoint, the shape in current model is torch.Size([2]).
```
Possibly related to https://github.com/huggingface/transformers/issues/36960. Doing some digging it seems like this was fixed a couple of weeks ago, will try from main.
### System Info
Env:
```
- `transformers` version: 4.52.4
- Platform: macOS-15.4.1-arm64-arm-64bit
- Python version: 3.12.3
- Huggingface_hub version: 0.31.2
- Safetensors version: 0.5.3
- Accelerate version: not installed
- Accelerate config: not found
- DeepSpeed version: not installed
- PyTorch version (GPU?): 2.7.0 (False)
- Tensorflow version (GPU?): not installed (NA)
- Flax version (CPU?/GPU?/TPU?): not installed (NA)
- Jax version: not installed
- JaxLib version: not installed
- Using distributed or parallel set-up in script?: No
```
### Who can help?
@amyeroberts @qubvel
### Information
- [x] The official example scripts
- [ ] 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
This example makes a 1-class model and then attempts to convert it to 2.
This fails:
```python
import tempfile
from transformers import DeformableDetrForObjectDetection, DeformableDetrConfig
with tempfile.TemporaryDirectory() as dirname:
configuration = DeformableDetrConfig(num_labels=1)
model = DeformableDetrForObjectDetection(configuration)
model.save_pretrained(dirname)
model = DeformableDetrForObjectDetection.from_pretrained(
dirname,
num_labels=2,
ignore_mismatched_sizes=True
)
```
```bash
File ~/miniconda3/envs/x/lib/python3.12/site-packages/torch/nn/modules/module.py:2593, in Module.load_state_dict(self, state_dict, strict, assign)
2585 error_msgs.insert(
2586 0,
2587 "Missing key(s) in state_dict: {}. ".format(
2588 ", ".join(f'"{k}"' for k in missing_keys)
2589 ),
2590 )
2592 if len(error_msgs) > 0:
-> 2593 raise RuntimeError(
2594 "Error(s) in loading state_dict for {}:\n\t{}".format(
2595 self.__class__.__name__, "\n\t".join(error_msgs)
2596 )
2597 )
2598 return _IncompatibleKeys(missing_keys, unexpected_keys)
RuntimeError: Error(s) in loading state_dict for Linear:
size mismatch for bias: copying a param with shape torch.Size([1]) from checkpoint, the shape in current model is torch.Size([2]).
```
Oddly, converting to a 10 class model is fine:
```python
with tempfile.TemporaryDirectory() as dirname:
configuration = DeformableDetrConfig(num_labels=1)
model = DeformableDetrForObjectDetection(configuration)
model.save_pretrained(dirname)
model = DeformableDetrForObjectDetection.from_pretrained(
dirname,
num_labels=10,
ignore_mismatched_sizes=True
)
```
```bash
Some weights of DeformableDetrForObjectDetection were not initialized from the model checkpoint at /var/folders/xp/26mwm6b951l39z_tvs352_c80000gn/T/tmpf8nrmwwy and are newly initialized because the shapes did not match:
- class_embed.0.bias: found shape torch.Size([1]) in the checkpoint and torch.Size([10]) in the model instantiated
- class_embed.0.weight: found shape torch.Size([1, 256]) in the checkpoint and torch.Size([10, 256]) in the model instantiated
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
```
It's also fine if we start with a different number of labels, which suggests it's specifically 1 -> 2 that's an issue?
```python
with tempfile.TemporaryDirectory() as dirname:
configuration = DeformableDetrConfig(num_labels=10)
model = DeformableDetrForObjectDetection(configuration)
model.save_pretrained(dirname)
model = DeformableDetrForObjectDetection.from_pretrained(
dirname,
num_labels=2,
ignore_mismatched_sizes=True
)
```
```bash
Some weights of DeformableDetrForObjectDetection were not initialized from the model checkpoint at /var/folders/xp/26mwm6b951l39z_tvs352_c80000gn/T/tmpwmopo1nf and are newly initialized because the shapes did not match:
- class_embed.0.bias: found shape torch.Size([10]) in the checkpoint and torch.Size([2]) in the model instantiated
- class_embed.0.weight: found shape torch.Size([10, 256]) in the checkpoint and torch.Size([2, 256]) in the model instantiated
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
```
Detr doesn't have this problem, so it seems to be an edge case in DeformableDetr specifically. Though here the class_labels_classifier bias goes from 2 -> 3.
```python
from transformers import DetrForObjectDetection, DetrConfig
with tempfile.TemporaryDirectory() as dirname:
configuration = DetrConfig(num_labels=1)
model = DetrForObjectDetection(configuration)
model.save_pretrained(dirname)
model = DetrForObjectDetection.from_pretrained(
dirname,
num_labels=2,
ignore_mismatched_sizes=True
)
```
```bash
Some weights of DetrForObjectDetection were not initialized from the model checkpoint at /var/folders/xp/26mwm6b951l39z_tvs352_c80000gn/T/tmp4yc1m7jh and are newly initialized because the shapes did not match:
- class_labels_classifier.bias: found shape torch.Size([2]) in the checkpoint and torch.Size([3]) in the model instantiated
- class_labels_classifier.weight: found shape torch.Size([2, 256]) in the checkpoint and torch.Size([3, 256]) in the model instantiated
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
```
### Expected behavior
`from_pretrained` should be able to handle this and return a model initialised with a 2-class detection head as requested. | {
"login": "jveitchmichaelis",
"id": 3159591,
"node_id": "MDQ6VXNlcjMxNTk1OTE=",
"avatar_url": "https://avatars.githubusercontent.com/u/3159591?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jveitchmichaelis",
"html_url": "https://github.com/jveitchmichaelis",
"followers_url": "https://api.github.com/users/jveitchmichaelis/followers",
"following_url": "https://api.github.com/users/jveitchmichaelis/following{/other_user}",
"gists_url": "https://api.github.com/users/jveitchmichaelis/gists{/gist_id}",
"starred_url": "https://api.github.com/users/jveitchmichaelis/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/jveitchmichaelis/subscriptions",
"organizations_url": "https://api.github.com/users/jveitchmichaelis/orgs",
"repos_url": "https://api.github.com/users/jveitchmichaelis/repos",
"events_url": "https://api.github.com/users/jveitchmichaelis/events{/privacy}",
"received_events_url": "https://api.github.com/users/jveitchmichaelis/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38823/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/38823/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38822 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38822/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38822/comments | https://api.github.com/repos/huggingface/transformers/issues/38822/events | https://github.com/huggingface/transformers/pull/38822 | 3,144,967,151 | PR_kwDOCUB6oc6ad4x5 | 38,822 | Corrections to PR #38642 and enhancements to Wav2Vec2Processor __call__ and pad docstrings | {
"login": "renet10",
"id": 12860891,
"node_id": "MDQ6VXNlcjEyODYwODkx",
"avatar_url": "https://avatars.githubusercontent.com/u/12860891?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/renet10",
"html_url": "https://github.com/renet10",
"followers_url": "https://api.github.com/users/renet10/followers",
"following_url": "https://api.github.com/users/renet10/following{/other_user}",
"gists_url": "https://api.github.com/users/renet10/gists{/gist_id}",
"starred_url": "https://api.github.com/users/renet10/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/renet10/subscriptions",
"organizations_url": "https://api.github.com/users/renet10/orgs",
"repos_url": "https://api.github.com/users/renet10/repos",
"events_url": "https://api.github.com/users/renet10/events{/privacy}",
"received_events_url": "https://api.github.com/users/renet10/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T22:33:54 | 2025-07-18T14:36:56 | 2025-07-16T21:13:07 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38822",
"html_url": "https://github.com/huggingface/transformers/pull/38822",
"diff_url": "https://github.com/huggingface/transformers/pull/38822.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38822.patch",
"merged_at": "2025-07-16T21:13:07"
} | # What does this PR do?
The modifications to Wav2Vec2Processor __call__ and pad docstrings adds back the reference of PreTrainedTokenizer __call__ and pad removed incorrectly in PR #38642. As the __call__ and pad methods are important public methods for Wav2Vec2Processor, this PR also enhances the docstrings with the invocation procedure post removal of "as_target_processor()" and adds information on return values.
Fixes # 38609
## 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#create-a-pull-request),
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.
- [ ] 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?
@stevhliu | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38822/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/38822/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38820 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38820/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38820/comments | https://api.github.com/repos/huggingface/transformers/issues/38820/events | https://github.com/huggingface/transformers/issues/38820 | 3,144,755,851 | I_kwDOCUB6oc67cSqL | 38,820 | CI skipped failures tracking issue | {
"login": "ivarflakstad",
"id": 69173633,
"node_id": "MDQ6VXNlcjY5MTczNjMz",
"avatar_url": "https://avatars.githubusercontent.com/u/69173633?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ivarflakstad",
"html_url": "https://github.com/ivarflakstad",
"followers_url": "https://api.github.com/users/ivarflakstad/followers",
"following_url": "https://api.github.com/users/ivarflakstad/following{/other_user}",
"gists_url": "https://api.github.com/users/ivarflakstad/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ivarflakstad/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ivarflakstad/subscriptions",
"organizations_url": "https://api.github.com/users/ivarflakstad/orgs",
"repos_url": "https://api.github.com/users/ivarflakstad/repos",
"events_url": "https://api.github.com/users/ivarflakstad/events{/privacy}",
"received_events_url": "https://api.github.com/users/ivarflakstad/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | {
"login": "ivarflakstad",
"id": 69173633,
"node_id": "MDQ6VXNlcjY5MTczNjMz",
"avatar_url": "https://avatars.githubusercontent.com/u/69173633?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ivarflakstad",
"html_url": "https://github.com/ivarflakstad",
"followers_url": "https://api.github.com/users/ivarflakstad/followers",
"following_url": "https://api.github.com/users/ivarflakstad/following{/other_user}",
"gists_url": "https://api.github.com/users/ivarflakstad/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ivarflakstad/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ivarflakstad/subscriptions",
"organizations_url": "https://api.github.com/users/ivarflakstad/orgs",
"repos_url": "https://api.github.com/users/ivarflakstad/repos",
"events_url": "https://api.github.com/users/ivarflakstad/events{/privacy}",
"received_events_url": "https://api.github.com/users/ivarflakstad/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"login": "ivarflakstad",
"id": 69173633,
"node_id": "MDQ6VXNlcjY5MTczNjMz",
"avatar_url": "https://avatars.githubusercontent.com/u/69173633?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ivarflakstad",
"html_url": "https://github.com/ivarflakstad",
"followers_url": "https://api.github.com/users/ivarflakstad/followers",
"following_url": "https://api.github.com/users/ivarflakstad/following{/other_user}",
"gists_url": "https://api.github.com/users/ivarflakstad/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ivarflakstad/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ivarflakstad/subscriptions",
"organizations_url": "https://api.github.com/users/ivarflakstad/orgs",
"repos_url": "https://api.github.com/users/ivarflakstad/repos",
"events_url": "https://api.github.com/users/ivarflakstad/events{/privacy}",
"received_events_url": "https://api.github.com/users/ivarflakstad/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
}
] | null | [] | 2025-06-13T20:56:55 | 2025-07-23T08:02:39 | 2025-07-23T08:02:39 | MEMBER | null | null | null | null | Some of the tests in our scheduled CI's are skipped because they require newer hardware (higher compute capability) than the setup we are currently using. That's not a problem on it's own, but after manually checking we have found that there are actually quite a few of these tests that fail on the hardware that can run said tests.
This issue is to be used to track work related to this category of tests.
The first step will be to skip said tests in the CI and add a reference to this issue in the reason for skipping. | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38820/reactions",
"total_count": 1,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 1
} | https://api.github.com/repos/huggingface/transformers/issues/38820/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38819 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38819/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38819/comments | https://api.github.com/repos/huggingface/transformers/issues/38819/events | https://github.com/huggingface/transformers/pull/38819 | 3,144,382,949 | PR_kwDOCUB6oc6ab2pg | 38,819 | Update trainer.py: add multiprocessing_context for mps devices | {
"login": "AmitMY",
"id": 5757359,
"node_id": "MDQ6VXNlcjU3NTczNTk=",
"avatar_url": "https://avatars.githubusercontent.com/u/5757359?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/AmitMY",
"html_url": "https://github.com/AmitMY",
"followers_url": "https://api.github.com/users/AmitMY/followers",
"following_url": "https://api.github.com/users/AmitMY/following{/other_user}",
"gists_url": "https://api.github.com/users/AmitMY/gists{/gist_id}",
"starred_url": "https://api.github.com/users/AmitMY/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/AmitMY/subscriptions",
"organizations_url": "https://api.github.com/users/AmitMY/orgs",
"repos_url": "https://api.github.com/users/AmitMY/repos",
"events_url": "https://api.github.com/users/AmitMY/events{/privacy}",
"received_events_url": "https://api.github.com/users/AmitMY/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-13T18:20:35 | 2025-06-13T18:20:35 | null | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38819",
"html_url": "https://github.com/huggingface/transformers/pull/38819",
"diff_url": "https://github.com/huggingface/transformers/pull/38819.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38819.patch",
"merged_at": null
} | # What does this PR do?
I have encountered an error during training on an M1 mac, similar to:
https://github.com/pytorch/pytorch/issues/87688
I have implemented the suggested fix, and it works.
## 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#create-a-pull-request),
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?
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38819/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/38819/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38818 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38818/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38818/comments | https://api.github.com/repos/huggingface/transformers/issues/38818/events | https://github.com/huggingface/transformers/pull/38818 | 3,144,286,910 | PR_kwDOCUB6oc6abhbl | 38,818 | [Feature] Support `is_split_into_words` in the `TokenClassificationPipeline`. | {
"login": "Vixel2006",
"id": 166058059,
"node_id": "U_kgDOCeXYSw",
"avatar_url": "https://avatars.githubusercontent.com/u/166058059?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Vixel2006",
"html_url": "https://github.com/Vixel2006",
"followers_url": "https://api.github.com/users/Vixel2006/followers",
"following_url": "https://api.github.com/users/Vixel2006/following{/other_user}",
"gists_url": "https://api.github.com/users/Vixel2006/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Vixel2006/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Vixel2006/subscriptions",
"organizations_url": "https://api.github.com/users/Vixel2006/orgs",
"repos_url": "https://api.github.com/users/Vixel2006/repos",
"events_url": "https://api.github.com/users/Vixel2006/events{/privacy}",
"received_events_url": "https://api.github.com/users/Vixel2006/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T17:42:37 | 2025-06-23T15:37:43 | 2025-06-23T15:31:32 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38818",
"html_url": "https://github.com/huggingface/transformers/pull/38818",
"diff_url": "https://github.com/huggingface/transformers/pull/38818.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38818.patch",
"merged_at": "2025-06-23T15:31:32"
} | # What does this PR do?
Fixes #30757
This PR solves the problem that when using an already tokenized input in the `TokenClassificationPipeline` by adding a `tokenizer_parameter` called `is_split_into_words`. I think we will need the aggregation code because some tokenizers are subword tokenizers, I have solved the problem that `start` and `end` was not correctly assigned when using the pipeline by reconstructing the sentence and create a map that stores the position of each word and pass it along with `words_id` to the `postprocess` and in `gather_pre_entities`, we use the map and `words_id` to produce the final position.
```py
from transformers import pipeline
from datasets import load_dataset
dataset = load_dataset("conll2003", split="test[:5]")
ner_pipeline = pipeline("token-classification", model="dslim/bert-base-NER", aggregation_strategy="first")
for example in dataset:
tokens = example['tokens']
predictions = ner_pipeline(tokens, is_split_into_words=True)
for entity in predictions:
print(entity)
```
Now this example doing inference on a batch of a preprocessed dataset. it gives this output.
```
[{'entity_group': 'MISC', 'score': np.float32(0.48014355), 'word': 'JAPAN', 'start': 9, 'end': 14}, {'entity_group': 'PER', 'score': np.float32(0.43156916), 'word': 'LUCKY', 'start': 19, 'end': 24}, {'entity_group': 'ORG', 'score': np.float32(0.64400136), 'word': 'CHINA', 'start': 31, 'end': 36}]
[{'entity_group': 'PER', 'score': np.float32(0.9979421), 'word': 'Nadim Ladki', 'start': 0, 'end': 11}]
[{'entity_group': 'LOC', 'score': np.float32(0.9851431), 'word': 'AL - AIN', 'start': 0, 'end': 6}, {'entity_group': 'LOC', 'score': np.float32(0.99941224), 'word': 'United Arab Emirates', 'start': 9, 'end': 29}]
[{'entity_group': 'LOC', 'score': np.float32(0.9998572), 'word': 'Japan', 'start': 0, 'end': 5}, {'entity_group': 'MISC', 'score': np.float32(0.99814206), 'word': 'Asian Cup', 'start': 33, 'end': 42}, {'entity_group': 'LOC', 'score': np.float32(0.9997607), 'word': 'Syria', 'start': 78, 'end': 83}, {'entity_group': 'MISC', 'score': np.float32(0.77703416), 'word': 'Group C', 'start': 89, 'end': 96}]
[{'entity_group': 'LOC', 'score': np.float32(0.9998673), 'word': 'China', 'start': 4, 'end': 9}, {'entity_group': 'LOC', 'score': np.float32(0.9997266), 'word': 'Uzbekistan', 'start': 119, 'end': 129}]
```
## Who can review?
@Rocketknight1 @ArthurZucker
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38818/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/38818/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38817 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38817/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38817/comments | https://api.github.com/repos/huggingface/transformers/issues/38817/events | https://github.com/huggingface/transformers/pull/38817 | 3,143,880,157 | PR_kwDOCUB6oc6aaImA | 38,817 | Use `wandb.run.url` instead of `wandb.run.get_url()` (deprecated) | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T15:06:44 | 2025-06-13T15:20:36 | 2025-06-13T15:20:04 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38817",
"html_url": "https://github.com/huggingface/transformers/pull/38817",
"diff_url": "https://github.com/huggingface/transformers/pull/38817.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38817.patch",
"merged_at": "2025-06-13T15:20:04"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "SunMarc",
"id": 57196510,
"node_id": "MDQ6VXNlcjU3MTk2NTEw",
"avatar_url": "https://avatars.githubusercontent.com/u/57196510?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SunMarc",
"html_url": "https://github.com/SunMarc",
"followers_url": "https://api.github.com/users/SunMarc/followers",
"following_url": "https://api.github.com/users/SunMarc/following{/other_user}",
"gists_url": "https://api.github.com/users/SunMarc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SunMarc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SunMarc/subscriptions",
"organizations_url": "https://api.github.com/users/SunMarc/orgs",
"repos_url": "https://api.github.com/users/SunMarc/repos",
"events_url": "https://api.github.com/users/SunMarc/events{/privacy}",
"received_events_url": "https://api.github.com/users/SunMarc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38817/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/38817/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38816 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38816/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38816/comments | https://api.github.com/repos/huggingface/transformers/issues/38816/events | https://github.com/huggingface/transformers/pull/38816 | 3,143,834,808 | PR_kwDOCUB6oc6aZ-zo | 38,816 | Fix `phi4_multimodal` tests | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T14:51:19 | 2025-06-18T07:39:20 | 2025-06-18T07:39:18 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38816",
"html_url": "https://github.com/huggingface/transformers/pull/38816",
"diff_url": "https://github.com/huggingface/transformers/pull/38816.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38816.patch",
"merged_at": "2025-06-18T07:39:18"
} | # What does this PR do?
Use revision as discussed offline (as [Hub PR](https://huggingface.co/microsoft/Phi-4-multimodal-instruct/discussions/70) is not merged) and also updated some expected values + one `require_torch_large_accelerator`.
All passing on T4/A10 with torch 2.7.1 | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38816/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/38816/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38815 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38815/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38815/comments | https://api.github.com/repos/huggingface/transformers/issues/38815/events | https://github.com/huggingface/transformers/pull/38815 | 3,143,642,843 | PR_kwDOCUB6oc6aZV3V | 38,815 | Fix a minor security issue | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T13:44:21 | 2025-06-13T15:37:48 | 2025-06-13T15:37:46 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38815",
"html_url": "https://github.com/huggingface/transformers/pull/38815",
"diff_url": "https://github.com/huggingface/transformers/pull/38815.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38815.patch",
"merged_at": "2025-06-13T15:37:46"
} | # What does this PR do?
There is an security report about this. It's not a big deal as `utils/check_self_hosted_runner.py` is only used in some AMD workflow files that could be only triggered by push event `.github/workflows/self-push-amd.yml` (not really used) or in
https://github.com/huggingface/hf-workflows/blob/main/.github/workflows/transformers_amd_ci_scheduled.yaml
But let's still update it anyway.
https://docs.python.org/3/library/subprocess.html
(search ` Unless otherwise stated, it is recommended to pass args as a sequence.`) | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38815/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/38815/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38814 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38814/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38814/comments | https://api.github.com/repos/huggingface/transformers/issues/38814/events | https://github.com/huggingface/transformers/pull/38814 | 3,143,553,787 | PR_kwDOCUB6oc6aZCMa | 38,814 | 🎯 Trackio integration | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T13:16:53 | 2025-07-25T11:42:32 | 2025-07-22T21:50:20 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38814",
"html_url": "https://github.com/huggingface/transformers/pull/38814",
"diff_url": "https://github.com/huggingface/transformers/pull/38814.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38814.patch",
"merged_at": "2025-07-22T21:50:20"
} | # What does this PR do?
```python
from trl import SFTTrainer, SFTConfig
from datasets import load_dataset
trainer = SFTTrainer(
model="Qwen/Qwen2.5-0.5B",
args=SFTConfig(report_to="trackio"),
train_dataset=load_dataset("trl-lib/Capybara", split="train"),
)
trainer.train()
```
<img width="1689" alt="Screenshot 2025-06-13 at 3 16 20 PM" src="https://github.com/user-attachments/assets/c34db895-b0e2-45c7-96ac-1610992fde30" />
## 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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38814/reactions",
"total_count": 4,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 4,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/38814/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38813 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38813/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38813/comments | https://api.github.com/repos/huggingface/transformers/issues/38813/events | https://github.com/huggingface/transformers/pull/38813 | 3,143,513,047 | PR_kwDOCUB6oc6aY5O6 | 38,813 | Fix `llava_next` tests | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T13:02:02 | 2025-06-13T13:19:43 | 2025-06-13T13:19:41 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38813",
"html_url": "https://github.com/huggingface/transformers/pull/38813",
"diff_url": "https://github.com/huggingface/transformers/pull/38813.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38813.patch",
"merged_at": "2025-06-13T13:19:41"
} | # What does this PR do?
Update expected values + a few dtype / device issue | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38813/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/38813/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38812 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38812/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38812/comments | https://api.github.com/repos/huggingface/transformers/issues/38812/events | https://github.com/huggingface/transformers/pull/38812 | 3,143,477,859 | PR_kwDOCUB6oc6aYxbW | 38,812 | [configuration] remove redundant `classmethod` | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T12:49:50 | 2025-07-28T10:38:49 | 2025-07-28T10:38:49 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38812",
"html_url": "https://github.com/huggingface/transformers/pull/38812",
"diff_url": "https://github.com/huggingface/transformers/pull/38812.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38812.patch",
"merged_at": "2025-07-28T10:38:49"
} | # What does this PR do?
As per title, another tiny clean up that deletes a `from_text_vision_configs` for initializing configs from its sub-configs. That can perfectly be done by `cls.__init__` and this no need to maintain a separate method for that
I added a warning with a deprecation cycle of 7 minor releases, to give users time to adjust in case anyone was using it | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38812/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/38812/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38811 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38811/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38811/comments | https://api.github.com/repos/huggingface/transformers/issues/38811/events | https://github.com/huggingface/transformers/pull/38811 | 3,143,393,195 | PR_kwDOCUB6oc6aYe6J | 38,811 | [internvl] fix video inference | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T12:17:51 | 2025-06-16T06:37:30 | 2025-06-16T06:37:30 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38811",
"html_url": "https://github.com/huggingface/transformers/pull/38811",
"diff_url": "https://github.com/huggingface/transformers/pull/38811.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38811.patch",
"merged_at": "2025-06-16T06:37:30"
} | # What does this PR do?
Video inference broke after https://github.com/huggingface/transformers/pull/38105 because the model was inferring number of frames before processing. Previously that worked fine, but with the new workflow the actual number of frames can be inferred only after processing and sampling | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38811/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/38811/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38810 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38810/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38810/comments | https://api.github.com/repos/huggingface/transformers/issues/38810/events | https://github.com/huggingface/transformers/pull/38810 | 3,143,119,263 | PR_kwDOCUB6oc6aXi1J | 38,810 | Add kwargs support in WhisperForConditionalGeneration | {
"login": "tanuj-rai",
"id": 84439872,
"node_id": "MDQ6VXNlcjg0NDM5ODcy",
"avatar_url": "https://avatars.githubusercontent.com/u/84439872?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/tanuj-rai",
"html_url": "https://github.com/tanuj-rai",
"followers_url": "https://api.github.com/users/tanuj-rai/followers",
"following_url": "https://api.github.com/users/tanuj-rai/following{/other_user}",
"gists_url": "https://api.github.com/users/tanuj-rai/gists{/gist_id}",
"starred_url": "https://api.github.com/users/tanuj-rai/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/tanuj-rai/subscriptions",
"organizations_url": "https://api.github.com/users/tanuj-rai/orgs",
"repos_url": "https://api.github.com/users/tanuj-rai/repos",
"events_url": "https://api.github.com/users/tanuj-rai/events{/privacy}",
"received_events_url": "https://api.github.com/users/tanuj-rai/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-13T10:36:08 | 2025-06-27T13:57:22 | null | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38810",
"html_url": "https://github.com/huggingface/transformers/pull/38810",
"diff_url": "https://github.com/huggingface/transformers/pull/38810.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38810.patch",
"merged_at": null
} | # What does this PR do?
This PR update the forward method to support kwargs for flexible loss customization.
Fixes #36119
## 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#create-a-pull-request),
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?
@ArthurZucker
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38810/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/38810/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38809 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38809/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38809/comments | https://api.github.com/repos/huggingface/transformers/issues/38809/events | https://github.com/huggingface/transformers/issues/38809 | 3,142,981,340 | I_kwDOCUB6oc67Vhbc | 38,809 | Text classification evaluation has_labels bug! | {
"login": "HERIUN",
"id": 25131767,
"node_id": "MDQ6VXNlcjI1MTMxNzY3",
"avatar_url": "https://avatars.githubusercontent.com/u/25131767?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/HERIUN",
"html_url": "https://github.com/HERIUN",
"followers_url": "https://api.github.com/users/HERIUN/followers",
"following_url": "https://api.github.com/users/HERIUN/following{/other_user}",
"gists_url": "https://api.github.com/users/HERIUN/gists{/gist_id}",
"starred_url": "https://api.github.com/users/HERIUN/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/HERIUN/subscriptions",
"organizations_url": "https://api.github.com/users/HERIUN/orgs",
"repos_url": "https://api.github.com/users/HERIUN/repos",
"events_url": "https://api.github.com/users/HERIUN/events{/privacy}",
"received_events_url": "https://api.github.com/users/HERIUN/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-13T09:48:49 | 2025-06-17T05:35:16 | 2025-06-17T05:35:16 | CONTRIBUTOR | null | null | null | null | in Trainer.prediction_step(...)
```python
def prediction_step(
self,
model: nn.Module,
inputs: dict[str, Union[torch.Tensor, Any]],
prediction_loss_only: bool,
ignore_keys: Optional[list[str]] = None,
) -> tuple[Optional[torch.Tensor], Optional[torch.Tensor], Optional[torch.Tensor]]:
has_labels = False if len(self.label_names) == 0 else all(inputs.get(k) is not None for k in self.label_names)
...
```
Inputs has "labels" but has_labels=False...
compute_metic dosen't work.
<img width="1111" alt="Image" src="https://github.com/user-attachments/assets/774c2483-bac9-4437-a4a3-f1a13014ee5e" />
<img width="1367" alt="Image" src="https://github.com/user-attachments/assets/9c23bf7a-6801-4a2f-8661-29b6bb7eaf36" />
### 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
```python
from datasets import load_from_disk, load_dataset
import numpy as np
import torch
import evaluate
from transformers import AutoTokenizer
from transformers import AutoModelForSequenceClassification, ModernBertForSequenceClassification
from transformers import TrainingArguments, Trainer
import os
import warnings
warnings.filterwarnings(action='ignore')
def main():
device = torch.device('mps' if torch.backends.mps.is_available() else 'cuda' if torch.cuda.is_available() else 'cpu')
print("device: ", device)
torch.cuda.empty_cache()
# loading dataset
print("Loading dataset...")
if os.path.exists("data/tokenized/wortschartz_30/"):
datasets = load_from_disk("data/tokenized/wortschartz_30/")
else:
datasets = load_dataset("Flitto/wortschartz_31", token=os.environ.get("HF_TOKEN"))
tokenizer = AutoTokenizer.from_pretrained("answerdotai/ModernBERT-base")
tokenizer.model_max_length = 512
def tokenize_function(examples):
return tokenizer(examples['text'], padding="max_length", truncation=True)
datasets = datasets.map(
tokenize_function,
batched=True,
num_proc=16,
)
datasets.save_to_disk('data/tokenized/wortschartz_30')
print("Done")
train_size = len(datasets["train"])
valid_size = len(datasets["validation"])
print("Dataset size(train): ", train_size)
print("Dataset size(validation): ", valid_size)
metric = evaluate.load("accuracy")
def compute_metrics(eval_pred):
logits, labels = eval_pred
predictions = np.argmax(logits, axis=-1)
return metric.compute(predictions=predictions, references=labels)
# loading base model
model = ModernBertForSequenceClassification.from_pretrained("answerdotai/ModernBERT-base", num_labels=31).to(device)
default_args = {
"output_dir": "test_trainer",
"num_train_epochs": 1,
"log_level": "debug",
"per_device_train_batch_size":25,
"per_device_eval_batch_size":50,
"label_names": datasets['train'].features['labels'].names,
"logging_steps": 100,
"do_train": True,
"do_eval": True,
"save_strategy": "steps",
"save_total_limit" : 5,
"load_best_model_at_end" : True,
"eval_strategy": "steps",
"eval_on_start": True,
# "dataloader_num_workers": 16,
"seed": 42,
"warmup_ratio": 0.1,
"fp16": True,
}
train_n_step = train_size / default_args["per_device_train_batch_size"]
valid_n_step = valid_size / default_args["per_device_eval_batch_size"]
default_args["save_steps"]= 5000 #train_n_step // 5
default_args["eval_steps"]= 5000 #valid_n_step // 5
training_args = TrainingArguments(**default_args)
trainer = Trainer(
model=model,
args=training_args,
train_dataset=datasets['train'],
eval_dataset=datasets['validation'],
compute_metrics=compute_metrics,
)
trainer.train()
if training_args.eval_strategy != "no":
metrics = trainer.evaluate()
trainer.log_metrics("eval", metrics)
trainer.save_metrics("eval", metrics)
# Save and push to hub
trainer.save_model(training_args.output_dir) # it only save adaptor, not merged model
# if training_args.push_to_hub:
# trainer.push_to_hub(dataset_name=training_args.dataset_name)
if __name__ == "__main__":
main()
```
### Expected behavior
has_labels=True
compute_metric works! | {
"login": "HERIUN",
"id": 25131767,
"node_id": "MDQ6VXNlcjI1MTMxNzY3",
"avatar_url": "https://avatars.githubusercontent.com/u/25131767?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/HERIUN",
"html_url": "https://github.com/HERIUN",
"followers_url": "https://api.github.com/users/HERIUN/followers",
"following_url": "https://api.github.com/users/HERIUN/following{/other_user}",
"gists_url": "https://api.github.com/users/HERIUN/gists{/gist_id}",
"starred_url": "https://api.github.com/users/HERIUN/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/HERIUN/subscriptions",
"organizations_url": "https://api.github.com/users/HERIUN/orgs",
"repos_url": "https://api.github.com/users/HERIUN/repos",
"events_url": "https://api.github.com/users/HERIUN/events{/privacy}",
"received_events_url": "https://api.github.com/users/HERIUN/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38809/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/38809/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38808 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38808/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38808/comments | https://api.github.com/repos/huggingface/transformers/issues/38808/events | https://github.com/huggingface/transformers/pull/38808 | 3,142,820,770 | PR_kwDOCUB6oc6aWg2r | 38,808 | Fix configs and doc for the Qwens | {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T09:02:43 | 2025-06-13T09:16:06 | 2025-06-13T09:10:56 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38808",
"html_url": "https://github.com/huggingface/transformers/pull/38808",
"diff_url": "https://github.com/huggingface/transformers/pull/38808.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38808.patch",
"merged_at": "2025-06-13T09:10:55"
} | Clarify the doc, and remove it for Qwen3Moe as it's never used | {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38808/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/38808/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38807 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38807/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38807/comments | https://api.github.com/repos/huggingface/transformers/issues/38807/events | https://github.com/huggingface/transformers/pull/38807 | 3,142,205,395 | PR_kwDOCUB6oc6aUdxE | 38,807 | change fsdp_strategy to fsdp in TrainingArguments in accelerate doc | {
"login": "PT-10",
"id": 102211549,
"node_id": "U_kgDOBhef3Q",
"avatar_url": "https://avatars.githubusercontent.com/u/102211549?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/PT-10",
"html_url": "https://github.com/PT-10",
"followers_url": "https://api.github.com/users/PT-10/followers",
"following_url": "https://api.github.com/users/PT-10/following{/other_user}",
"gists_url": "https://api.github.com/users/PT-10/gists{/gist_id}",
"starred_url": "https://api.github.com/users/PT-10/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/PT-10/subscriptions",
"organizations_url": "https://api.github.com/users/PT-10/orgs",
"repos_url": "https://api.github.com/users/PT-10/repos",
"events_url": "https://api.github.com/users/PT-10/events{/privacy}",
"received_events_url": "https://api.github.com/users/PT-10/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T04:22:59 | 2025-06-13T15:33:09 | 2025-06-13T15:32:40 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38807",
"html_url": "https://github.com/huggingface/transformers/pull/38807",
"diff_url": "https://github.com/huggingface/transformers/pull/38807.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38807.patch",
"merged_at": "2025-06-13T15:32:40"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
Fixes #38776
## 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#create-a-pull-request),
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.
LINK - https://github.com/huggingface/transformers/issues/38776#issuecomment-29666039699
- [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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "SunMarc",
"id": 57196510,
"node_id": "MDQ6VXNlcjU3MTk2NTEw",
"avatar_url": "https://avatars.githubusercontent.com/u/57196510?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SunMarc",
"html_url": "https://github.com/SunMarc",
"followers_url": "https://api.github.com/users/SunMarc/followers",
"following_url": "https://api.github.com/users/SunMarc/following{/other_user}",
"gists_url": "https://api.github.com/users/SunMarc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SunMarc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SunMarc/subscriptions",
"organizations_url": "https://api.github.com/users/SunMarc/orgs",
"repos_url": "https://api.github.com/users/SunMarc/repos",
"events_url": "https://api.github.com/users/SunMarc/events{/privacy}",
"received_events_url": "https://api.github.com/users/SunMarc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38807/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/38807/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38806 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38806/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38806/comments | https://api.github.com/repos/huggingface/transformers/issues/38806/events | https://github.com/huggingface/transformers/pull/38806 | 3,142,177,378 | PR_kwDOCUB6oc6aUYAN | 38,806 | Remove fsdp_strategy from TrainingArguments in accelerate.md doc | {
"login": "PT-10",
"id": 102211549,
"node_id": "U_kgDOBhef3Q",
"avatar_url": "https://avatars.githubusercontent.com/u/102211549?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/PT-10",
"html_url": "https://github.com/PT-10",
"followers_url": "https://api.github.com/users/PT-10/followers",
"following_url": "https://api.github.com/users/PT-10/following{/other_user}",
"gists_url": "https://api.github.com/users/PT-10/gists{/gist_id}",
"starred_url": "https://api.github.com/users/PT-10/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/PT-10/subscriptions",
"organizations_url": "https://api.github.com/users/PT-10/orgs",
"repos_url": "https://api.github.com/users/PT-10/repos",
"events_url": "https://api.github.com/users/PT-10/events{/privacy}",
"received_events_url": "https://api.github.com/users/PT-10/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T04:02:01 | 2025-06-13T04:12:07 | 2025-06-13T04:11:10 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38806",
"html_url": "https://github.com/huggingface/transformers/pull/38806",
"diff_url": "https://github.com/huggingface/transformers/pull/38806.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38806.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
Fixes #38776
## 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#create-a-pull-request), 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.
LINK - https://github.com/huggingface/transformers/issues/38776#issuecomment-2966603969
- [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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "PT-10",
"id": 102211549,
"node_id": "U_kgDOBhef3Q",
"avatar_url": "https://avatars.githubusercontent.com/u/102211549?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/PT-10",
"html_url": "https://github.com/PT-10",
"followers_url": "https://api.github.com/users/PT-10/followers",
"following_url": "https://api.github.com/users/PT-10/following{/other_user}",
"gists_url": "https://api.github.com/users/PT-10/gists{/gist_id}",
"starred_url": "https://api.github.com/users/PT-10/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/PT-10/subscriptions",
"organizations_url": "https://api.github.com/users/PT-10/orgs",
"repos_url": "https://api.github.com/users/PT-10/repos",
"events_url": "https://api.github.com/users/PT-10/events{/privacy}",
"received_events_url": "https://api.github.com/users/PT-10/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38806/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/38806/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38805 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38805/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38805/comments | https://api.github.com/repos/huggingface/transformers/issues/38805/events | https://github.com/huggingface/transformers/pull/38805 | 3,142,030,371 | PR_kwDOCUB6oc6aT48c | 38,805 | Add Dust3R | {
"login": "jadechoghari",
"id": 78852495,
"node_id": "MDQ6VXNlcjc4ODUyNDk1",
"avatar_url": "https://avatars.githubusercontent.com/u/78852495?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/jadechoghari",
"html_url": "https://github.com/jadechoghari",
"followers_url": "https://api.github.com/users/jadechoghari/followers",
"following_url": "https://api.github.com/users/jadechoghari/following{/other_user}",
"gists_url": "https://api.github.com/users/jadechoghari/gists{/gist_id}",
"starred_url": "https://api.github.com/users/jadechoghari/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/jadechoghari/subscriptions",
"organizations_url": "https://api.github.com/users/jadechoghari/orgs",
"repos_url": "https://api.github.com/users/jadechoghari/repos",
"events_url": "https://api.github.com/users/jadechoghari/events{/privacy}",
"received_events_url": "https://api.github.com/users/jadechoghari/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-13T02:16:59 | 2025-07-22T01:01:56 | null | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38805",
"html_url": "https://github.com/huggingface/transformers/pull/38805",
"diff_url": "https://github.com/huggingface/transformers/pull/38805.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38805.patch",
"merged_at": null
} | # What does this PR do?
Draft PR to add Dust3R
Cc @raf-fonseca
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38805/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/38805/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38804 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38804/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38804/comments | https://api.github.com/repos/huggingface/transformers/issues/38804/events | https://github.com/huggingface/transformers/pull/38804 | 3,141,848,711 | PR_kwDOCUB6oc6aTSBO | 38,804 | Add Korean translation for glossary.md | {
"login": "JoosunH",
"id": 127996020,
"node_id": "U_kgDOB6EQdA",
"avatar_url": "https://avatars.githubusercontent.com/u/127996020?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/JoosunH",
"html_url": "https://github.com/JoosunH",
"followers_url": "https://api.github.com/users/JoosunH/followers",
"following_url": "https://api.github.com/users/JoosunH/following{/other_user}",
"gists_url": "https://api.github.com/users/JoosunH/gists{/gist_id}",
"starred_url": "https://api.github.com/users/JoosunH/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/JoosunH/subscriptions",
"organizations_url": "https://api.github.com/users/JoosunH/orgs",
"repos_url": "https://api.github.com/users/JoosunH/repos",
"events_url": "https://api.github.com/users/JoosunH/events{/privacy}",
"received_events_url": "https://api.github.com/users/JoosunH/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-13T00:12:24 | 2025-07-07T16:12:55 | 2025-07-07T16:12:55 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38804",
"html_url": "https://github.com/huggingface/transformers/pull/38804",
"diff_url": "https://github.com/huggingface/transformers/pull/38804.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38804.patch",
"merged_at": "2025-07-07T16:12:55"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38804/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/38804/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38803 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38803/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38803/comments | https://api.github.com/repos/huggingface/transformers/issues/38803/events | https://github.com/huggingface/transformers/issues/38803 | 3,141,716,627 | I_kwDOCUB6oc67QsqT | 38,803 | DTensor issues when running Llama4ForConditionalGeneration with tensor parallel. | {
"login": "czkkkkkk",
"id": 17905585,
"node_id": "MDQ6VXNlcjE3OTA1NTg1",
"avatar_url": "https://avatars.githubusercontent.com/u/17905585?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/czkkkkkk",
"html_url": "https://github.com/czkkkkkk",
"followers_url": "https://api.github.com/users/czkkkkkk/followers",
"following_url": "https://api.github.com/users/czkkkkkk/following{/other_user}",
"gists_url": "https://api.github.com/users/czkkkkkk/gists{/gist_id}",
"starred_url": "https://api.github.com/users/czkkkkkk/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/czkkkkkk/subscriptions",
"organizations_url": "https://api.github.com/users/czkkkkkk/orgs",
"repos_url": "https://api.github.com/users/czkkkkkk/repos",
"events_url": "https://api.github.com/users/czkkkkkk/events{/privacy}",
"received_events_url": "https://api.github.com/users/czkkkkkk/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-12T22:34:24 | 2025-07-21T08:03:24 | 2025-07-21T08:03:24 | NONE | null | null | null | null | ### System Info
transformers version: 4.52.4
pytorch version: 2.6
### Who can help?
transformers version: 4.52.4
pytorch version: 2.6
When running Llama4 with tensor parallel, [torch.nn.Unfold used in llama4 ](https://github.com/huggingface/transformers/blob/v4.52.4/src/transformers/models/llama4/modeling_llama4.py#L1320) isn't compatible with DTensor. So I got this error:
`NotImplementedError: Operator aten.im2col.default does not have a sharding strategy registered.` Looks like it is because the latest transformers use replicate DTensor for layers without tp_plan but `Unfold` isn't compatible with DTensor.
To workaround this error, I manually changed the input tensor to regular Tensor.
```
device_mesh = hidden_states.device_mesh if isinstance(hidden_states, DTensor) else None
placements = hidden_states.placements if isinstance(hidden_states, DTensor) else None
hidden_states = hidden_states.to_local()
hidden_states = self.unfold(hidden_states)
hidden_states = DTensor.from_local(hidden_states, device_mesh, placements)
```
After the change, I got `AttributeError: 'BaseModelOutput' object has no attribute 'to_local'` when running `vision_model` https://github.com/huggingface/transformers/blob/v4.52.4/src/transformers/models/llama4/modeling_llama4.py#L1543.
### Information
- [ ] The official example scripts
- [x] 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
Minimal script to reproduce the error
```py
# test.py
import torch
from transformers.models.llama4.modeling_llama4 import Llama4ForConditionalGeneration
if __name__ == '__main__':
model_id = "meta-llama/Llama-4-Scout-17B-16E-Instruct"
model = Llama4ForConditionalGeneration.from_pretrained(
model_id,
torch_dtype=torch.bfloat16,
low_cpu_mem_usage=True,
device_map="auto",
)
B = 1
S = 128
input_ids = torch.randint(0, 1000, (B, S))
attention_mask = torch.ones((B, S))
pixel_values = torch.randn((5, 3, 336, 336)).to(torch.bfloat16)
model(input_ids=input_ids, attention_mask=attention_mask, pixel_values=pixel_values)
```
```sh
torchrun --nproc-per-node=8 test.py
```
### Expected behavior
Expect the program to finish successfully. | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38803/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/38803/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38802 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38802/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38802/comments | https://api.github.com/repos/huggingface/transformers/issues/38802/events | https://github.com/huggingface/transformers/pull/38802 | 3,141,594,055 | PR_kwDOCUB6oc6aSbNi | 38,802 | feat: Add granite architectures to auto tokenizer name mappings | {
"login": "gabe-l-hart",
"id": 1254484,
"node_id": "MDQ6VXNlcjEyNTQ0ODQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/1254484?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/gabe-l-hart",
"html_url": "https://github.com/gabe-l-hart",
"followers_url": "https://api.github.com/users/gabe-l-hart/followers",
"following_url": "https://api.github.com/users/gabe-l-hart/following{/other_user}",
"gists_url": "https://api.github.com/users/gabe-l-hart/gists{/gist_id}",
"starred_url": "https://api.github.com/users/gabe-l-hart/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/gabe-l-hart/subscriptions",
"organizations_url": "https://api.github.com/users/gabe-l-hart/orgs",
"repos_url": "https://api.github.com/users/gabe-l-hart/repos",
"events_url": "https://api.github.com/users/gabe-l-hart/events{/privacy}",
"received_events_url": "https://api.github.com/users/gabe-l-hart/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T21:22:50 | 2025-06-20T16:21:03 | 2025-06-19T14:20:42 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38802",
"html_url": "https://github.com/huggingface/transformers/pull/38802",
"diff_url": "https://github.com/huggingface/transformers/pull/38802.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38802.patch",
"merged_at": "2025-06-19T14:20:42"
} | Branch: GraniteTokenizerMapping
# What does this PR do?
This PR adds the `granite`, `granitemoe`, `granitemoeshared`, and `granitemoehybrid` model architectures to the `TOKENIZER_MAPPING_NAMES` mapping so that using `pipeline` will correctly infer the tokenizer class when loading the models without needing to add it to the `config.json` in each model ([this line](https://github.com/huggingface/transformers/blob/main/src/transformers/pipelines/__init__.py#L955))
## 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#create-a-pull-request),
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?
@ArthurZucker for text model / tokenizer pipeline review
@Rocketknight1 for pipeline review
@mayank31398 for Granite models
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38802/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/38802/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38801 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38801/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38801/comments | https://api.github.com/repos/huggingface/transformers/issues/38801/events | https://github.com/huggingface/transformers/pull/38801 | 3,141,425,334 | PR_kwDOCUB6oc6aR3AA | 38,801 | GraniteMoeHybrid: Allow for only shared expert case. | {
"login": "shawntan",
"id": 119799,
"node_id": "MDQ6VXNlcjExOTc5OQ==",
"avatar_url": "https://avatars.githubusercontent.com/u/119799?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/shawntan",
"html_url": "https://github.com/shawntan",
"followers_url": "https://api.github.com/users/shawntan/followers",
"following_url": "https://api.github.com/users/shawntan/following{/other_user}",
"gists_url": "https://api.github.com/users/shawntan/gists{/gist_id}",
"starred_url": "https://api.github.com/users/shawntan/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/shawntan/subscriptions",
"organizations_url": "https://api.github.com/users/shawntan/orgs",
"repos_url": "https://api.github.com/users/shawntan/repos",
"events_url": "https://api.github.com/users/shawntan/events{/privacy}",
"received_events_url": "https://api.github.com/users/shawntan/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T20:01:25 | 2025-06-16T15:15:42 | 2025-06-16T15:15:42 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38801",
"html_url": "https://github.com/huggingface/transformers/pull/38801",
"diff_url": "https://github.com/huggingface/transformers/pull/38801.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38801.patch",
"merged_at": "2025-06-16T15:15:42"
} | # What does this PR do?
Allows for the case where `GraniteMoeHybrid` models have `num_local_experts = 0`
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38801/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/38801/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38800 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38800/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38800/comments | https://api.github.com/repos/huggingface/transformers/issues/38800/events | https://github.com/huggingface/transformers/pull/38800 | 3,141,336,885 | PR_kwDOCUB6oc6aRjdU | 38,800 | [docs] Update docs moved to the course | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T19:25:12 | 2025-06-13T19:02:31 | 2025-06-13T19:02:28 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38800",
"html_url": "https://github.com/huggingface/transformers/pull/38800",
"diff_url": "https://github.com/huggingface/transformers/pull/38800.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38800.patch",
"merged_at": "2025-06-13T19:02:28"
} | Adds a new section on `index.md` to link to the LLM course for more in-depth details about how transformer models work. This also removes some redundant docs in Transformers that have since been migrated to the course (see https://github.com/huggingface/course/issues/839). | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38800/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/38800/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38799 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38799/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38799/comments | https://api.github.com/repos/huggingface/transformers/issues/38799/events | https://github.com/huggingface/transformers/pull/38799 | 3,141,251,680 | PR_kwDOCUB6oc6aRRB3 | 38,799 | Merge subwords when align_to_words=True | {
"login": "DuyguA",
"id": 8277232,
"node_id": "MDQ6VXNlcjgyNzcyMzI=",
"avatar_url": "https://avatars.githubusercontent.com/u/8277232?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/DuyguA",
"html_url": "https://github.com/DuyguA",
"followers_url": "https://api.github.com/users/DuyguA/followers",
"following_url": "https://api.github.com/users/DuyguA/following{/other_user}",
"gists_url": "https://api.github.com/users/DuyguA/gists{/gist_id}",
"starred_url": "https://api.github.com/users/DuyguA/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/DuyguA/subscriptions",
"organizations_url": "https://api.github.com/users/DuyguA/orgs",
"repos_url": "https://api.github.com/users/DuyguA/repos",
"events_url": "https://api.github.com/users/DuyguA/events{/privacy}",
"received_events_url": "https://api.github.com/users/DuyguA/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T18:44:24 | 2025-06-23T08:49:29 | 2025-06-14T08:18:01 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38799",
"html_url": "https://github.com/huggingface/transformers/pull/38799",
"diff_url": "https://github.com/huggingface/transformers/pull/38799.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38799.patch",
"merged_at": null
} | # What does this PR do?
Fixes #26286
For the issue basically when `align_to_words=True` , I added start/end subword probabilities per each word and assigned to the first subword of the word (also assigned 0s for start/end probs to rest subwords), this way assured that the first subword will always be chosen provided this is the start/end of the answer. More graphically consider start probabilities:
before:
```
Ang #ela Merkel is the chancellor of Germany
0.01 0.01 ...................................
```
after:
```
Ang #ela Merkel is the chancellor of Germany
0.02 0.00 ...................................
```
This way I was able to keep the rest of the span calculation changes minimal.
Also I ruffes the 2 files, that introduced some diff too, please discard those.
## 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#create-a-pull-request),
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).
- [x ] Did you write any new necessary tests?
## Who can review?
@ArthurZucker , @MichelBartels
| {
"login": "DuyguA",
"id": 8277232,
"node_id": "MDQ6VXNlcjgyNzcyMzI=",
"avatar_url": "https://avatars.githubusercontent.com/u/8277232?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/DuyguA",
"html_url": "https://github.com/DuyguA",
"followers_url": "https://api.github.com/users/DuyguA/followers",
"following_url": "https://api.github.com/users/DuyguA/following{/other_user}",
"gists_url": "https://api.github.com/users/DuyguA/gists{/gist_id}",
"starred_url": "https://api.github.com/users/DuyguA/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/DuyguA/subscriptions",
"organizations_url": "https://api.github.com/users/DuyguA/orgs",
"repos_url": "https://api.github.com/users/DuyguA/repos",
"events_url": "https://api.github.com/users/DuyguA/events{/privacy}",
"received_events_url": "https://api.github.com/users/DuyguA/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38799/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/38799/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38798 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38798/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38798/comments | https://api.github.com/repos/huggingface/transformers/issues/38798/events | https://github.com/huggingface/transformers/pull/38798 | 3,140,917,730 | PR_kwDOCUB6oc6aQIiL | 38,798 | fixed docstring in modular_qwen2_5_vl.py | {
"login": "lawrencefeng17",
"id": 80508784,
"node_id": "MDQ6VXNlcjgwNTA4Nzg0",
"avatar_url": "https://avatars.githubusercontent.com/u/80508784?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/lawrencefeng17",
"html_url": "https://github.com/lawrencefeng17",
"followers_url": "https://api.github.com/users/lawrencefeng17/followers",
"following_url": "https://api.github.com/users/lawrencefeng17/following{/other_user}",
"gists_url": "https://api.github.com/users/lawrencefeng17/gists{/gist_id}",
"starred_url": "https://api.github.com/users/lawrencefeng17/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/lawrencefeng17/subscriptions",
"organizations_url": "https://api.github.com/users/lawrencefeng17/orgs",
"repos_url": "https://api.github.com/users/lawrencefeng17/repos",
"events_url": "https://api.github.com/users/lawrencefeng17/events{/privacy}",
"received_events_url": "https://api.github.com/users/lawrencefeng17/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T16:27:35 | 2025-06-13T18:09:51 | 2025-06-13T18:09:51 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38798",
"html_url": "https://github.com/huggingface/transformers/pull/38798",
"diff_url": "https://github.com/huggingface/transformers/pull/38798.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38798.patch",
"merged_at": "2025-06-13T18:09:51"
} | # What does this PR do?
Fixes a small typo in the docstring reference to `Qwen2VLImageProcessor`. The incorrect reference (`Qwen2_5_VLImageProcessor`).
## Before submitting
- [x] This PR fixes a typo or improves the docs.
- [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request)?
- [x] Did you make sure to update the documentation with your changes?
- [ ] Did you write any new necessary tests? *(Not applicable)*
## Who can review?
@stevhliu for documentation | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38798/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/38798/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38797 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38797/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38797/comments | https://api.github.com/repos/huggingface/transformers/issues/38797/events | https://github.com/huggingface/transformers/pull/38797 | 3,140,873,982 | PR_kwDOCUB6oc6aP-1G | 38,797 | No more Tuple, List, Dict | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T16:13:22 | 2025-06-18T07:14:49 | 2025-06-17T18:37:18 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38797",
"html_url": "https://github.com/huggingface/transformers/pull/38797",
"diff_url": "https://github.com/huggingface/transformers/pull/38797.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38797.patch",
"merged_at": "2025-06-17T18:37:18"
} | This PR uses @cyyever's work to update types from Tuple/List/Dict to tuple/list/dict. We also add the UP006 ruff check to require this in future. | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38797/reactions",
"total_count": 1,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 1,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/38797/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38796 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38796/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38796/comments | https://api.github.com/repos/huggingface/transformers/issues/38796/events | https://github.com/huggingface/transformers/pull/38796 | 3,140,864,428 | PR_kwDOCUB6oc6aP80l | 38,796 | No more Tuple, List, Dict | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T16:09:31 | 2025-06-12T16:22:46 | 2025-06-12T16:10:58 | MEMBER | null | null | true | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38796",
"html_url": "https://github.com/huggingface/transformers/pull/38796",
"diff_url": "https://github.com/huggingface/transformers/pull/38796.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38796.patch",
"merged_at": null
} | This PR uses @cyyever's work to update types from Tuple/List/Dict to tuple/list/dict. We also add the `UP006` ruff check to require this in future. | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38796/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/38796/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38795 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38795/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38795/comments | https://api.github.com/repos/huggingface/transformers/issues/38795/events | https://github.com/huggingface/transformers/pull/38795 | 3,140,776,288 | PR_kwDOCUB6oc6aPpq6 | 38,795 | LlamaAttention forward function type hint is incorrect #38739 | {
"login": "ArkVex",
"id": 159469387,
"node_id": "U_kgDOCYFPSw",
"avatar_url": "https://avatars.githubusercontent.com/u/159469387?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ArkVex",
"html_url": "https://github.com/ArkVex",
"followers_url": "https://api.github.com/users/ArkVex/followers",
"following_url": "https://api.github.com/users/ArkVex/following{/other_user}",
"gists_url": "https://api.github.com/users/ArkVex/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ArkVex/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ArkVex/subscriptions",
"organizations_url": "https://api.github.com/users/ArkVex/orgs",
"repos_url": "https://api.github.com/users/ArkVex/repos",
"events_url": "https://api.github.com/users/ArkVex/events{/privacy}",
"received_events_url": "https://api.github.com/users/ArkVex/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T15:39:00 | 2025-06-16T17:34:13 | 2025-06-16T17:34:13 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38795",
"html_url": "https://github.com/huggingface/transformers/pull/38795",
"diff_url": "https://github.com/huggingface/transformers/pull/38795.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38795.patch",
"merged_at": null
} | Hi, this PR fixes a small issue in the LlamaAttention class. The return type in the forward method currently shows three values, but the function actually returns only two. This seems to have been missed during the attention refactor (possibly in PR #35235).
I’ve updated the type hint to reflect the actual return values, just to avoid confusion for anyone reading or using the code. Let me know if any other changes are needed. Happy to help! | {
"login": "ArkVex",
"id": 159469387,
"node_id": "U_kgDOCYFPSw",
"avatar_url": "https://avatars.githubusercontent.com/u/159469387?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ArkVex",
"html_url": "https://github.com/ArkVex",
"followers_url": "https://api.github.com/users/ArkVex/followers",
"following_url": "https://api.github.com/users/ArkVex/following{/other_user}",
"gists_url": "https://api.github.com/users/ArkVex/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ArkVex/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ArkVex/subscriptions",
"organizations_url": "https://api.github.com/users/ArkVex/orgs",
"repos_url": "https://api.github.com/users/ArkVex/repos",
"events_url": "https://api.github.com/users/ArkVex/events{/privacy}",
"received_events_url": "https://api.github.com/users/ArkVex/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38795/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/38795/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38794 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38794/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38794/comments | https://api.github.com/repos/huggingface/transformers/issues/38794/events | https://github.com/huggingface/transformers/pull/38794 | 3,140,738,389 | PR_kwDOCUB6oc6aPha1 | 38,794 | Fix erroneous docstring for the ordering of SWA layers | {
"login": "norpadon",
"id": 6224581,
"node_id": "MDQ6VXNlcjYyMjQ1ODE=",
"avatar_url": "https://avatars.githubusercontent.com/u/6224581?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/norpadon",
"html_url": "https://github.com/norpadon",
"followers_url": "https://api.github.com/users/norpadon/followers",
"following_url": "https://api.github.com/users/norpadon/following{/other_user}",
"gists_url": "https://api.github.com/users/norpadon/gists{/gist_id}",
"starred_url": "https://api.github.com/users/norpadon/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/norpadon/subscriptions",
"organizations_url": "https://api.github.com/users/norpadon/orgs",
"repos_url": "https://api.github.com/users/norpadon/repos",
"events_url": "https://api.github.com/users/norpadon/events{/privacy}",
"received_events_url": "https://api.github.com/users/norpadon/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T15:27:18 | 2025-06-13T08:46:45 | 2025-06-13T08:46:45 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38794",
"html_url": "https://github.com/huggingface/transformers/pull/38794",
"diff_url": "https://github.com/huggingface/transformers/pull/38794.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38794.patch",
"merged_at": "2025-06-13T08:46:45"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
Fixes #38787
## 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#create-a-pull-request),
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).
- [x] Did you write any new necessary tests?
## Who can review?
@Cyrilvallez @Rocketknight1
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38794/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/38794/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38793 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38793/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38793/comments | https://api.github.com/repos/huggingface/transformers/issues/38793/events | https://github.com/huggingface/transformers/pull/38793 | 3,140,588,300 | PR_kwDOCUB6oc6aPAZ6 | 38,793 | Fix Typos in Comments and Improve Clarity | {
"login": "kilavvy",
"id": 140459108,
"node_id": "U_kgDOCF88ZA",
"avatar_url": "https://avatars.githubusercontent.com/u/140459108?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/kilavvy",
"html_url": "https://github.com/kilavvy",
"followers_url": "https://api.github.com/users/kilavvy/followers",
"following_url": "https://api.github.com/users/kilavvy/following{/other_user}",
"gists_url": "https://api.github.com/users/kilavvy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/kilavvy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/kilavvy/subscriptions",
"organizations_url": "https://api.github.com/users/kilavvy/orgs",
"repos_url": "https://api.github.com/users/kilavvy/repos",
"events_url": "https://api.github.com/users/kilavvy/events{/privacy}",
"received_events_url": "https://api.github.com/users/kilavvy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-12T14:43:18 | 2025-06-12T21:08:43 | null | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38793",
"html_url": "https://github.com/huggingface/transformers/pull/38793",
"diff_url": "https://github.com/huggingface/transformers/pull/38793.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38793.patch",
"merged_at": null
} |
Description:
This pull request fixes minor typos in comments across two files:
- In `modeling_swin.py`, the word "disible" was corrected to "divisible" for clarity in the padding comment.
- In `pt_utils.py`, the typo "lookig" was corrected to "looking" in a comment explaining iterator behavior.
These changes improve code readability and maintain documentation quality. No functional code was modified. | null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38793/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/38793/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38792 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38792/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38792/comments | https://api.github.com/repos/huggingface/transformers/issues/38792/events | https://github.com/huggingface/transformers/pull/38792 | 3,140,464,970 | PR_kwDOCUB6oc6aOlZp | 38,792 | Remove all traces of `low_cpu_mem_usage` | {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T14:08:41 | 2025-06-12T14:42:58 | 2025-06-12T14:39:34 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38792",
"html_url": "https://github.com/huggingface/transformers/pull/38792",
"diff_url": "https://github.com/huggingface/transformers/pull/38792.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38792.patch",
"merged_at": "2025-06-12T14:39:34"
} | # What does this PR do?
`low_cpu_mem_usage` argument was removed in https://github.com/huggingface/transformers/pull/36963 - this PR removes all leftover everywhere in the library (the arg is ignored already)
We only keep it in the `peft` integration, as it is used downstream in `peft` | {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38792/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/38792/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38791 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38791/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38791/comments | https://api.github.com/repos/huggingface/transformers/issues/38791/events | https://github.com/huggingface/transformers/pull/38791 | 3,140,168,863 | PR_kwDOCUB6oc6aNkGL | 38,791 | Fix `llava_onevision` tests | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T12:49:32 | 2025-06-12T13:06:51 | 2025-06-12T13:06:49 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38791",
"html_url": "https://github.com/huggingface/transformers/pull/38791",
"diff_url": "https://github.com/huggingface/transformers/pull/38791.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38791.patch",
"merged_at": "2025-06-12T13:06:49"
} | # What does this PR do?
As usual, just some updates of expected values.
Now pass both on A10 / T4 torch 2.7.1 | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38791/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/38791/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38790 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38790/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38790/comments | https://api.github.com/repos/huggingface/transformers/issues/38790/events | https://github.com/huggingface/transformers/pull/38790 | 3,140,139,561 | PR_kwDOCUB6oc6aNduP | 38,790 | Gaudi3 CI | {
"login": "IlyasMoutawwakil",
"id": 57442720,
"node_id": "MDQ6VXNlcjU3NDQyNzIw",
"avatar_url": "https://avatars.githubusercontent.com/u/57442720?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/IlyasMoutawwakil",
"html_url": "https://github.com/IlyasMoutawwakil",
"followers_url": "https://api.github.com/users/IlyasMoutawwakil/followers",
"following_url": "https://api.github.com/users/IlyasMoutawwakil/following{/other_user}",
"gists_url": "https://api.github.com/users/IlyasMoutawwakil/gists{/gist_id}",
"starred_url": "https://api.github.com/users/IlyasMoutawwakil/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/IlyasMoutawwakil/subscriptions",
"organizations_url": "https://api.github.com/users/IlyasMoutawwakil/orgs",
"repos_url": "https://api.github.com/users/IlyasMoutawwakil/repos",
"events_url": "https://api.github.com/users/IlyasMoutawwakil/events{/privacy}",
"received_events_url": "https://api.github.com/users/IlyasMoutawwakil/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T12:40:00 | 2025-06-23T08:56:54 | 2025-06-23T08:56:52 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38790",
"html_url": "https://github.com/huggingface/transformers/pull/38790",
"diff_url": "https://github.com/huggingface/transformers/pull/38790.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38790.patch",
"merged_at": "2025-06-23T08:56:52"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
@regisss @ydshieh
## 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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface, @SunMarc and @qgallouedec
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "IlyasMoutawwakil",
"id": 57442720,
"node_id": "MDQ6VXNlcjU3NDQyNzIw",
"avatar_url": "https://avatars.githubusercontent.com/u/57442720?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/IlyasMoutawwakil",
"html_url": "https://github.com/IlyasMoutawwakil",
"followers_url": "https://api.github.com/users/IlyasMoutawwakil/followers",
"following_url": "https://api.github.com/users/IlyasMoutawwakil/following{/other_user}",
"gists_url": "https://api.github.com/users/IlyasMoutawwakil/gists{/gist_id}",
"starred_url": "https://api.github.com/users/IlyasMoutawwakil/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/IlyasMoutawwakil/subscriptions",
"organizations_url": "https://api.github.com/users/IlyasMoutawwakil/orgs",
"repos_url": "https://api.github.com/users/IlyasMoutawwakil/repos",
"events_url": "https://api.github.com/users/IlyasMoutawwakil/events{/privacy}",
"received_events_url": "https://api.github.com/users/IlyasMoutawwakil/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38790/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/38790/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38789 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38789/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38789/comments | https://api.github.com/repos/huggingface/transformers/issues/38789/events | https://github.com/huggingface/transformers/pull/38789 | 3,140,136,396 | PR_kwDOCUB6oc6aNdA8 | 38,789 | build: :pushpin: Remove upper bound on PyTorch | {
"login": "KyleMylonakisProtopia",
"id": 122286752,
"node_id": "U_kgDOB0nyoA",
"avatar_url": "https://avatars.githubusercontent.com/u/122286752?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/KyleMylonakisProtopia",
"html_url": "https://github.com/KyleMylonakisProtopia",
"followers_url": "https://api.github.com/users/KyleMylonakisProtopia/followers",
"following_url": "https://api.github.com/users/KyleMylonakisProtopia/following{/other_user}",
"gists_url": "https://api.github.com/users/KyleMylonakisProtopia/gists{/gist_id}",
"starred_url": "https://api.github.com/users/KyleMylonakisProtopia/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/KyleMylonakisProtopia/subscriptions",
"organizations_url": "https://api.github.com/users/KyleMylonakisProtopia/orgs",
"repos_url": "https://api.github.com/users/KyleMylonakisProtopia/repos",
"events_url": "https://api.github.com/users/KyleMylonakisProtopia/events{/privacy}",
"received_events_url": "https://api.github.com/users/KyleMylonakisProtopia/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T12:39:08 | 2025-06-12T17:18:45 | 2025-06-12T14:34:14 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38789",
"html_url": "https://github.com/huggingface/transformers/pull/38789",
"diff_url": "https://github.com/huggingface/transformers/pull/38789.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38789.patch",
"merged_at": "2025-06-12T14:34:14"
} | # What does this PR do?
Removes the upper bound on the PyTorch dependency, as the original issue in PyTorch (https://github.com/pytorch/pytorch/issues/152275) has been resolved and released.
Fixes # (issue)
Fixes https://github.com/huggingface/transformers/issues/37824
## 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#create-a-pull-request),
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?
@anijain2305
| {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38789/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/38789/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38788 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38788/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38788/comments | https://api.github.com/repos/huggingface/transformers/issues/38788/events | https://github.com/huggingface/transformers/pull/38788 | 3,140,041,216 | PR_kwDOCUB6oc6aNIRr | 38,788 | Add V-JEPA for video classification model | {
"login": "qubvel",
"id": 31920396,
"node_id": "MDQ6VXNlcjMxOTIwMzk2",
"avatar_url": "https://avatars.githubusercontent.com/u/31920396?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/qubvel",
"html_url": "https://github.com/qubvel",
"followers_url": "https://api.github.com/users/qubvel/followers",
"following_url": "https://api.github.com/users/qubvel/following{/other_user}",
"gists_url": "https://api.github.com/users/qubvel/gists{/gist_id}",
"starred_url": "https://api.github.com/users/qubvel/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/qubvel/subscriptions",
"organizations_url": "https://api.github.com/users/qubvel/orgs",
"repos_url": "https://api.github.com/users/qubvel/repos",
"events_url": "https://api.github.com/users/qubvel/events{/privacy}",
"received_events_url": "https://api.github.com/users/qubvel/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T12:08:42 | 2025-06-13T17:16:02 | 2025-06-13T16:56:15 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38788",
"html_url": "https://github.com/huggingface/transformers/pull/38788",
"diff_url": "https://github.com/huggingface/transformers/pull/38788.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38788.patch",
"merged_at": "2025-06-13T16:56:15"
} | # What does this PR do?
Adds a classification model compatible with the released pretrained classifiers on Something-Something-v2 and Diving48 datasets.
Epic Kitchen is a bit special, with 3 different classifiers on top of pooler (noun, verb, action), so it might require an additional model class to handle the case (I'm not sure if we have to add it, let's see on the V-JEPA usage).
cc @yonigozlan @koustuvsinha
P.S. commit history is a bit weird cause I started working on it before model merge, but diff is OK | {
"login": "qubvel",
"id": 31920396,
"node_id": "MDQ6VXNlcjMxOTIwMzk2",
"avatar_url": "https://avatars.githubusercontent.com/u/31920396?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/qubvel",
"html_url": "https://github.com/qubvel",
"followers_url": "https://api.github.com/users/qubvel/followers",
"following_url": "https://api.github.com/users/qubvel/following{/other_user}",
"gists_url": "https://api.github.com/users/qubvel/gists{/gist_id}",
"starred_url": "https://api.github.com/users/qubvel/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/qubvel/subscriptions",
"organizations_url": "https://api.github.com/users/qubvel/orgs",
"repos_url": "https://api.github.com/users/qubvel/repos",
"events_url": "https://api.github.com/users/qubvel/events{/privacy}",
"received_events_url": "https://api.github.com/users/qubvel/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38788/reactions",
"total_count": 6,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 2,
"confused": 0,
"heart": 4,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/38788/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38787 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38787/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38787/comments | https://api.github.com/repos/huggingface/transformers/issues/38787/events | https://github.com/huggingface/transformers/issues/38787 | 3,140,007,148 | I_kwDOCUB6oc67KLTs | 38,787 | Qwen3 Incorrect order of sliding window layers | {
"login": "norpadon",
"id": 6224581,
"node_id": "MDQ6VXNlcjYyMjQ1ODE=",
"avatar_url": "https://avatars.githubusercontent.com/u/6224581?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/norpadon",
"html_url": "https://github.com/norpadon",
"followers_url": "https://api.github.com/users/norpadon/followers",
"following_url": "https://api.github.com/users/norpadon/following{/other_user}",
"gists_url": "https://api.github.com/users/norpadon/gists{/gist_id}",
"starred_url": "https://api.github.com/users/norpadon/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/norpadon/subscriptions",
"organizations_url": "https://api.github.com/users/norpadon/orgs",
"repos_url": "https://api.github.com/users/norpadon/repos",
"events_url": "https://api.github.com/users/norpadon/events{/privacy}",
"received_events_url": "https://api.github.com/users/norpadon/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-12T11:57:17 | 2025-06-13T08:46:46 | 2025-06-13T08:46:45 | CONTRIBUTOR | null | null | null | null | ### System Info
Irrelevant
### Who can help?
@ArthurZucker
Documentation [says](https://github.com/huggingface/transformers/blob/89c46b648d82b670cc7286a25fa64d2d92770418/src/transformers/models/qwen3/configuration_qwen3.py#L115C9-L116C134):
"The number of layers that use SWA (Sliding Window Attention). The bottom layers use SWA while the top use full attention."
But the code [does the opposite thing](https://github.com/huggingface/transformers/blob/89c46b648d82b670cc7286a25fa64d2d92770418/src/transformers/models/qwen3/configuration_qwen3.py#L210C9-L216C14):
```python
if self.layer_types is None:
self.layer_types = [
"sliding_attention"
if self.sliding_window is not None and i >= self.max_window_layers
else "full_attention"
for i in range(self.num_hidden_layers)
]
```
WAT? Which one is correct?
### Information
- [x] The official example scripts
- [x] 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
Look at the code, look at the documentation, look at the code again, have existential crisis
### Expected behavior
No existential crisis | {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38787/reactions",
"total_count": 5,
"+1": 3,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 2,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/38787/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38786 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38786/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38786/comments | https://api.github.com/repos/huggingface/transformers/issues/38786/events | https://github.com/huggingface/transformers/pull/38786 | 3,139,685,284 | PR_kwDOCUB6oc6aL5yw | 38,786 | Provide clearer instructions on how to specify target language. | {
"login": "khof312",
"id": 6764017,
"node_id": "MDQ6VXNlcjY3NjQwMTc=",
"avatar_url": "https://avatars.githubusercontent.com/u/6764017?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/khof312",
"html_url": "https://github.com/khof312",
"followers_url": "https://api.github.com/users/khof312/followers",
"following_url": "https://api.github.com/users/khof312/following{/other_user}",
"gists_url": "https://api.github.com/users/khof312/gists{/gist_id}",
"starred_url": "https://api.github.com/users/khof312/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/khof312/subscriptions",
"organizations_url": "https://api.github.com/users/khof312/orgs",
"repos_url": "https://api.github.com/users/khof312/repos",
"events_url": "https://api.github.com/users/khof312/events{/privacy}",
"received_events_url": "https://api.github.com/users/khof312/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-12T10:18:14 | 2025-07-21T16:13:50 | null | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38786",
"html_url": "https://github.com/huggingface/transformers/pull/38786",
"diff_url": "https://github.com/huggingface/transformers/pull/38786.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38786.patch",
"merged_at": null
} | In the current documentation, I believe there is an error in the sample code. While the language token of the input is specified as `<2pt>` (Portuguese), the returned translation is in Italian, so actually the language should be `<2it>`.
I didn't initially understand that this was a token for the target language, so I have added an explanation of this to the documentation as well as instructions on how to get the available languages based on the [code](https://huggingface.co/spaces/jbochi/madlad400-3b-mt/blob/main/app.py) in Juarez Bochi's demo space.
It would be good to check that I have retrieved the correct range of tokens (4-459) and not omitted any (there were some additional tokens that looked to refer to scripts instead of languages, but I was not sure).
If these changes are approved, then they should also be perpetuated to the documentation of the different model checkpoints from google and jbochi.
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38786/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/38786/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38785 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38785/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38785/comments | https://api.github.com/repos/huggingface/transformers/issues/38785/events | https://github.com/huggingface/transformers/pull/38785 | 3,139,560,690 | PR_kwDOCUB6oc6aLeYX | 38,785 | Reduce verbosity for `average_tokens_across_devices=True` and `world size = 1` | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T09:40:26 | 2025-06-12T12:02:54 | 2025-06-12T12:02:53 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38785",
"html_url": "https://github.com/huggingface/transformers/pull/38785",
"diff_url": "https://github.com/huggingface/transformers/pull/38785.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38785.patch",
"merged_at": "2025-06-12T12:02:53"
} | # What does this PR do?
Averaging a single value is not strictly speaking incorrect. So I suggest rephrasing, and reducing the level to info. This removes recurring warnings when True is used as the default for this argument.
## 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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "SunMarc",
"id": 57196510,
"node_id": "MDQ6VXNlcjU3MTk2NTEw",
"avatar_url": "https://avatars.githubusercontent.com/u/57196510?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SunMarc",
"html_url": "https://github.com/SunMarc",
"followers_url": "https://api.github.com/users/SunMarc/followers",
"following_url": "https://api.github.com/users/SunMarc/following{/other_user}",
"gists_url": "https://api.github.com/users/SunMarc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SunMarc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SunMarc/subscriptions",
"organizations_url": "https://api.github.com/users/SunMarc/orgs",
"repos_url": "https://api.github.com/users/SunMarc/repos",
"events_url": "https://api.github.com/users/SunMarc/events{/privacy}",
"received_events_url": "https://api.github.com/users/SunMarc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38785/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/38785/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38784 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38784/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38784/comments | https://api.github.com/repos/huggingface/transformers/issues/38784/events | https://github.com/huggingface/transformers/pull/38784 | 3,139,409,328 | PR_kwDOCUB6oc6aK9iO | 38,784 | break style | {
"login": "ydshieh2",
"id": 183479141,
"node_id": "U_kgDOCu-rZQ",
"avatar_url": "https://avatars.githubusercontent.com/u/183479141?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ydshieh2",
"html_url": "https://github.com/ydshieh2",
"followers_url": "https://api.github.com/users/ydshieh2/followers",
"following_url": "https://api.github.com/users/ydshieh2/following{/other_user}",
"gists_url": "https://api.github.com/users/ydshieh2/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ydshieh2/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ydshieh2/subscriptions",
"organizations_url": "https://api.github.com/users/ydshieh2/orgs",
"repos_url": "https://api.github.com/users/ydshieh2/repos",
"events_url": "https://api.github.com/users/ydshieh2/events{/privacy}",
"received_events_url": "https://api.github.com/users/ydshieh2/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-12T08:51:40 | 2025-06-12T08:52:58 | null | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38784",
"html_url": "https://github.com/huggingface/transformers/pull/38784",
"diff_url": "https://github.com/huggingface/transformers/pull/38784.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38784.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38784/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/38784/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38783 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38783/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38783/comments | https://api.github.com/repos/huggingface/transformers/issues/38783/events | https://github.com/huggingface/transformers/pull/38783 | 3,139,406,443 | PR_kwDOCUB6oc6aK86M | 38,783 | Fix masking utils | {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T08:50:40 | 2025-06-12T09:06:12 | 2025-06-12T09:00:46 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38783",
"html_url": "https://github.com/huggingface/transformers/pull/38783",
"diff_url": "https://github.com/huggingface/transformers/pull/38783.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38783.patch",
"merged_at": "2025-06-12T09:00:46"
} | # What does this PR do?
https://github.com/huggingface/transformers/pull/38561 removed the try/catch statement but changed True to False in sliding and chunked mask logic. This PR fixes it.
cc @zucchini-nlp for viz | {
"login": "Cyrilvallez",
"id": 71554963,
"node_id": "MDQ6VXNlcjcxNTU0OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/71554963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Cyrilvallez",
"html_url": "https://github.com/Cyrilvallez",
"followers_url": "https://api.github.com/users/Cyrilvallez/followers",
"following_url": "https://api.github.com/users/Cyrilvallez/following{/other_user}",
"gists_url": "https://api.github.com/users/Cyrilvallez/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Cyrilvallez/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Cyrilvallez/subscriptions",
"organizations_url": "https://api.github.com/users/Cyrilvallez/orgs",
"repos_url": "https://api.github.com/users/Cyrilvallez/repos",
"events_url": "https://api.github.com/users/Cyrilvallez/events{/privacy}",
"received_events_url": "https://api.github.com/users/Cyrilvallez/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38783/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/38783/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38782 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38782/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38782/comments | https://api.github.com/repos/huggingface/transformers/issues/38782/events | https://github.com/huggingface/transformers/pull/38782 | 3,139,354,636 | PR_kwDOCUB6oc6aKxyK | 38,782 | debug style bot 2 | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T08:32:11 | 2025-06-12T09:18:03 | 2025-06-12T09:18:03 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38782",
"html_url": "https://github.com/huggingface/transformers/pull/38782",
"diff_url": "https://github.com/huggingface/transformers/pull/38782.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38782.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38782/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/38782/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38781 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38781/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38781/comments | https://api.github.com/repos/huggingface/transformers/issues/38781/events | https://github.com/huggingface/transformers/pull/38781 | 3,139,333,797 | PR_kwDOCUB6oc6aKtRE | 38,781 | Release: v4.45.0 | {
"login": "byungchul-sqzb",
"id": 211779483,
"node_id": "U_kgDODJ9_mw",
"avatar_url": "https://avatars.githubusercontent.com/u/211779483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/byungchul-sqzb",
"html_url": "https://github.com/byungchul-sqzb",
"followers_url": "https://api.github.com/users/byungchul-sqzb/followers",
"following_url": "https://api.github.com/users/byungchul-sqzb/following{/other_user}",
"gists_url": "https://api.github.com/users/byungchul-sqzb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/byungchul-sqzb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/byungchul-sqzb/subscriptions",
"organizations_url": "https://api.github.com/users/byungchul-sqzb/orgs",
"repos_url": "https://api.github.com/users/byungchul-sqzb/repos",
"events_url": "https://api.github.com/users/byungchul-sqzb/events{/privacy}",
"received_events_url": "https://api.github.com/users/byungchul-sqzb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T08:24:48 | 2025-06-12T08:25:39 | 2025-06-12T08:25:39 | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38781",
"html_url": "https://github.com/huggingface/transformers/pull/38781",
"diff_url": "https://github.com/huggingface/transformers/pull/38781.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38781.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "byungchul-sqzb",
"id": 211779483,
"node_id": "U_kgDODJ9_mw",
"avatar_url": "https://avatars.githubusercontent.com/u/211779483?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/byungchul-sqzb",
"html_url": "https://github.com/byungchul-sqzb",
"followers_url": "https://api.github.com/users/byungchul-sqzb/followers",
"following_url": "https://api.github.com/users/byungchul-sqzb/following{/other_user}",
"gists_url": "https://api.github.com/users/byungchul-sqzb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/byungchul-sqzb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/byungchul-sqzb/subscriptions",
"organizations_url": "https://api.github.com/users/byungchul-sqzb/orgs",
"repos_url": "https://api.github.com/users/byungchul-sqzb/repos",
"events_url": "https://api.github.com/users/byungchul-sqzb/events{/privacy}",
"received_events_url": "https://api.github.com/users/byungchul-sqzb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38781/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/38781/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38780 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38780/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38780/comments | https://api.github.com/repos/huggingface/transformers/issues/38780/events | https://github.com/huggingface/transformers/pull/38780 | 3,139,322,659 | PR_kwDOCUB6oc6aKq2t | 38,780 | Fix(informer): Correct tensor shape for input_size=1 | {
"login": "Flink-ddd",
"id": 180720690,
"node_id": "U_kgDOCsWUMg",
"avatar_url": "https://avatars.githubusercontent.com/u/180720690?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Flink-ddd",
"html_url": "https://github.com/Flink-ddd",
"followers_url": "https://api.github.com/users/Flink-ddd/followers",
"following_url": "https://api.github.com/users/Flink-ddd/following{/other_user}",
"gists_url": "https://api.github.com/users/Flink-ddd/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Flink-ddd/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Flink-ddd/subscriptions",
"organizations_url": "https://api.github.com/users/Flink-ddd/orgs",
"repos_url": "https://api.github.com/users/Flink-ddd/repos",
"events_url": "https://api.github.com/users/Flink-ddd/events{/privacy}",
"received_events_url": "https://api.github.com/users/Flink-ddd/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T08:20:56 | 2025-06-17T05:49:32 | 2025-06-17T05:49:32 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38780",
"html_url": "https://github.com/huggingface/transformers/pull/38780",
"diff_url": "https://github.com/huggingface/transformers/pull/38780.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38780.patch",
"merged_at": null
} | What does this PR do?
This PR fixes a RuntimeError that occurs in InformerModel when config.input_size is set to 1.
When input_size=1, the loc and scale tensors calculated by the scaler retained an extra dimension (e.g., shape [B, 1, 1] instead of [B, 1]). This incorrect shape caused a dimension mismatch error during the expand() operation later when creating expanded_static_feat.
This fix applies .squeeze(-1) to both the loc and scale tensors to ensure they have the correct dimensionality before being used. This resolves the crash and allows the model to run correctly with univariate time series data.
Fixes https://github.com/huggingface/transformers/issues/38745 | {
"login": "Flink-ddd",
"id": 180720690,
"node_id": "U_kgDOCsWUMg",
"avatar_url": "https://avatars.githubusercontent.com/u/180720690?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Flink-ddd",
"html_url": "https://github.com/Flink-ddd",
"followers_url": "https://api.github.com/users/Flink-ddd/followers",
"following_url": "https://api.github.com/users/Flink-ddd/following{/other_user}",
"gists_url": "https://api.github.com/users/Flink-ddd/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Flink-ddd/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Flink-ddd/subscriptions",
"organizations_url": "https://api.github.com/users/Flink-ddd/orgs",
"repos_url": "https://api.github.com/users/Flink-ddd/repos",
"events_url": "https://api.github.com/users/Flink-ddd/events{/privacy}",
"received_events_url": "https://api.github.com/users/Flink-ddd/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38780/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/38780/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38779 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38779/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38779/comments | https://api.github.com/repos/huggingface/transformers/issues/38779/events | https://github.com/huggingface/transformers/pull/38779 | 3,139,316,950 | PR_kwDOCUB6oc6aKpnS | 38,779 | [Hotfix] Fix style bot | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T08:18:51 | 2025-06-12T08:31:51 | 2025-06-12T08:20:37 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38779",
"html_url": "https://github.com/huggingface/transformers/pull/38779",
"diff_url": "https://github.com/huggingface/transformers/pull/38779.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38779.patch",
"merged_at": "2025-06-12T08:20:37"
} | # What does this PR do?
[Quickly merged to check everything is working now]
#38754 My comment https://github.com/huggingface/transformers/pull/38754#issuecomment-2963796287
is actually incorrect as my experiment on my own forked repository forgot to leave a comment after the style bot push.
We still some permission for this.
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38779/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/38779/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38778 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38778/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38778/comments | https://api.github.com/repos/huggingface/transformers/issues/38778/events | https://github.com/huggingface/transformers/pull/38778 | 3,139,177,063 | PR_kwDOCUB6oc6aKK3s | 38,778 | fix(generation): stop beam search per-instance when heuristic satisfied | {
"login": "guang-yng",
"id": 24318026,
"node_id": "MDQ6VXNlcjI0MzE4MDI2",
"avatar_url": "https://avatars.githubusercontent.com/u/24318026?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/guang-yng",
"html_url": "https://github.com/guang-yng",
"followers_url": "https://api.github.com/users/guang-yng/followers",
"following_url": "https://api.github.com/users/guang-yng/following{/other_user}",
"gists_url": "https://api.github.com/users/guang-yng/gists{/gist_id}",
"starred_url": "https://api.github.com/users/guang-yng/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/guang-yng/subscriptions",
"organizations_url": "https://api.github.com/users/guang-yng/orgs",
"repos_url": "https://api.github.com/users/guang-yng/repos",
"events_url": "https://api.github.com/users/guang-yng/events{/privacy}",
"received_events_url": "https://api.github.com/users/guang-yng/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T07:30:46 | 2025-07-08T09:26:27 | 2025-07-08T08:59:38 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38778",
"html_url": "https://github.com/huggingface/transformers/pull/38778",
"diff_url": "https://github.com/huggingface/transformers/pull/38778.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38778.patch",
"merged_at": "2025-07-08T08:59:38"
} | # What does this PR do?
This PR fixes a bug in beam search generation where early stopping heuristics (when `early_stopping=False`) was incorrectly applied across the entire batch, instead of per instance.
## 🔍 Problem
When `early_stopping=False`, the generation heuristic is supposed to stop generating once it’s unlikely that any beam will improve. However, the current behavior waits until all batch instances satisfy this heuristic before halting. This causes:
- Instances that are already “done” (according to the heuristic) to continue generating,
- Unnecessarily long and repetitive outputs,
- Inconsistent behavior depending on batch composition.
## ✅ Fix
We now apply the early stopping heuristic per-instance. As soon as a single instance has no beams left that can improve, generation for that instance is not used for updating answers. This restores expected behavior and leads to:
- Consistency between single-instance and batched generation,
- Parity with behavior in transformers < 4.50.
## 🧪 Reproduction Example
**Working case (single input)**
```py
from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig
olmo_model = AutoModelForCausalLM.from_pretrained("allenai/OLMo-2-0425-1B-Instruct")
olmo_model = olmo_model.to("cuda")
tokenizer = AutoTokenizer.from_pretrained("allenai/OLMo-2-0425-1B-Instruct", padding_side="left")
generation_config = GenerationConfig(
num_beams=10,
max_new_tokens=256,
length_penalty=2,
)
question = [ {"role": "user", "content": "What is 3+5?"} ]
question = tokenizer.apply_chat_template(
question, tokenize=False, add_generation_prompt=True, return_tensors="pt"
)
inputs = tokenizer(question, return_tensors="pt", padding=True).to("cuda")
outputs = olmo_model.generate(
**inputs,
generation_config=generation_config,
)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)
```
Produces clean output:
```
The sum of 3 and 5 is 8.
So, 3 + 5 = 8.
...
The sum of 3 and 5 is \(\boxed{8}\).
```
**Broken case (batched input)**
```py
question = [ {"role": "user", "content": "What is 3+5?"} ]
cot_question = [ {"role": "user", "content": "What is 3+5? Explain your reasoning step by step, and provide the final answer at the end."} ]
question = tokenizer.apply_chat_template(
question, tokenize=False, add_generation_prompt=True, return_tensors="pt"
)
cot_question = tokenizer.apply_chat_template(
cot_question, tokenize=False, add_generation_prompt=True, return_tensors="pt"
)
inputs = tokenizer([question, cot_question], return_tensors="pt", padding=True).to("cuda")
outputs = olmo_model.generate(
**inputs,
generation_config=generation_config,
)
responses = tokenizer.batch_decode(outputs, skip_special_tokens=True)
print(responses[0])
```
Produces repetitive output:
```
The sum of 3 and 5 is 8.
...
The sum of \(3 + 5\) is \(\boxed{8}\).
If you have any more questions or need further assistance, feel free to ask!
The sum of \(3 + 5\) is \(\boxed{8}\).
If you have any more questions or need further assistance, feel free to ask!
The sum of \(3 + 5\) is \(\boxed{8}\).
If you have any more questions or need further assistance, feel free to ask!
The sum of \(3 + 5\) is \(\boxed{8}\).
If you have any more questions or need further assistance, feel free to ask!
The sum of \(3 + 5\) is \(\boxed{8}\).
If you have any more questions or need further assistance, feel free to ask!
```
This undesirable repetition happens only when batched with longer examples. It can occur even with **default settings** like `length_penalty=1`.
This bug appears in recent versions with vectorized beam search. It **does not appear in transformers < 4.50.0**.
## Who can review?
@gante Could you please take a look at this? Thanks! | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38778/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/38778/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38777 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38777/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38777/comments | https://api.github.com/repos/huggingface/transformers/issues/38777/events | https://github.com/huggingface/transformers/pull/38777 | 3,139,168,759 | PR_kwDOCUB6oc6aKJAl | 38,777 | [docs] updated roberta model card | {
"login": "allmight05",
"id": 166301619,
"node_id": "U_kgDOCemPsw",
"avatar_url": "https://avatars.githubusercontent.com/u/166301619?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/allmight05",
"html_url": "https://github.com/allmight05",
"followers_url": "https://api.github.com/users/allmight05/followers",
"following_url": "https://api.github.com/users/allmight05/following{/other_user}",
"gists_url": "https://api.github.com/users/allmight05/gists{/gist_id}",
"starred_url": "https://api.github.com/users/allmight05/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/allmight05/subscriptions",
"organizations_url": "https://api.github.com/users/allmight05/orgs",
"repos_url": "https://api.github.com/users/allmight05/repos",
"events_url": "https://api.github.com/users/allmight05/events{/privacy}",
"received_events_url": "https://api.github.com/users/allmight05/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T07:28:39 | 2025-06-13T19:02:44 | 2025-06-13T19:02:44 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38777",
"html_url": "https://github.com/huggingface/transformers/pull/38777",
"diff_url": "https://github.com/huggingface/transformers/pull/38777.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38777.patch",
"merged_at": "2025-06-13T19:02:44"
} | # What does this PR do?
As suggested in this issue - #36979 - this PR updates the documentation of the Roberta model, which will now be aligned with the standardized format for all the docs.
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
## 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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
@stevhliu would you please review my PR? | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38777/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/38777/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38776 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38776/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38776/comments | https://api.github.com/repos/huggingface/transformers/issues/38776/events | https://github.com/huggingface/transformers/issues/38776 | 3,139,065,414 | I_kwDOCUB6oc67GlZG | 38,776 | [Bug][DOCS] TrainingArguments.__init__() got an unexpected keyword argument 'fsdp_strategy' | {
"login": "PT-10",
"id": 102211549,
"node_id": "U_kgDOBhef3Q",
"avatar_url": "https://avatars.githubusercontent.com/u/102211549?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/PT-10",
"html_url": "https://github.com/PT-10",
"followers_url": "https://api.github.com/users/PT-10/followers",
"following_url": "https://api.github.com/users/PT-10/following{/other_user}",
"gists_url": "https://api.github.com/users/PT-10/gists{/gist_id}",
"starred_url": "https://api.github.com/users/PT-10/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/PT-10/subscriptions",
"organizations_url": "https://api.github.com/users/PT-10/orgs",
"repos_url": "https://api.github.com/users/PT-10/repos",
"events_url": "https://api.github.com/users/PT-10/events{/privacy}",
"received_events_url": "https://api.github.com/users/PT-10/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-12T06:54:36 | 2025-06-13T15:32:41 | 2025-06-13T15:32:41 | CONTRIBUTOR | null | null | null | null | ### System Info
transformers v4.52.3
The docs at https://huggingface.co/docs/transformers/accelerate show -
However when working with accelerate in TrainingArgs I get the following issue of fsdp strategy not available
I checked the source code at https://github.com/huggingface/transformers/blob/v4.52.3/src/transformers/training_args.py and transformers v4.52.3 does not actually have 'fsdp_strategy' as an argument in the training args. I believe the docs are to be updated.
### Who can help?
@SunMarc @zach-huggingface
### 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
Task information:
Finetune "mistralai/Mistral-7B-Instruct-v0.1" using PEFT on "Josephgflowers/Finance-Instruct-500k"
1. Create config yaml as per like
```
fsdp_yaml = """
compute_environment: LOCAL_MACHINE
debug: false
distributed_type: FSDP
downcast_bf16: 'no'
fsdp_config:
fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
fsdp_backward_prefetch_policy: BACKWARD_PRE
fsdp_forward_prefetch: false
fsdp_cpu_ram_efficient_loading: true
fsdp_offload_params: false
fsdp_sharding_strategy: FULL_SHARD
fsdp_state_dict_type: SHARDED_STATE_DICT
fsdp_sync_module_states: true
fsdp_transformer_layer_cls_to_wrap: MistralDecoderLayer
fsdp_use_orig_params: true
machine_rank: 0
main_training_function: main
mixed_precision: fp16
num_machines: 1
num_processes: 2
rdzv_backend: static
same_network: true
tpu_env: []
tpu_use_cluster: false
tpu_use_sudo: false
use_cpu: false
"""
with open("path/to/config/yaml", "w") as f:
f.write(fsdp_yaml)
```
2. Initialize trainer arguments
```
from transformers import TrainingArguments, Trainer, DataCollatorForLanguageModeling
training_args = TrainingArguments(
output_dir="path/to/output_dir",
per_device_train_batch_size=2,
gradient_accumulation_steps=4,
learning_rate=2e-4,
num_train_epochs=1,
fsdp_config="path/to/config/yaml",
fsdp_strategy="full_shard",
fp16=True,
logging_steps=10,
save_steps=100,
save_total_limit=1
)
```
### Expected behavior
Training args being initialized as per docs and then being able to pass them to the Trainer class | {
"login": "SunMarc",
"id": 57196510,
"node_id": "MDQ6VXNlcjU3MTk2NTEw",
"avatar_url": "https://avatars.githubusercontent.com/u/57196510?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SunMarc",
"html_url": "https://github.com/SunMarc",
"followers_url": "https://api.github.com/users/SunMarc/followers",
"following_url": "https://api.github.com/users/SunMarc/following{/other_user}",
"gists_url": "https://api.github.com/users/SunMarc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SunMarc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SunMarc/subscriptions",
"organizations_url": "https://api.github.com/users/SunMarc/orgs",
"repos_url": "https://api.github.com/users/SunMarc/repos",
"events_url": "https://api.github.com/users/SunMarc/events{/privacy}",
"received_events_url": "https://api.github.com/users/SunMarc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38776/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/38776/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38775 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38775/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38775/comments | https://api.github.com/repos/huggingface/transformers/issues/38775/events | https://github.com/huggingface/transformers/pull/38775 | 3,139,013,249 | PR_kwDOCUB6oc6aJm5j | 38,775 | [docs] update cache docs with new info | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T06:32:12 | 2025-06-13T07:10:57 | 2025-06-13T07:10:56 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38775",
"html_url": "https://github.com/huggingface/transformers/pull/38775",
"diff_url": "https://github.com/huggingface/transformers/pull/38775.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38775.patch",
"merged_at": "2025-06-13T07:10:56"
} | # What does this PR do?
Fixes https://github.com/huggingface/transformers/issues/38727 and adds tiny block about possible issues with chat templates
| {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38775/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/38775/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38774 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38774/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38774/comments | https://api.github.com/repos/huggingface/transformers/issues/38774/events | https://github.com/huggingface/transformers/pull/38774 | 3,138,946,566 | PR_kwDOCUB6oc6aJYiX | 38,774 | updated Roberta model card Documentation | {
"login": "allmight05",
"id": 166301619,
"node_id": "U_kgDOCemPsw",
"avatar_url": "https://avatars.githubusercontent.com/u/166301619?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/allmight05",
"html_url": "https://github.com/allmight05",
"followers_url": "https://api.github.com/users/allmight05/followers",
"following_url": "https://api.github.com/users/allmight05/following{/other_user}",
"gists_url": "https://api.github.com/users/allmight05/gists{/gist_id}",
"starred_url": "https://api.github.com/users/allmight05/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/allmight05/subscriptions",
"organizations_url": "https://api.github.com/users/allmight05/orgs",
"repos_url": "https://api.github.com/users/allmight05/repos",
"events_url": "https://api.github.com/users/allmight05/events{/privacy}",
"received_events_url": "https://api.github.com/users/allmight05/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T05:59:22 | 2025-06-12T07:02:49 | 2025-06-12T07:01:17 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38774",
"html_url": "https://github.com/huggingface/transformers/pull/38774",
"diff_url": "https://github.com/huggingface/transformers/pull/38774.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38774.patch",
"merged_at": null
} | # What does this PR do?
As suggested in this issue - https://github.com/huggingface/transformers/issues/36979 - this PR updates the documentation of the Roberta model, which will now be aligned with the standardized format for all the docs.
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
@stevhliu can you please review it? | {
"login": "allmight05",
"id": 166301619,
"node_id": "U_kgDOCemPsw",
"avatar_url": "https://avatars.githubusercontent.com/u/166301619?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/allmight05",
"html_url": "https://github.com/allmight05",
"followers_url": "https://api.github.com/users/allmight05/followers",
"following_url": "https://api.github.com/users/allmight05/following{/other_user}",
"gists_url": "https://api.github.com/users/allmight05/gists{/gist_id}",
"starred_url": "https://api.github.com/users/allmight05/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/allmight05/subscriptions",
"organizations_url": "https://api.github.com/users/allmight05/orgs",
"repos_url": "https://api.github.com/users/allmight05/repos",
"events_url": "https://api.github.com/users/allmight05/events{/privacy}",
"received_events_url": "https://api.github.com/users/allmight05/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38774/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/38774/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38773 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38773/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38773/comments | https://api.github.com/repos/huggingface/transformers/issues/38773/events | https://github.com/huggingface/transformers/pull/38773 | 3,138,682,059 | PR_kwDOCUB6oc6aIe4f | 38,773 | Add missing div in Pegasus model card | {
"login": "dross20",
"id": 73395516,
"node_id": "MDQ6VXNlcjczMzk1NTE2",
"avatar_url": "https://avatars.githubusercontent.com/u/73395516?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/dross20",
"html_url": "https://github.com/dross20",
"followers_url": "https://api.github.com/users/dross20/followers",
"following_url": "https://api.github.com/users/dross20/following{/other_user}",
"gists_url": "https://api.github.com/users/dross20/gists{/gist_id}",
"starred_url": "https://api.github.com/users/dross20/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/dross20/subscriptions",
"organizations_url": "https://api.github.com/users/dross20/orgs",
"repos_url": "https://api.github.com/users/dross20/repos",
"events_url": "https://api.github.com/users/dross20/events{/privacy}",
"received_events_url": "https://api.github.com/users/dross20/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-12T03:19:44 | 2025-06-12T17:27:07 | 2025-06-12T17:27:07 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38773",
"html_url": "https://github.com/huggingface/transformers/pull/38773",
"diff_url": "https://github.com/huggingface/transformers/pull/38773.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38773.patch",
"merged_at": "2025-06-12T17:27:07"
} | # What does this PR do?
I updated the Pegasus model card in #38675 but missed a single `<div>`, causing the badges to appear misaligned on the page. This PR adds the missing div and fixes the alignment issue.
## Before submitting
- [X] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
## Who can review?
@stevhliu
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38773/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/38773/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38772 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38772/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38772/comments | https://api.github.com/repos/huggingface/transformers/issues/38772/events | https://github.com/huggingface/transformers/pull/38772 | 3,138,357,982 | PR_kwDOCUB6oc6aHanv | 38,772 | Simplify and update trl examples | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T23:36:00 | 2025-06-13T12:03:51 | 2025-06-13T12:03:50 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38772",
"html_url": "https://github.com/huggingface/transformers/pull/38772",
"diff_url": "https://github.com/huggingface/transformers/pull/38772.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38772.patch",
"merged_at": "2025-06-13T12:03:50"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38772/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/38772/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38771 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38771/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38771/comments | https://api.github.com/repos/huggingface/transformers/issues/38771/events | https://github.com/huggingface/transformers/issues/38771 | 3,138,280,332 | I_kwDOCUB6oc67DluM | 38,771 | device_map='auto' coupled with tp_plan='auto' | {
"login": "weifengpy",
"id": 134637289,
"node_id": "U_kgDOCAZm6Q",
"avatar_url": "https://avatars.githubusercontent.com/u/134637289?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/weifengpy",
"html_url": "https://github.com/weifengpy",
"followers_url": "https://api.github.com/users/weifengpy/followers",
"following_url": "https://api.github.com/users/weifengpy/following{/other_user}",
"gists_url": "https://api.github.com/users/weifengpy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/weifengpy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/weifengpy/subscriptions",
"organizations_url": "https://api.github.com/users/weifengpy/orgs",
"repos_url": "https://api.github.com/users/weifengpy/repos",
"events_url": "https://api.github.com/users/weifengpy/events{/privacy}",
"received_events_url": "https://api.github.com/users/weifengpy/received_events",
"type": "User",
"user_view_type": "public",
"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 | [] | null | [] | 2025-06-11T22:38:16 | 2025-07-11T04:54:37 | 2025-07-03T09:07:13 | NONE | null | null | null | null | ### Feature request
Hi from pytorch distributed! Thanks for showcasing pytorch APIs
device_map='auto' and tp_plan='auto' are somehow coupled right now:
https://github.com/huggingface/transformers/blob/main/src/transformers/modeling_utils.py#L4280-L4283
if people load model with `device_map='auto'`, they will ended up with a TP-enabled model with model.parameters() as DTensor. Once they train it, DDP will complain about DTensor. If they set `device_map='cuda'`, the model is loaded as plain module without TP and it works with DDP
Curious if we want to decouple `device_map='auto'` with `tp_plan='auto'`? I found this from pytorch users https://github.com/pytorch/pytorch/issues/155463
### Motivation
d
### Your contribution
d | {
"login": "SunMarc",
"id": 57196510,
"node_id": "MDQ6VXNlcjU3MTk2NTEw",
"avatar_url": "https://avatars.githubusercontent.com/u/57196510?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SunMarc",
"html_url": "https://github.com/SunMarc",
"followers_url": "https://api.github.com/users/SunMarc/followers",
"following_url": "https://api.github.com/users/SunMarc/following{/other_user}",
"gists_url": "https://api.github.com/users/SunMarc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SunMarc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SunMarc/subscriptions",
"organizations_url": "https://api.github.com/users/SunMarc/orgs",
"repos_url": "https://api.github.com/users/SunMarc/repos",
"events_url": "https://api.github.com/users/SunMarc/events{/privacy}",
"received_events_url": "https://api.github.com/users/SunMarc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38771/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/38771/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38770 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38770/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38770/comments | https://api.github.com/repos/huggingface/transformers/issues/38770/events | https://github.com/huggingface/transformers/pull/38770 | 3,138,249,292 | PR_kwDOCUB6oc6aHDBf | 38,770 | Update PULL_REQUEST_TEMPLATE.md | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T22:15:06 | 2025-06-12T12:03:35 | 2025-06-12T12:03:33 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38770",
"html_url": "https://github.com/huggingface/transformers/pull/38770",
"diff_url": "https://github.com/huggingface/transformers/pull/38770.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38770.patch",
"merged_at": "2025-06-12T12:03:33"
} | Sometimes I miss important trainer iterations. See https://github.com/huggingface/trl/pull/3568 | {
"login": "SunMarc",
"id": 57196510,
"node_id": "MDQ6VXNlcjU3MTk2NTEw",
"avatar_url": "https://avatars.githubusercontent.com/u/57196510?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/SunMarc",
"html_url": "https://github.com/SunMarc",
"followers_url": "https://api.github.com/users/SunMarc/followers",
"following_url": "https://api.github.com/users/SunMarc/following{/other_user}",
"gists_url": "https://api.github.com/users/SunMarc/gists{/gist_id}",
"starred_url": "https://api.github.com/users/SunMarc/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/SunMarc/subscriptions",
"organizations_url": "https://api.github.com/users/SunMarc/orgs",
"repos_url": "https://api.github.com/users/SunMarc/repos",
"events_url": "https://api.github.com/users/SunMarc/events{/privacy}",
"received_events_url": "https://api.github.com/users/SunMarc/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38770/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/38770/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38769 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38769/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38769/comments | https://api.github.com/repos/huggingface/transformers/issues/38769/events | https://github.com/huggingface/transformers/pull/38769 | 3,138,182,865 | PR_kwDOCUB6oc6aG0Nq | 38,769 | LlamaAttention forward function type hint is incorrect #38739 | {
"login": "ArkVex",
"id": 159469387,
"node_id": "U_kgDOCYFPSw",
"avatar_url": "https://avatars.githubusercontent.com/u/159469387?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ArkVex",
"html_url": "https://github.com/ArkVex",
"followers_url": "https://api.github.com/users/ArkVex/followers",
"following_url": "https://api.github.com/users/ArkVex/following{/other_user}",
"gists_url": "https://api.github.com/users/ArkVex/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ArkVex/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ArkVex/subscriptions",
"organizations_url": "https://api.github.com/users/ArkVex/orgs",
"repos_url": "https://api.github.com/users/ArkVex/repos",
"events_url": "https://api.github.com/users/ArkVex/events{/privacy}",
"received_events_url": "https://api.github.com/users/ArkVex/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T21:39:47 | 2025-06-12T11:25:49 | 2025-06-12T11:25:49 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38769",
"html_url": "https://github.com/huggingface/transformers/pull/38769",
"diff_url": "https://github.com/huggingface/transformers/pull/38769.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38769.patch",
"merged_at": null
} | Hi, this PR fixes a small issue in the LlamaAttention class. The return type in the forward method currently shows three values, but the function actually returns only two. This seems to have been missed during the attention refactor (possibly in PR #35235).
I’ve updated the type hint to reflect the actual return values, just to avoid confusion for anyone reading or using the code. Let me know if any other changes are needed. Happy to help! | {
"login": "ArkVex",
"id": 159469387,
"node_id": "U_kgDOCYFPSw",
"avatar_url": "https://avatars.githubusercontent.com/u/159469387?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ArkVex",
"html_url": "https://github.com/ArkVex",
"followers_url": "https://api.github.com/users/ArkVex/followers",
"following_url": "https://api.github.com/users/ArkVex/following{/other_user}",
"gists_url": "https://api.github.com/users/ArkVex/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ArkVex/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ArkVex/subscriptions",
"organizations_url": "https://api.github.com/users/ArkVex/orgs",
"repos_url": "https://api.github.com/users/ArkVex/repos",
"events_url": "https://api.github.com/users/ArkVex/events{/privacy}",
"received_events_url": "https://api.github.com/users/ArkVex/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38769/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/38769/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38768 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38768/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38768/comments | https://api.github.com/repos/huggingface/transformers/issues/38768/events | https://github.com/huggingface/transformers/pull/38768 | 3,137,961,529 | PR_kwDOCUB6oc6aGC5y | 38,768 | Initialize flash attn flag | {
"login": "farnasirim",
"id": 8123364,
"node_id": "MDQ6VXNlcjgxMjMzNjQ=",
"avatar_url": "https://avatars.githubusercontent.com/u/8123364?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/farnasirim",
"html_url": "https://github.com/farnasirim",
"followers_url": "https://api.github.com/users/farnasirim/followers",
"following_url": "https://api.github.com/users/farnasirim/following{/other_user}",
"gists_url": "https://api.github.com/users/farnasirim/gists{/gist_id}",
"starred_url": "https://api.github.com/users/farnasirim/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/farnasirim/subscriptions",
"organizations_url": "https://api.github.com/users/farnasirim/orgs",
"repos_url": "https://api.github.com/users/farnasirim/repos",
"events_url": "https://api.github.com/users/farnasirim/events{/privacy}",
"received_events_url": "https://api.github.com/users/farnasirim/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T20:01:21 | 2025-06-12T14:06:37 | 2025-06-12T14:06:13 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38768",
"html_url": "https://github.com/huggingface/transformers/pull/38768",
"diff_url": "https://github.com/huggingface/transformers/pull/38768.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38768.patch",
"merged_at": "2025-06-12T14:06:13"
} | _flash_supports_window_size is used further down in this file and relied on by e.g. [ring-flash-attention](https://github.com/zhuzilin/ring-flash-attention/blob/123f924/ring_flash_attn/adapters/hf_adapter.py#L9-L11). Even though it is an unexported name, it still makes sense to keep the state of `globals()` in this file consistent.
# What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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).
No
- [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request),
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).
- [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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38768/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/38768/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38767 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38767/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38767/comments | https://api.github.com/repos/huggingface/transformers/issues/38767/events | https://github.com/huggingface/transformers/pull/38767 | 3,137,799,574 | PR_kwDOCUB6oc6aFfbG | 38,767 | Debug style | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T18:46:45 | 2025-06-12T09:13:49 | 2025-06-12T09:13:49 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38767",
"html_url": "https://github.com/huggingface/transformers/pull/38767",
"diff_url": "https://github.com/huggingface/transformers/pull/38767.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38767.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38767/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/38767/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38766 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38766/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38766/comments | https://api.github.com/repos/huggingface/transformers/issues/38766/events | https://github.com/huggingface/transformers/pull/38766 | 3,137,689,160 | PR_kwDOCUB6oc6aFHQH | 38,766 | Fix Typos in Comments: "quantitation" → "quantization", "averege" → "average" | {
"login": "leopardracer",
"id": 136604165,
"node_id": "U_kgDOCCRqBQ",
"avatar_url": "https://avatars.githubusercontent.com/u/136604165?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/leopardracer",
"html_url": "https://github.com/leopardracer",
"followers_url": "https://api.github.com/users/leopardracer/followers",
"following_url": "https://api.github.com/users/leopardracer/following{/other_user}",
"gists_url": "https://api.github.com/users/leopardracer/gists{/gist_id}",
"starred_url": "https://api.github.com/users/leopardracer/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/leopardracer/subscriptions",
"organizations_url": "https://api.github.com/users/leopardracer/orgs",
"repos_url": "https://api.github.com/users/leopardracer/repos",
"events_url": "https://api.github.com/users/leopardracer/events{/privacy}",
"received_events_url": "https://api.github.com/users/leopardracer/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T18:03:33 | 2025-06-12T14:05:21 | 2025-06-12T14:04:39 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38766",
"html_url": "https://github.com/huggingface/transformers/pull/38766",
"diff_url": "https://github.com/huggingface/transformers/pull/38766.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38766.patch",
"merged_at": "2025-06-12T14:04:39"
} |
Description:
This pull request corrects minor spelling mistakes in code comments across two files:
- In `convert_llama4_weights_to_hf.py`, the word "quantitation" is corrected to "quantization".
- In `modeling_visual_bert.py`, the word "averege" is corrected to "average".
These changes improve code readability and maintain consistency in terminology. No functional code changes are included.
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38766/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/38766/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38765 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38765/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38765/comments | https://api.github.com/repos/huggingface/transformers/issues/38765/events | https://github.com/huggingface/transformers/issues/38765 | 3,137,586,966 | I_kwDOCUB6oc67A8cW | 38,765 | Add FastVLM from CVPR 2025 | {
"login": "kamila-chay",
"id": 201148875,
"node_id": "U_kgDOC_1Jyw",
"avatar_url": "https://avatars.githubusercontent.com/u/201148875?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/kamila-chay",
"html_url": "https://github.com/kamila-chay",
"followers_url": "https://api.github.com/users/kamila-chay/followers",
"following_url": "https://api.github.com/users/kamila-chay/following{/other_user}",
"gists_url": "https://api.github.com/users/kamila-chay/gists{/gist_id}",
"starred_url": "https://api.github.com/users/kamila-chay/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/kamila-chay/subscriptions",
"organizations_url": "https://api.github.com/users/kamila-chay/orgs",
"repos_url": "https://api.github.com/users/kamila-chay/repos",
"events_url": "https://api.github.com/users/kamila-chay/events{/privacy}",
"received_events_url": "https://api.github.com/users/kamila-chay/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 1843244711,
"node_id": "MDU6TGFiZWwxODQzMjQ0NzEx",
"url": "https://api.github.com/repos/huggingface/transformers/labels/New%20model",
"name": "New model",
"color": "fbca04",
"default": false,
"description": ""
}
] | open | false | null | [] | null | [] | 2025-06-11T17:21:03 | 2025-09-18T19:49:14 | null | NONE | null | null | null | null | ### Model description
Apple recently released FastVLM, a new vision-language model introduced at CVPR 2025, which significantly improves on previous models in the LLaVA family.
The smallest FastVLM variant outperforms LLaVA-OneVision-0.5B, achieving:
- 85× faster Time-to-First-Token (TTFT)
- 3.4× smaller vision encoder
Its ultra-low latency enables real-time deployment on mobile and edge devices, as demonstrated in the official demo. Given its strong performance and efficiency, I believe FastVLM would be a valuable addition to this repo.
If you also think so, I'd love to contribute this model.
### Open source status
- [x] The model implementation is available
- [x] The model weights are available
### Provide useful links for the implementation
Paper: https://www.arxiv.org/abs/2412.13303
Official repo: https://github.com/apple/ml-fastvlm | null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38765/reactions",
"total_count": 1,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 1
} | https://api.github.com/repos/huggingface/transformers/issues/38765/timeline | null | null | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | false |
https://api.github.com/repos/huggingface/transformers/issues/38764 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38764/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38764/comments | https://api.github.com/repos/huggingface/transformers/issues/38764/events | https://github.com/huggingface/transformers/pull/38764 | 3,137,398,701 | PR_kwDOCUB6oc6aEHag | 38,764 | Fixes issue #38709 for CLIP and XCLIP | {
"login": "Vishu26",
"id": 24605821,
"node_id": "MDQ6VXNlcjI0NjA1ODIx",
"avatar_url": "https://avatars.githubusercontent.com/u/24605821?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Vishu26",
"html_url": "https://github.com/Vishu26",
"followers_url": "https://api.github.com/users/Vishu26/followers",
"following_url": "https://api.github.com/users/Vishu26/following{/other_user}",
"gists_url": "https://api.github.com/users/Vishu26/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Vishu26/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Vishu26/subscriptions",
"organizations_url": "https://api.github.com/users/Vishu26/orgs",
"repos_url": "https://api.github.com/users/Vishu26/repos",
"events_url": "https://api.github.com/users/Vishu26/events{/privacy}",
"received_events_url": "https://api.github.com/users/Vishu26/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-11T16:10:22 | 2025-06-12T14:16:51 | null | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38764",
"html_url": "https://github.com/huggingface/transformers/pull/38764",
"diff_url": "https://github.com/huggingface/transformers/pull/38764.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38764.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
Fixes #38709 (at least for CLIP and XCLIP for now and is fully backward compatible and existing functionality remains unaffected).
Previously, `get_<>_features` function in CLIP and XCLIP always return the `pooled_output`. Now, if `output_hidden_states` is set, the function returns a dict. It is set to `None` by default, so the existing behavior will be unaffected.
## 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#create-a-pull-request),
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.
- [ ] 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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38764/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/38764/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38763 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38763/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38763/comments | https://api.github.com/repos/huggingface/transformers/issues/38763/events | https://github.com/huggingface/transformers/pull/38763 | 3,137,319,967 | PR_kwDOCUB6oc6aD18P | 38,763 | Update repo consistency check | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T15:45:30 | 2025-06-11T16:02:05 | 2025-06-11T16:02:04 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38763",
"html_url": "https://github.com/huggingface/transformers/pull/38763",
"diff_url": "https://github.com/huggingface/transformers/pull/38763.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38763.patch",
"merged_at": "2025-06-11T16:02:04"
} | The old check is wrong, and doesn't actually test things! My bad, this PR fixes it. | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38763/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/38763/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38762 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38762/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38762/comments | https://api.github.com/repos/huggingface/transformers/issues/38762/events | https://github.com/huggingface/transformers/pull/38762 | 3,137,303,835 | PR_kwDOCUB6oc6aDyZA | 38,762 | Reword README in light of model definitions | {
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "https://api.github.com/users/LysandreJik/followers",
"following_url": "https://api.github.com/users/LysandreJik/following{/other_user}",
"gists_url": "https://api.github.com/users/LysandreJik/gists{/gist_id}",
"starred_url": "https://api.github.com/users/LysandreJik/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/LysandreJik/subscriptions",
"organizations_url": "https://api.github.com/users/LysandreJik/orgs",
"repos_url": "https://api.github.com/users/LysandreJik/repos",
"events_url": "https://api.github.com/users/LysandreJik/events{/privacy}",
"received_events_url": "https://api.github.com/users/LysandreJik/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T15:40:09 | 2025-06-12T13:43:32 | 2025-06-12T13:43:31 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38762",
"html_url": "https://github.com/huggingface/transformers/pull/38762",
"diff_url": "https://github.com/huggingface/transformers/pull/38762.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38762.patch",
"merged_at": "2025-06-12T13:43:31"
} | null | {
"login": "LysandreJik",
"id": 30755778,
"node_id": "MDQ6VXNlcjMwNzU1Nzc4",
"avatar_url": "https://avatars.githubusercontent.com/u/30755778?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/LysandreJik",
"html_url": "https://github.com/LysandreJik",
"followers_url": "https://api.github.com/users/LysandreJik/followers",
"following_url": "https://api.github.com/users/LysandreJik/following{/other_user}",
"gists_url": "https://api.github.com/users/LysandreJik/gists{/gist_id}",
"starred_url": "https://api.github.com/users/LysandreJik/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/LysandreJik/subscriptions",
"organizations_url": "https://api.github.com/users/LysandreJik/orgs",
"repos_url": "https://api.github.com/users/LysandreJik/repos",
"events_url": "https://api.github.com/users/LysandreJik/events{/privacy}",
"received_events_url": "https://api.github.com/users/LysandreJik/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38762/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/38762/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38761 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38761/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38761/comments | https://api.github.com/repos/huggingface/transformers/issues/38761/events | https://github.com/huggingface/transformers/pull/38761 | 3,137,241,165 | PR_kwDOCUB6oc6aDkrg | 38,761 | [BugFix] QA pipeline edge case: `align_to_words=True` in `QuestionAnsweringPipeline` can lead to duplicate answers | {
"login": "Vixel2006",
"id": 166058059,
"node_id": "U_kgDOCeXYSw",
"avatar_url": "https://avatars.githubusercontent.com/u/166058059?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Vixel2006",
"html_url": "https://github.com/Vixel2006",
"followers_url": "https://api.github.com/users/Vixel2006/followers",
"following_url": "https://api.github.com/users/Vixel2006/following{/other_user}",
"gists_url": "https://api.github.com/users/Vixel2006/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Vixel2006/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Vixel2006/subscriptions",
"organizations_url": "https://api.github.com/users/Vixel2006/orgs",
"repos_url": "https://api.github.com/users/Vixel2006/repos",
"events_url": "https://api.github.com/users/Vixel2006/events{/privacy}",
"received_events_url": "https://api.github.com/users/Vixel2006/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T15:21:08 | 2025-07-03T06:28:00 | 2025-06-16T15:01:22 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38761",
"html_url": "https://github.com/huggingface/transformers/pull/38761",
"diff_url": "https://github.com/huggingface/transformers/pull/38761.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38761.patch",
"merged_at": "2025-06-16T15:01:22"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
Fixes #26286
Solves the duplicate answers problem when `align_to_words=True` which is default in the `question_answering` pipeline by keeping track of the answers history and if the answer is already in the answers list we just merge the two answers by summing their probabilities.
```py
from transformers import pipeline
answers = pipeline("question-answering", model="deepset/tinyroberta-squad2")(
question="Who is the chancellor of Germany?",
context="Angela Merkel was the chancellor of Germany.",
top_k=10,
)
for answer in answers:
print(answer)
```
This solution makes sure that there is no duplicate answers in the output but we don't know how many answers are returned.
## Who can review?
[@Rocketknight1](https://github.com/Rocketknight1) [@ArthurZucker ](https://github.com/ArthurZucker)
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38761/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/38761/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38760 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38760/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38760/comments | https://api.github.com/repos/huggingface/transformers/issues/38760/events | https://github.com/huggingface/transformers/pull/38760 | 3,137,110,308 | PR_kwDOCUB6oc6aDHaH | 38,760 | Prepare for TF+Jax deprecation | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T14:44:39 | 2025-06-11T15:03:33 | 2025-06-11T15:03:31 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38760",
"html_url": "https://github.com/huggingface/transformers/pull/38760",
"diff_url": "https://github.com/huggingface/transformers/pull/38760.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38760.patch",
"merged_at": "2025-06-11T15:03:31"
} | This PR removes a lot of TF + Jax testing in `repo-consistency` and `.circleci`. In a follow-up PR, I'll delete a lot of the test files, but this needs to be in a separate PR because repo-consistency checks run against the target branch, not the PR branch, and so they will fail if we remove other tests before merging this. | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38760/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/38760/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38759 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38759/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38759/comments | https://api.github.com/repos/huggingface/transformers/issues/38759/events | https://github.com/huggingface/transformers/pull/38759 | 3,137,091,339 | PR_kwDOCUB6oc6aDDQJ | 38,759 | Add chatterbox | {
"login": "eustlb",
"id": 94853470,
"node_id": "U_kgDOBadZXg",
"avatar_url": "https://avatars.githubusercontent.com/u/94853470?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/eustlb",
"html_url": "https://github.com/eustlb",
"followers_url": "https://api.github.com/users/eustlb/followers",
"following_url": "https://api.github.com/users/eustlb/following{/other_user}",
"gists_url": "https://api.github.com/users/eustlb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/eustlb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/eustlb/subscriptions",
"organizations_url": "https://api.github.com/users/eustlb/orgs",
"repos_url": "https://api.github.com/users/eustlb/repos",
"events_url": "https://api.github.com/users/eustlb/events{/privacy}",
"received_events_url": "https://api.github.com/users/eustlb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-11T14:39:06 | 2025-06-12T09:00:49 | null | CONTRIBUTOR | null | null | true | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38759",
"html_url": "https://github.com/huggingface/transformers/pull/38759",
"diff_url": "https://github.com/huggingface/transformers/pull/38759.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38759.patch",
"merged_at": null
} | # What does this PR do?
⚠️ This PR depends on and shouldn't be undrafted until both are merged:
https://github.com/huggingface/transformers/pull/38757
https://github.com/huggingface/transformers/pull/38756
Should be regularly rebased from the temporary aggregated branch [add-chatterbox-base](https://github.com/eustlb/transformers/tree/add-chatterbox-base) following the workflow:
```bash
git checkout main
git pull
git branch -D add-chatterbox-base
git checkout -b add-chatterbox-base
git merge add-hiftnet
git merge add-s3tokenizer
git checkout add-chatterbox
git rebase add-chatterbox-base
git push --force-with-lease
```
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38759/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/38759/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38758 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38758/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38758/comments | https://api.github.com/repos/huggingface/transformers/issues/38758/events | https://github.com/huggingface/transformers/pull/38758 | 3,136,986,995 | PR_kwDOCUB6oc6aCsBN | 38,758 | Deprecate TF + JAX | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T14:10:07 | 2025-06-11T16:28:07 | 2025-06-11T16:28:06 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38758",
"html_url": "https://github.com/huggingface/transformers/pull/38758",
"diff_url": "https://github.com/huggingface/transformers/pull/38758.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38758.patch",
"merged_at": "2025-06-11T16:28:06"
} | The time has finally come :gun: :tumbler_glass: | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38758/reactions",
"total_count": 79,
"+1": 2,
"-1": 10,
"laugh": 0,
"hooray": 0,
"confused": 15,
"heart": 16,
"rocket": 35,
"eyes": 1
} | https://api.github.com/repos/huggingface/transformers/issues/38758/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38757 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38757/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38757/comments | https://api.github.com/repos/huggingface/transformers/issues/38757/events | https://github.com/huggingface/transformers/pull/38757 | 3,136,974,692 | PR_kwDOCUB6oc6aCpTw | 38,757 | Add S3 tokenizer | {
"login": "eustlb",
"id": 94853470,
"node_id": "U_kgDOBadZXg",
"avatar_url": "https://avatars.githubusercontent.com/u/94853470?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/eustlb",
"html_url": "https://github.com/eustlb",
"followers_url": "https://api.github.com/users/eustlb/followers",
"following_url": "https://api.github.com/users/eustlb/following{/other_user}",
"gists_url": "https://api.github.com/users/eustlb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/eustlb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/eustlb/subscriptions",
"organizations_url": "https://api.github.com/users/eustlb/orgs",
"repos_url": "https://api.github.com/users/eustlb/repos",
"events_url": "https://api.github.com/users/eustlb/events{/privacy}",
"received_events_url": "https://api.github.com/users/eustlb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-11T14:06:14 | 2025-06-11T14:47:36 | null | CONTRIBUTOR | null | null | true | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38757",
"html_url": "https://github.com/huggingface/transformers/pull/38757",
"diff_url": "https://github.com/huggingface/transformers/pull/38757.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38757.patch",
"merged_at": null
} | # What does this PR do?
WIP
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38757/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/38757/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38756 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38756/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38756/comments | https://api.github.com/repos/huggingface/transformers/issues/38756/events | https://github.com/huggingface/transformers/pull/38756 | 3,136,973,024 | PR_kwDOCUB6oc6aCo82 | 38,756 | Add HiFTNet | {
"login": "eustlb",
"id": 94853470,
"node_id": "U_kgDOBadZXg",
"avatar_url": "https://avatars.githubusercontent.com/u/94853470?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/eustlb",
"html_url": "https://github.com/eustlb",
"followers_url": "https://api.github.com/users/eustlb/followers",
"following_url": "https://api.github.com/users/eustlb/following{/other_user}",
"gists_url": "https://api.github.com/users/eustlb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/eustlb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/eustlb/subscriptions",
"organizations_url": "https://api.github.com/users/eustlb/orgs",
"repos_url": "https://api.github.com/users/eustlb/repos",
"events_url": "https://api.github.com/users/eustlb/events{/privacy}",
"received_events_url": "https://api.github.com/users/eustlb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-11T14:05:41 | 2025-06-17T12:02:29 | null | CONTRIBUTOR | null | null | true | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38756",
"html_url": "https://github.com/huggingface/transformers/pull/38756",
"diff_url": "https://github.com/huggingface/transformers/pull/38756.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38756.patch",
"merged_at": null
} | # What does this PR do?
WIP | null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38756/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/38756/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38755 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38755/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38755/comments | https://api.github.com/repos/huggingface/transformers/issues/38755/events | https://github.com/huggingface/transformers/pull/38755 | 3,136,921,271 | PR_kwDOCUB6oc6aCde1 | 38,755 | Add SmolLM3 | {
"login": "anton-l",
"id": 26864830,
"node_id": "MDQ6VXNlcjI2ODY0ODMw",
"avatar_url": "https://avatars.githubusercontent.com/u/26864830?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/anton-l",
"html_url": "https://github.com/anton-l",
"followers_url": "https://api.github.com/users/anton-l/followers",
"following_url": "https://api.github.com/users/anton-l/following{/other_user}",
"gists_url": "https://api.github.com/users/anton-l/gists{/gist_id}",
"starred_url": "https://api.github.com/users/anton-l/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/anton-l/subscriptions",
"organizations_url": "https://api.github.com/users/anton-l/orgs",
"repos_url": "https://api.github.com/users/anton-l/repos",
"events_url": "https://api.github.com/users/anton-l/events{/privacy}",
"received_events_url": "https://api.github.com/users/anton-l/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T13:51:42 | 2025-06-25T15:12:16 | 2025-06-25T15:12:15 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38755",
"html_url": "https://github.com/huggingface/transformers/pull/38755",
"diff_url": "https://github.com/huggingface/transformers/pull/38755.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38755.patch",
"merged_at": "2025-06-25T15:12:15"
} | # What does this PR do?
Adds the SmolLM3 modular arch
## 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#create-a-pull-request),
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?
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38755/reactions",
"total_count": 8,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 8,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/38755/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38754 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38754/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38754/comments | https://api.github.com/repos/huggingface/transformers/issues/38754/events | https://github.com/huggingface/transformers/pull/38754 | 3,136,771,269 | PR_kwDOCUB6oc6aB8DH | 38,754 | Make style bot trigger CI after push | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T13:09:06 | 2025-06-11T18:40:07 | 2025-06-11T18:40:05 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38754",
"html_url": "https://github.com/huggingface/transformers/pull/38754",
"diff_url": "https://github.com/huggingface/transformers/pull/38754.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38754.patch",
"merged_at": "2025-06-11T18:40:05"
} | # What does this PR do?
Let's see if it works. I tested on my own ydshieh/transformers with an external contributor `ydshieh2`. | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38754/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/38754/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38753 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38753/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38753/comments | https://api.github.com/repos/huggingface/transformers/issues/38753/events | https://github.com/huggingface/transformers/issues/38753 | 3,136,712,978 | I_kwDOCUB6oc669nES | 38,753 | Silent Overwrite of Custom Optimizer When Using DeepSpeed with Transformers Trainer | {
"login": "VladPyzh",
"id": 61621694,
"node_id": "MDQ6VXNlcjYxNjIxNjk0",
"avatar_url": "https://avatars.githubusercontent.com/u/61621694?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/VladPyzh",
"html_url": "https://github.com/VladPyzh",
"followers_url": "https://api.github.com/users/VladPyzh/followers",
"following_url": "https://api.github.com/users/VladPyzh/following{/other_user}",
"gists_url": "https://api.github.com/users/VladPyzh/gists{/gist_id}",
"starred_url": "https://api.github.com/users/VladPyzh/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/VladPyzh/subscriptions",
"organizations_url": "https://api.github.com/users/VladPyzh/orgs",
"repos_url": "https://api.github.com/users/VladPyzh/repos",
"events_url": "https://api.github.com/users/VladPyzh/events{/privacy}",
"received_events_url": "https://api.github.com/users/VladPyzh/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-11T12:50:26 | 2025-07-21T08:03:43 | 2025-07-21T08:03:43 | NONE | null | null | null | null | ### System Info
- `transformers` version: 4.49.0
- Platform: Linux-5.4.0-216-generic-x86_64-with-glibc2.31
- Python version: 3.13.2
- Huggingface_hub version: 0.29.2
- Safetensors version: 0.5.3
- Accelerate version: 1.7.0
- Accelerate config: not found
- DeepSpeed version: 0.16.8
- PyTorch version (GPU?): 2.6.0+cu124 (True)
- Tensorflow version (GPU?): not installed (NA)
- Flax version (CPU?/GPU?/TPU?): not installed (NA)
- Jax version: not installed
- JaxLib version: not installed
- Using distributed or parallel set-up in script?: No
- Using GPU in script?: Yes
- GPU type: Tesla V100-PCIE-32GB
### Who can help?
@zach-huggingface @SunMarc
### 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
To reproduce the problem:
1. Pass some DeepSpeed config file to the TrainingArguments
2. Pass custom optimizer to the Trainer
Example:
```
class CustomOptim(torch.optim.Adam):
def step(self, *args, **kwargs):
print("steped!")
super().step(*args, **kwargs)
model_name = "facebook/opt-125m"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)
dataset = load_dataset("Salesforce/wikitext", "wikitext-103-raw-v1")
train_ds = dataset["train"]
eval_ds = dataset["validation"]
def tokenize(examples):
return tokenizer(
examples["text"], return_special_tokens_mask=True, truncation=True
)
train_ds = train_ds.map(tokenize, batched=True, remove_columns=["text"])
eval_ds = eval_ds.map(tokenize, batched=True, remove_columns=["text"])
data_collator = DataCollatorForLanguageModeling(tokenizer=tokenizer, mlm=False)
training_args = TrainingArguments(
output_dir="./opt-125m-finetuned",
do_eval=False,
overwrite_output_dir=True,
num_train_epochs=1,
per_device_train_batch_size=32,
per_device_eval_batch_size=32,
max_steps=1,
push_to_hub=False,
deepspeed="/data/vladislav/ds_config.json",
)
optimizer = CustomOptim(model.parameters())
scheduler = get_constant_schedule(optimizer)
trainer = Trainer(
model=model,
args=training_args,
optimizers=(optimizer, scheduler),
train_dataset=train_ds,
eval_dataset=eval_ds,
data_collator=data_collator,
)
trainer.train()
```
DeepSpeed config:
```
{
"train_batch_size": 32,
"gradient_accumulation_steps": 1,
"gradient_clipping": 1.0,
"zero_optimization": {
"stage": 0,
"contiguous_gradients": true,
"overlap_comm": true,
"reduce_bucket_size": 50000000,
"allgather_bucket_size": 500000000
}
}
```
In this example CustomOptimizer will be overwritten and no print would appear.
### Expected behavior
Hi HuggingFace team,
While working with the `transformers.Trainer` to train a LLM, I encountered an unexpected behavior when combining a custom optimizer with a DeepSpeed config in TrainingArguments.
It seems that when a custom optimizer is passed to the Trainer, but a DeepSpeed config is also provided, the custom optimizer is silently ignored and overwritten by DeepSpeed. This was quite surprising, and from what I can tell, this behavior is not mentioned in the documentation for either Trainer or TrainingArguments.
To avoid confusion and help users catch this early, it would be helpful if an explicit warning or exception were raised in such scenarios.
Thanks for the great work on the library! | {
"login": "github-actions[bot]",
"id": 41898282,
"node_id": "MDM6Qm90NDE4OTgyODI=",
"avatar_url": "https://avatars.githubusercontent.com/in/15368?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/github-actions%5Bbot%5D",
"html_url": "https://github.com/apps/github-actions",
"followers_url": "https://api.github.com/users/github-actions%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/github-actions%5Bbot%5D/following{/other_user}",
"gists_url": "https://api.github.com/users/github-actions%5Bbot%5D/gists{/gist_id}",
"starred_url": "https://api.github.com/users/github-actions%5Bbot%5D/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/github-actions%5Bbot%5D/subscriptions",
"organizations_url": "https://api.github.com/users/github-actions%5Bbot%5D/orgs",
"repos_url": "https://api.github.com/users/github-actions%5Bbot%5D/repos",
"events_url": "https://api.github.com/users/github-actions%5Bbot%5D/events{/privacy}",
"received_events_url": "https://api.github.com/users/github-actions%5Bbot%5D/received_events",
"type": "Bot",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38753/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/38753/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38752 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38752/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38752/comments | https://api.github.com/repos/huggingface/transformers/issues/38752/events | https://github.com/huggingface/transformers/pull/38752 | 3,136,677,135 | PR_kwDOCUB6oc6aBnL_ | 38,752 | Fix/gpu memory leak autoprocessor | {
"login": "Obssaa",
"id": 103986577,
"node_id": "U_kgDOBjK1kQ",
"avatar_url": "https://avatars.githubusercontent.com/u/103986577?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Obssaa",
"html_url": "https://github.com/Obssaa",
"followers_url": "https://api.github.com/users/Obssaa/followers",
"following_url": "https://api.github.com/users/Obssaa/following{/other_user}",
"gists_url": "https://api.github.com/users/Obssaa/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Obssaa/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Obssaa/subscriptions",
"organizations_url": "https://api.github.com/users/Obssaa/orgs",
"repos_url": "https://api.github.com/users/Obssaa/repos",
"events_url": "https://api.github.com/users/Obssaa/events{/privacy}",
"received_events_url": "https://api.github.com/users/Obssaa/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | open | false | null | [] | null | [] | 2025-06-11T12:37:57 | 2025-06-13T15:37:58 | null | NONE | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38752",
"html_url": "https://github.com/huggingface/transformers/pull/38752",
"diff_url": "https://github.com/huggingface/transformers/pull/38752.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38752.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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).
- [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request),
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?
## 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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38752/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/38752/timeline | null | null | null | null | true | false |
https://api.github.com/repos/huggingface/transformers/issues/38751 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38751/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38751/comments | https://api.github.com/repos/huggingface/transformers/issues/38751/events | https://github.com/huggingface/transformers/pull/38751 | 3,136,574,735 | PR_kwDOCUB6oc6aBQnI | 38,751 | 🚨 Don't use cache in non-generative models | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T12:02:39 | 2025-07-01T09:08:21 | 2025-07-01T09:08:21 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38751",
"html_url": "https://github.com/huggingface/transformers/pull/38751",
"diff_url": "https://github.com/huggingface/transformers/pull/38751.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38751.patch",
"merged_at": "2025-07-01T09:08:21"
} | # What does this PR do?
When working on https://github.com/huggingface/transformers/pull/38635, I found that there are some models which have `past_key_values` in their signature, even though they cannot generate. The reason is that models were all copying from Bert
This PR clean it up and changes the copy statement to Align model and adds support for new attention API in all those models | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38751/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/38751/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38750 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38750/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38750/comments | https://api.github.com/repos/huggingface/transformers/issues/38750/events | https://github.com/huggingface/transformers/issues/38750 | 3,136,416,174 | I_kwDOCUB6oc668emu | 38,750 | Is it a good choice to early error when `output_attentions=True` and attn implementation not equal to `eager` | {
"login": "kaixuanliu",
"id": 13268042,
"node_id": "MDQ6VXNlcjEzMjY4MDQy",
"avatar_url": "https://avatars.githubusercontent.com/u/13268042?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/kaixuanliu",
"html_url": "https://github.com/kaixuanliu",
"followers_url": "https://api.github.com/users/kaixuanliu/followers",
"following_url": "https://api.github.com/users/kaixuanliu/following{/other_user}",
"gists_url": "https://api.github.com/users/kaixuanliu/gists{/gist_id}",
"starred_url": "https://api.github.com/users/kaixuanliu/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/kaixuanliu/subscriptions",
"organizations_url": "https://api.github.com/users/kaixuanliu/orgs",
"repos_url": "https://api.github.com/users/kaixuanliu/repos",
"events_url": "https://api.github.com/users/kaixuanliu/events{/privacy}",
"received_events_url": "https://api.github.com/users/kaixuanliu/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-11T11:05:48 | 2025-06-25T08:00:06 | 2025-06-11T11:10:11 | CONTRIBUTOR | null | null | null | null | ### System Info
Before this PR [38288](https://github.com/huggingface/transformers/pull/38288), the program will run smoothly even when we set `output_attentions=True` and the attn implementation is not `eager`, as it will fallback to use eager mode, after this PR, it will throw error directly: [L342](https://github.com/huggingface/transformers/blob/main/src/transformers/configuration_utils.py#L342). I think it would be better if we just throw a warning and fallback to `eager` attn. Is it possible to revert it back or make small direct change based on this PR?
### Who can help?
@ArthurZucker
### 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
N/A
### Expected behavior
We want to make sure program can run without crash even we set `output_attentions=True` and attn implementation not equal to `eager` | {
"login": "kaixuanliu",
"id": 13268042,
"node_id": "MDQ6VXNlcjEzMjY4MDQy",
"avatar_url": "https://avatars.githubusercontent.com/u/13268042?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/kaixuanliu",
"html_url": "https://github.com/kaixuanliu",
"followers_url": "https://api.github.com/users/kaixuanliu/followers",
"following_url": "https://api.github.com/users/kaixuanliu/following{/other_user}",
"gists_url": "https://api.github.com/users/kaixuanliu/gists{/gist_id}",
"starred_url": "https://api.github.com/users/kaixuanliu/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/kaixuanliu/subscriptions",
"organizations_url": "https://api.github.com/users/kaixuanliu/orgs",
"repos_url": "https://api.github.com/users/kaixuanliu/repos",
"events_url": "https://api.github.com/users/kaixuanliu/events{/privacy}",
"received_events_url": "https://api.github.com/users/kaixuanliu/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38750/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/38750/timeline | null | not_planned | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38749 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38749/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38749/comments | https://api.github.com/repos/huggingface/transformers/issues/38749/events | https://github.com/huggingface/transformers/pull/38749 | 3,136,317,962 | PR_kwDOCUB6oc6aAYKI | 38,749 | Updated aya_vision.md | {
"login": "1himan",
"id": 140396762,
"node_id": "U_kgDOCF5I2g",
"avatar_url": "https://avatars.githubusercontent.com/u/140396762?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/1himan",
"html_url": "https://github.com/1himan",
"followers_url": "https://api.github.com/users/1himan/followers",
"following_url": "https://api.github.com/users/1himan/following{/other_user}",
"gists_url": "https://api.github.com/users/1himan/gists{/gist_id}",
"starred_url": "https://api.github.com/users/1himan/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/1himan/subscriptions",
"organizations_url": "https://api.github.com/users/1himan/orgs",
"repos_url": "https://api.github.com/users/1himan/repos",
"events_url": "https://api.github.com/users/1himan/events{/privacy}",
"received_events_url": "https://api.github.com/users/1himan/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T10:31:42 | 2025-06-17T05:46:11 | 2025-06-16T17:46:31 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38749",
"html_url": "https://github.com/huggingface/transformers/pull/38749",
"diff_url": "https://github.com/huggingface/transformers/pull/38749.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38749.patch",
"merged_at": "2025-06-16T17:46:31"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
As a part of this - https://github.com/huggingface/transformers/issues/36979
## 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#create-a-pull-request),
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?
@stevhliu
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38749/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/38749/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38748 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38748/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38748/comments | https://api.github.com/repos/huggingface/transformers/issues/38748/events | https://github.com/huggingface/transformers/pull/38748 | 3,136,292,113 | PR_kwDOCUB6oc6aASfC | 38,748 | Fix(informer): Correct tensor shape for input_size=1 | {
"login": "Flink-ddd",
"id": 180720690,
"node_id": "U_kgDOCsWUMg",
"avatar_url": "https://avatars.githubusercontent.com/u/180720690?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Flink-ddd",
"html_url": "https://github.com/Flink-ddd",
"followers_url": "https://api.github.com/users/Flink-ddd/followers",
"following_url": "https://api.github.com/users/Flink-ddd/following{/other_user}",
"gists_url": "https://api.github.com/users/Flink-ddd/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Flink-ddd/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Flink-ddd/subscriptions",
"organizations_url": "https://api.github.com/users/Flink-ddd/orgs",
"repos_url": "https://api.github.com/users/Flink-ddd/repos",
"events_url": "https://api.github.com/users/Flink-ddd/events{/privacy}",
"received_events_url": "https://api.github.com/users/Flink-ddd/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T10:22:27 | 2025-06-12T03:36:04 | 2025-06-12T03:36:04 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38748",
"html_url": "https://github.com/huggingface/transformers/pull/38748",
"diff_url": "https://github.com/huggingface/transformers/pull/38748.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38748.patch",
"merged_at": null
} | What does this PR do?
This PR fixes a RuntimeError that occurs in InformerModel when config.input_size is set to 1.
When input_size=1, the loc and scale tensors calculated by the scaler retained an extra dimension (e.g., shape [B, 1, 1] instead of [B, 1]). This incorrect shape caused a dimension mismatch error during the expand() operation later when creating expanded_static_feat.
This fix applies .squeeze(-1) to both the loc and scale tensors to ensure they have the correct dimensionality before being used. This resolves the crash and allows the model to run correctly with univariate time series data.
Fixes #38745 | {
"login": "Flink-ddd",
"id": 180720690,
"node_id": "U_kgDOCsWUMg",
"avatar_url": "https://avatars.githubusercontent.com/u/180720690?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Flink-ddd",
"html_url": "https://github.com/Flink-ddd",
"followers_url": "https://api.github.com/users/Flink-ddd/followers",
"following_url": "https://api.github.com/users/Flink-ddd/following{/other_user}",
"gists_url": "https://api.github.com/users/Flink-ddd/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Flink-ddd/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Flink-ddd/subscriptions",
"organizations_url": "https://api.github.com/users/Flink-ddd/orgs",
"repos_url": "https://api.github.com/users/Flink-ddd/repos",
"events_url": "https://api.github.com/users/Flink-ddd/events{/privacy}",
"received_events_url": "https://api.github.com/users/Flink-ddd/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38748/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/38748/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38747 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38747/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38747/comments | https://api.github.com/repos/huggingface/transformers/issues/38747/events | https://github.com/huggingface/transformers/pull/38747 | 3,136,143,517 | PR_kwDOCUB6oc6Z_yC0 | 38,747 | Add chatterbox model | {
"login": "eustlb",
"id": 94853470,
"node_id": "U_kgDOBadZXg",
"avatar_url": "https://avatars.githubusercontent.com/u/94853470?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/eustlb",
"html_url": "https://github.com/eustlb",
"followers_url": "https://api.github.com/users/eustlb/followers",
"following_url": "https://api.github.com/users/eustlb/following{/other_user}",
"gists_url": "https://api.github.com/users/eustlb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/eustlb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/eustlb/subscriptions",
"organizations_url": "https://api.github.com/users/eustlb/orgs",
"repos_url": "https://api.github.com/users/eustlb/repos",
"events_url": "https://api.github.com/users/eustlb/events{/privacy}",
"received_events_url": "https://api.github.com/users/eustlb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T09:31:41 | 2025-06-11T14:16:25 | 2025-06-11T14:15:16 | CONTRIBUTOR | null | null | true | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38747",
"html_url": "https://github.com/huggingface/transformers/pull/38747",
"diff_url": "https://github.com/huggingface/transformers/pull/38747.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38747.patch",
"merged_at": null
} | # What does this PR do?
Adds chatterbox text-to-speech model!
[Original code](https://github.com/resemble-ai/chatterbox)
| {
"login": "eustlb",
"id": 94853470,
"node_id": "U_kgDOBadZXg",
"avatar_url": "https://avatars.githubusercontent.com/u/94853470?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/eustlb",
"html_url": "https://github.com/eustlb",
"followers_url": "https://api.github.com/users/eustlb/followers",
"following_url": "https://api.github.com/users/eustlb/following{/other_user}",
"gists_url": "https://api.github.com/users/eustlb/gists{/gist_id}",
"starred_url": "https://api.github.com/users/eustlb/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/eustlb/subscriptions",
"organizations_url": "https://api.github.com/users/eustlb/orgs",
"repos_url": "https://api.github.com/users/eustlb/repos",
"events_url": "https://api.github.com/users/eustlb/events{/privacy}",
"received_events_url": "https://api.github.com/users/eustlb/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38747/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/38747/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38746 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38746/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38746/comments | https://api.github.com/repos/huggingface/transformers/issues/38746/events | https://github.com/huggingface/transformers/pull/38746 | 3,136,044,015 | PR_kwDOCUB6oc6Z_chX | 38,746 | Add V-JEPA 2 | {
"login": "qubvel",
"id": 31920396,
"node_id": "MDQ6VXNlcjMxOTIwMzk2",
"avatar_url": "https://avatars.githubusercontent.com/u/31920396?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/qubvel",
"html_url": "https://github.com/qubvel",
"followers_url": "https://api.github.com/users/qubvel/followers",
"following_url": "https://api.github.com/users/qubvel/following{/other_user}",
"gists_url": "https://api.github.com/users/qubvel/gists{/gist_id}",
"starred_url": "https://api.github.com/users/qubvel/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/qubvel/subscriptions",
"organizations_url": "https://api.github.com/users/qubvel/orgs",
"repos_url": "https://api.github.com/users/qubvel/repos",
"events_url": "https://api.github.com/users/qubvel/events{/privacy}",
"received_events_url": "https://api.github.com/users/qubvel/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T08:59:56 | 2025-06-11T14:00:08 | 2025-06-11T14:00:08 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38746",
"html_url": "https://github.com/huggingface/transformers/pull/38746",
"diff_url": "https://github.com/huggingface/transformers/pull/38746.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38746.patch",
"merged_at": "2025-06-11T14:00:08"
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "qubvel",
"id": 31920396,
"node_id": "MDQ6VXNlcjMxOTIwMzk2",
"avatar_url": "https://avatars.githubusercontent.com/u/31920396?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/qubvel",
"html_url": "https://github.com/qubvel",
"followers_url": "https://api.github.com/users/qubvel/followers",
"following_url": "https://api.github.com/users/qubvel/following{/other_user}",
"gists_url": "https://api.github.com/users/qubvel/gists{/gist_id}",
"starred_url": "https://api.github.com/users/qubvel/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/qubvel/subscriptions",
"organizations_url": "https://api.github.com/users/qubvel/orgs",
"repos_url": "https://api.github.com/users/qubvel/repos",
"events_url": "https://api.github.com/users/qubvel/events{/privacy}",
"received_events_url": "https://api.github.com/users/qubvel/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38746/reactions",
"total_count": 1,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 1,
"eyes": 0
} | https://api.github.com/repos/huggingface/transformers/issues/38746/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38745 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38745/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38745/comments | https://api.github.com/repos/huggingface/transformers/issues/38745/events | https://github.com/huggingface/transformers/issues/38745 | 3,135,770,652 | I_kwDOCUB6oc666BAc | 38,745 | [Bug][InformerForPredict] The shape will cause a problem | {
"login": "2004learner",
"id": 117647339,
"node_id": "U_kgDOBwMn6w",
"avatar_url": "https://avatars.githubusercontent.com/u/117647339?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/2004learner",
"html_url": "https://github.com/2004learner",
"followers_url": "https://api.github.com/users/2004learner/followers",
"following_url": "https://api.github.com/users/2004learner/following{/other_user}",
"gists_url": "https://api.github.com/users/2004learner/gists{/gist_id}",
"starred_url": "https://api.github.com/users/2004learner/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/2004learner/subscriptions",
"organizations_url": "https://api.github.com/users/2004learner/orgs",
"repos_url": "https://api.github.com/users/2004learner/repos",
"events_url": "https://api.github.com/users/2004learner/events{/privacy}",
"received_events_url": "https://api.github.com/users/2004learner/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-11T07:22:06 | 2025-07-20T11:41:45 | 2025-06-23T09:50:53 | NONE | null | null | null | null | ### System Info
When I set the infomerconfig.input_size = 1, I find a bug, but I don't know how to fix it.
- Function Name : `create_network_inputs`
```
time_feat = (
torch.cat(
(
past_time_features[:, self._past_length - self.config.context_length :, ...],
future_time_features,
),
dim=1,
)
if future_values is not None
else past_time_features[:, self._past_length - self.config.context_length :, ...]
)
print(self._past_length)
# target
if past_observed_mask is None:
past_observed_mask = torch.ones_like(past_values)
context = past_values[:, -self.config.context_length :]
observed_context = past_observed_mask[:, -self.config.context_length :]
_, loc, scale = self.scaler(context, observed_context)
inputs = (
(torch.cat((past_values, future_values), dim=1) - loc) / scale
if future_values is not None
else (past_values - loc) / scale
)
print(loc.shape, scale.shape, inputs.shape)
# static features
log_abs_loc = loc.abs().log1p() if self.config.input_size == 1 else loc.squeeze(1).abs().log1p()
log_scale = scale.log() if self.config.input_size == 1 else scale.squeeze(1).log()
print(f"log_abs_loc: {log_abs_loc.shape}, {log_scale.shape}")
print(time_feat.shape, self.config.input_size)
static_feat = torch.cat((log_abs_loc, log_scale), dim=1)
print(time_feat.shape, static_feat.shape)
if static_real_features is not None:
static_feat = torch.cat((static_real_features, static_feat), dim=1)
if static_categorical_features is not None:
embedded_cat = self.embedder(static_categorical_features)
static_feat = torch.cat((embedded_cat, static_feat), dim=1)
print(time_feat.shape, static_feat.shape)
expanded_static_feat = static_feat.unsqueeze(1).expand(-1, time_feat.shape[1], -1)
# all features
features = torch.cat((expanded_static_feat, time_feat), dim=-1)
# lagged features
subsequences_length = (
self.config.context_length + self.config.prediction_length
if future_values is not None
else self.config.context_length
)
lagged_sequence = self.get_lagged_subsequences(sequence=inputs, subsequences_length=subsequences_length)
lags_shape = lagged_sequence.shape
reshaped_lagged_sequence = lagged_sequence.reshape(lags_shape[0], lags_shape[1], -1)
if reshaped_lagged_sequence.shape[1] != time_feat.shape[1]:
raise ValueError(
f"input length {reshaped_lagged_sequence.shape[1]} and time feature lengths {time_feat.shape[1]} does not match"
)
# transformer inputs
transformer_inputs = torch.cat((reshaped_lagged_sequence, features), dim=-1)
return transformer_inputs, loc, scale, static_feat
```
As we can see, I add some `print` sentence in the library to see the shape, now the bug is:
```
Traceback (most recent call last):
File "/home/wjt/luck/FinalWork/alert_models/informer_based_model_3_cpu.py", line 820, in <module>
pipline.train_model()
File "/home/wjt/luck/FinalWork/alert_models/informer_based_model_3_cpu.py", line 466, in train_model
outputs = model(
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
return forward_call(*args, **kwargs)
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/transformers/models/informer/modeling_informer.py", line 1844, in forward
outputs = self.model(
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
return forward_call(*args, **kwargs)
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/transformers/models/informer/modeling_informer.py", line 1568, in forward
transformer_inputs, loc, scale, static_feat = self.create_network_inputs(
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/transformers/models/informer/modeling_informer.py", line 1386, in create_network_inputs
expanded_static_feat = static_feat.unsqueeze(1).expand(-1, time_feat.shape[1], -1)
RuntimeError: expand(torch.cuda.FloatTensor{[32, 1, 2, 1]}, size=[-1, 27, -1]): the number of sizes provided (3) must be greater or equal to the number of dimensions in the tensor (4)
```
- First
```
log_abs_loc = loc.abs().log1p() if self.config.input_size == 1 else loc.squeeze(1).abs().log1p() log_scale = scale.log() if self.config.input_size == 1 else scale.squeeze(1).log()
```
If the input_size is 1, `loc.squeeze` will not execute, but now the loc.shape is `(bsz, 1, 1)`, so it will be a hidden issue, and
- second
```
static_feat.unsqueeze(1) will execute, then the error will be reported
```
I don't know whether it's clear to clarify the problem, but I really need a help.
- environment
```
Name: transformers
Version: 4.52.3
Summary: State-of-the-art Machine Learning for JAX, PyTorch and TensorFlow
Home-page: https://github.com/huggingface/transformers
Author: The Hugging Face team (past and future) with the help of all our contributors (https://github.com/huggingface/transformers/graphs/contributors)
Author-email: transformers@huggingface.co
License: Apache 2.0 License
Location: /home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages
Requires: filelock, huggingface-hub, numpy, pack
Linux
python3.9.21
```
### 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
1. set the InformerConfig.size = 1
2. maybe do not include any other static features
3. maybe InformerConfig.scale = None
### Expected behavior
Traceback (most recent call last):
File "/home/wjt/luck/FinalWork/alert_models/informer_based_model_3_cpu.py", line 820, in <module>
pipline.train_model()
File "/home/wjt/luck/FinalWork/alert_models/informer_based_model_3_cpu.py", line 466, in train_model
outputs = model(
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
return forward_call(*args, **kwargs)
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/transformers/models/informer/modeling_informer.py", line 1844, in forward
outputs = self.model(
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
return forward_call(*args, **kwargs)
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/transformers/models/informer/modeling_informer.py", line 1568, in forward
transformer_inputs, loc, scale, static_feat = self.create_network_inputs(
File "/home/wjt/.conda/envs/luckluck/lib/python3.9/site-packages/transformers/models/informer/modeling_informer.py", line 1386, in create_network_inputs
expanded_static_feat = static_feat.unsqueeze(1).expand(-1, time_feat.shape[1], -1)
RuntimeError: expand(torch.cuda.FloatTensor{[32, 1, 2, 1]}, size=[-1, 27, -1]): the number of sizes provided (3) must be greater or equal to the number of dimensions in the tensor (4) | {
"login": "kashif",
"id": 8100,
"node_id": "MDQ6VXNlcjgxMDA=",
"avatar_url": "https://avatars.githubusercontent.com/u/8100?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/kashif",
"html_url": "https://github.com/kashif",
"followers_url": "https://api.github.com/users/kashif/followers",
"following_url": "https://api.github.com/users/kashif/following{/other_user}",
"gists_url": "https://api.github.com/users/kashif/gists{/gist_id}",
"starred_url": "https://api.github.com/users/kashif/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/kashif/subscriptions",
"organizations_url": "https://api.github.com/users/kashif/orgs",
"repos_url": "https://api.github.com/users/kashif/repos",
"events_url": "https://api.github.com/users/kashif/events{/privacy}",
"received_events_url": "https://api.github.com/users/kashif/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38745/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/38745/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38744 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38744/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38744/comments | https://api.github.com/repos/huggingface/transformers/issues/38744/events | https://github.com/huggingface/transformers/pull/38744 | 3,135,648,985 | PR_kwDOCUB6oc6Z-Hmy | 38,744 | chore: support passing kwargs when using gradient checkpointing | {
"login": "starcatmeow",
"id": 19618963,
"node_id": "MDQ6VXNlcjE5NjE4OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/19618963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/starcatmeow",
"html_url": "https://github.com/starcatmeow",
"followers_url": "https://api.github.com/users/starcatmeow/followers",
"following_url": "https://api.github.com/users/starcatmeow/following{/other_user}",
"gists_url": "https://api.github.com/users/starcatmeow/gists{/gist_id}",
"starred_url": "https://api.github.com/users/starcatmeow/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/starcatmeow/subscriptions",
"organizations_url": "https://api.github.com/users/starcatmeow/orgs",
"repos_url": "https://api.github.com/users/starcatmeow/repos",
"events_url": "https://api.github.com/users/starcatmeow/events{/privacy}",
"received_events_url": "https://api.github.com/users/starcatmeow/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T06:30:25 | 2025-06-11T06:47:37 | 2025-06-11T06:47:37 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38744",
"html_url": "https://github.com/huggingface/transformers/pull/38744",
"diff_url": "https://github.com/huggingface/transformers/pull/38744.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38744.patch",
"merged_at": null
} | # What does this PR do?
Similar to the changes made in https://github.com/huggingface/transformers/pull/37037, this commit allows passing kwargs to layers when using gradient checkpointing for more models.
For example, Pixtral needs to [pass](https://github.com/huggingface/transformers/blob/main/src/transformers/models/pixtral/modeling_pixtral.py#L214-L217) `position_ids` to [_flash_attention_forward](https://github.com/huggingface/transformers/blob/main/src/transformers/modeling_flash_attention_utils.py#L292), without this change we cannot train Pixtral with gradient checkpointing & flash attention.
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
Related PR https://github.com/huggingface/transformers/pull/37037, issue https://github.com/huggingface/transformers/issues/35509.
## Before submitting
- [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request),
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.
## Who can review?
- @efsotr
- @ArthurZucker
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "starcatmeow",
"id": 19618963,
"node_id": "MDQ6VXNlcjE5NjE4OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/19618963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/starcatmeow",
"html_url": "https://github.com/starcatmeow",
"followers_url": "https://api.github.com/users/starcatmeow/followers",
"following_url": "https://api.github.com/users/starcatmeow/following{/other_user}",
"gists_url": "https://api.github.com/users/starcatmeow/gists{/gist_id}",
"starred_url": "https://api.github.com/users/starcatmeow/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/starcatmeow/subscriptions",
"organizations_url": "https://api.github.com/users/starcatmeow/orgs",
"repos_url": "https://api.github.com/users/starcatmeow/repos",
"events_url": "https://api.github.com/users/starcatmeow/events{/privacy}",
"received_events_url": "https://api.github.com/users/starcatmeow/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38744/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/38744/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38743 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38743/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38743/comments | https://api.github.com/repos/huggingface/transformers/issues/38743/events | https://github.com/huggingface/transformers/pull/38743 | 3,135,645,264 | PR_kwDOCUB6oc6Z-Gze | 38,743 | [DeepSeek-V3] implement when q_lora_rank is None | {
"login": "bzantium",
"id": 19511788,
"node_id": "MDQ6VXNlcjE5NTExNzg4",
"avatar_url": "https://avatars.githubusercontent.com/u/19511788?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/bzantium",
"html_url": "https://github.com/bzantium",
"followers_url": "https://api.github.com/users/bzantium/followers",
"following_url": "https://api.github.com/users/bzantium/following{/other_user}",
"gists_url": "https://api.github.com/users/bzantium/gists{/gist_id}",
"starred_url": "https://api.github.com/users/bzantium/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/bzantium/subscriptions",
"organizations_url": "https://api.github.com/users/bzantium/orgs",
"repos_url": "https://api.github.com/users/bzantium/repos",
"events_url": "https://api.github.com/users/bzantium/events{/privacy}",
"received_events_url": "https://api.github.com/users/bzantium/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T06:28:38 | 2025-06-11T12:43:14 | 2025-06-11T12:35:10 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38743",
"html_url": "https://github.com/huggingface/transformers/pull/38743",
"diff_url": "https://github.com/huggingface/transformers/pull/38743.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38743.patch",
"merged_at": "2025-06-11T12:35:10"
} | # What does this PR do?
Implement handling for configurations where the `q_lora_rank` parameter is set to `None`.
Fixes #38742
## 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#create-a-pull-request),
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?
@ArthurZucker | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38743/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/38743/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38742 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38742/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38742/comments | https://api.github.com/repos/huggingface/transformers/issues/38742/events | https://github.com/huggingface/transformers/issues/38742 | 3,135,614,001 | I_kwDOCUB6oc665awx | 38,742 | [DeepSeek-V3] take care of the case `q_lora_rank is None` | {
"login": "bzantium",
"id": 19511788,
"node_id": "MDQ6VXNlcjE5NTExNzg4",
"avatar_url": "https://avatars.githubusercontent.com/u/19511788?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/bzantium",
"html_url": "https://github.com/bzantium",
"followers_url": "https://api.github.com/users/bzantium/followers",
"following_url": "https://api.github.com/users/bzantium/following{/other_user}",
"gists_url": "https://api.github.com/users/bzantium/gists{/gist_id}",
"starred_url": "https://api.github.com/users/bzantium/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/bzantium/subscriptions",
"organizations_url": "https://api.github.com/users/bzantium/orgs",
"repos_url": "https://api.github.com/users/bzantium/repos",
"events_url": "https://api.github.com/users/bzantium/events{/privacy}",
"received_events_url": "https://api.github.com/users/bzantium/received_events",
"type": "User",
"user_view_type": "public",
"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 | [] | null | [] | 2025-06-11T06:13:51 | 2025-06-11T12:35:11 | 2025-06-11T12:35:11 | CONTRIBUTOR | null | null | null | null | ### Feature request
Implement handling for configurations where the `q_lora_rank` parameter is set to `None`.
### Motivation
1. [DeepSeek-V2-Lite](https://huggingface.co/deepseek-ai/DeepSeek-V2-Lite/blob/main/config.json) model has `q_lora_rank=None` so we can support this model with this modeling.
2. [Original remote modeling code](https://huggingface.co/deepseek-ai/DeepSeek-V3) also supports this feature.
### Your contribution
Implement handling for configurations where the `q_lora_rank` parameter is set to `None`. | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38742/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/38742/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38741 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38741/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38741/comments | https://api.github.com/repos/huggingface/transformers/issues/38741/events | https://github.com/huggingface/transformers/pull/38741 | 3,135,533,179 | PR_kwDOCUB6oc6Z9uUG | 38,741 | chore(pixtral): emit block attention mask when using flash attention | {
"login": "starcatmeow",
"id": 19618963,
"node_id": "MDQ6VXNlcjE5NjE4OTYz",
"avatar_url": "https://avatars.githubusercontent.com/u/19618963?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/starcatmeow",
"html_url": "https://github.com/starcatmeow",
"followers_url": "https://api.github.com/users/starcatmeow/followers",
"following_url": "https://api.github.com/users/starcatmeow/following{/other_user}",
"gists_url": "https://api.github.com/users/starcatmeow/gists{/gist_id}",
"starred_url": "https://api.github.com/users/starcatmeow/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/starcatmeow/subscriptions",
"organizations_url": "https://api.github.com/users/starcatmeow/orgs",
"repos_url": "https://api.github.com/users/starcatmeow/repos",
"events_url": "https://api.github.com/users/starcatmeow/events{/privacy}",
"received_events_url": "https://api.github.com/users/starcatmeow/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-11T05:28:26 | 2025-06-11T18:55:57 | 2025-06-11T18:55:24 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38741",
"html_url": "https://github.com/huggingface/transformers/pull/38741",
"diff_url": "https://github.com/huggingface/transformers/pull/38741.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38741.patch",
"merged_at": "2025-06-11T18:55:24"
} | # What does this PR do?
Since flash_attention_2 relies solely on position_ids, emitting the block attention mask avoids unnecessary memory usage and prevents OOM on large inputs.
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
## Before submitting
- [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request),
Pull Request section?
## Who can review?
- @uminaty
- @amyeroberts
- @qubvel
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38741/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/38741/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38740 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38740/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38740/comments | https://api.github.com/repos/huggingface/transformers/issues/38740/events | https://github.com/huggingface/transformers/issues/38740 | 3,135,466,321 | I_kwDOCUB6oc6642tR | 38,740 | [DOCS] Add `pruna` as optimization framework | {
"login": "davidberenstein1957",
"id": 25269220,
"node_id": "MDQ6VXNlcjI1MjY5MjIw",
"avatar_url": "https://avatars.githubusercontent.com/u/25269220?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/davidberenstein1957",
"html_url": "https://github.com/davidberenstein1957",
"followers_url": "https://api.github.com/users/davidberenstein1957/followers",
"following_url": "https://api.github.com/users/davidberenstein1957/following{/other_user}",
"gists_url": "https://api.github.com/users/davidberenstein1957/gists{/gist_id}",
"starred_url": "https://api.github.com/users/davidberenstein1957/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/davidberenstein1957/subscriptions",
"organizations_url": "https://api.github.com/users/davidberenstein1957/orgs",
"repos_url": "https://api.github.com/users/davidberenstein1957/repos",
"events_url": "https://api.github.com/users/davidberenstein1957/events{/privacy}",
"received_events_url": "https://api.github.com/users/davidberenstein1957/received_events",
"type": "User",
"user_view_type": "public",
"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 | [] | null | [] | 2025-06-11T04:52:33 | 2025-07-16T08:56:52 | null | NONE | null | null | null | null | ### Feature request
Have a section on Pruna AI within the documentation. We did [a similar PR for diffusers](https://github.com/huggingface/diffusers/pull/11688) and thought it would be nice to show how to optimize transformers models too.
.
### Motivation
Have a section on Pruna AI within the documentation to show how to optimize LLMs for inference.
### Your contribution
We could do everything for the PR. | null | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38740/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/38740/timeline | null | null | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | false |
https://api.github.com/repos/huggingface/transformers/issues/38739 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38739/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38739/comments | https://api.github.com/repos/huggingface/transformers/issues/38739/events | https://github.com/huggingface/transformers/issues/38739 | 3,134,486,533 | I_kwDOCUB6oc661HgF | 38,739 | LlamaAttention forward function type hint is incorrect | {
"login": "nhatkhtn",
"id": 61368343,
"node_id": "MDQ6VXNlcjYxMzY4MzQz",
"avatar_url": "https://avatars.githubusercontent.com/u/61368343?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nhatkhtn",
"html_url": "https://github.com/nhatkhtn",
"followers_url": "https://api.github.com/users/nhatkhtn/followers",
"following_url": "https://api.github.com/users/nhatkhtn/following{/other_user}",
"gists_url": "https://api.github.com/users/nhatkhtn/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nhatkhtn/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nhatkhtn/subscriptions",
"organizations_url": "https://api.github.com/users/nhatkhtn/orgs",
"repos_url": "https://api.github.com/users/nhatkhtn/repos",
"events_url": "https://api.github.com/users/nhatkhtn/events{/privacy}",
"received_events_url": "https://api.github.com/users/nhatkhtn/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [
{
"id": 3817266200,
"node_id": "MDU6TGFiZWwzODE3MjY2MjAw",
"url": "https://api.github.com/repos/huggingface/transformers/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": null
}
] | closed | false | null | [] | null | [] | 2025-06-10T18:17:48 | 2025-07-16T09:52:24 | 2025-07-15T20:02:41 | NONE | null | null | null | null | ### System Info
For the current version (4.52.4), in the `LlamaAttention` class, the type hint for the forward function
https://github.com/huggingface/transformers/blob/aa798b7ac9ff5018b3578eb927dc438671ab6a3e/src/transformers/models/llama/modeling_llama.py#L231
and what it actually returns
https://github.com/huggingface/transformers/blob/aa798b7ac9ff5018b3578eb927dc438671ab6a3e/src/transformers/models/llama/modeling_llama.py#L264
mismatch. Looking at the git blame, this is probably the result of the attention refactor introduced in #35235.
### Who can help?
@ArthurZucker
### 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
Look at the code
### Expected behavior
The type hint should be correct to avoid confusion | {
"login": "nhatkhtn",
"id": 61368343,
"node_id": "MDQ6VXNlcjYxMzY4MzQz",
"avatar_url": "https://avatars.githubusercontent.com/u/61368343?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/nhatkhtn",
"html_url": "https://github.com/nhatkhtn",
"followers_url": "https://api.github.com/users/nhatkhtn/followers",
"following_url": "https://api.github.com/users/nhatkhtn/following{/other_user}",
"gists_url": "https://api.github.com/users/nhatkhtn/gists{/gist_id}",
"starred_url": "https://api.github.com/users/nhatkhtn/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/nhatkhtn/subscriptions",
"organizations_url": "https://api.github.com/users/nhatkhtn/orgs",
"repos_url": "https://api.github.com/users/nhatkhtn/repos",
"events_url": "https://api.github.com/users/nhatkhtn/events{/privacy}",
"received_events_url": "https://api.github.com/users/nhatkhtn/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38739/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/38739/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
https://api.github.com/repos/huggingface/transformers/issues/38738 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38738/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38738/comments | https://api.github.com/repos/huggingface/transformers/issues/38738/events | https://github.com/huggingface/transformers/pull/38738 | 3,134,406,448 | PR_kwDOCUB6oc6Z56xl | 38,738 | Disable custom MRA kernels for ROCm | {
"login": "ahadnagy",
"id": 21314428,
"node_id": "MDQ6VXNlcjIxMzE0NDI4",
"avatar_url": "https://avatars.githubusercontent.com/u/21314428?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ahadnagy",
"html_url": "https://github.com/ahadnagy",
"followers_url": "https://api.github.com/users/ahadnagy/followers",
"following_url": "https://api.github.com/users/ahadnagy/following{/other_user}",
"gists_url": "https://api.github.com/users/ahadnagy/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ahadnagy/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ahadnagy/subscriptions",
"organizations_url": "https://api.github.com/users/ahadnagy/orgs",
"repos_url": "https://api.github.com/users/ahadnagy/repos",
"events_url": "https://api.github.com/users/ahadnagy/events{/privacy}",
"received_events_url": "https://api.github.com/users/ahadnagy/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-10T17:46:01 | 2025-06-13T10:25:28 | 2025-06-13T10:25:28 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38738",
"html_url": "https://github.com/huggingface/transformers/pull/38738",
"diff_url": "https://github.com/huggingface/transformers/pull/38738.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38738.patch",
"merged_at": "2025-06-13T10:25:28"
} | # What does this PR do?
This PR is a clone of https://github.com/huggingface/transformers/pull/38578 with a small fix, this time targeting main. It's part of a series of PRs propagating back fixes for AMD devices from https://github.com/huggingface/transformers/tree/amd-hf-ci-branch .
TLDR; it disables the MRA CUDA extensions for AMD devices as the kernels are broken on the platform at the moment.
## 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#create-a-pull-request),
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.
<!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @
If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**.
Please tag fewer than 3 people.
Models:
- text models: @ArthurZucker
- vision models: @amyeroberts, @qubvel
- speech models: @eustlb
- graph models: @clefourrier
Library:
- flax: @gante and @Rocketknight1
- generate: @zucchini-nlp (visual-language models) or @gante (all others)
- pipelines: @Rocketknight1
- tensorflow: @gante and @Rocketknight1
- tokenizers: @ArthurZucker
- trainer: @zach-huggingface and @SunMarc
- chat templates: @Rocketknight1
Integrations:
- deepspeed: HF Trainer/Accelerate: @SunMarc @zach-huggingface
- ray/raytune: @richardliaw, @amogkam
- Big Model Inference: @SunMarc
- quantization (bitsandbytes, autogpt): @SunMarc @MekkCyber
Documentation: @stevhliu
HF projects:
- accelerate: [different repo](https://github.com/huggingface/accelerate)
- datasets: [different repo](https://github.com/huggingface/datasets)
- diffusers: [different repo](https://github.com/huggingface/diffusers)
- rust tokenizers: [different repo](https://github.com/huggingface/tokenizers)
Maintained examples (not research project or legacy):
- Flax: @Rocketknight1
- PyTorch: See Models above and tag the person corresponding to the modality of the example.
- TensorFlow: @Rocketknight1
-->
| {
"login": "ivarflakstad",
"id": 69173633,
"node_id": "MDQ6VXNlcjY5MTczNjMz",
"avatar_url": "https://avatars.githubusercontent.com/u/69173633?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/ivarflakstad",
"html_url": "https://github.com/ivarflakstad",
"followers_url": "https://api.github.com/users/ivarflakstad/followers",
"following_url": "https://api.github.com/users/ivarflakstad/following{/other_user}",
"gists_url": "https://api.github.com/users/ivarflakstad/gists{/gist_id}",
"starred_url": "https://api.github.com/users/ivarflakstad/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/ivarflakstad/subscriptions",
"organizations_url": "https://api.github.com/users/ivarflakstad/orgs",
"repos_url": "https://api.github.com/users/ivarflakstad/repos",
"events_url": "https://api.github.com/users/ivarflakstad/events{/privacy}",
"received_events_url": "https://api.github.com/users/ivarflakstad/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38738/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/38738/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38737 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38737/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38737/comments | https://api.github.com/repos/huggingface/transformers/issues/38737/events | https://github.com/huggingface/transformers/pull/38737 | 3,134,326,065 | PR_kwDOCUB6oc6Z5otG | 38,737 | Fixing the Problem: `align_to_words=True` in `QuestionAnsweringPipeline` can lead to duplicate answers | {
"login": "Vixel2006",
"id": 166058059,
"node_id": "U_kgDOCeXYSw",
"avatar_url": "https://avatars.githubusercontent.com/u/166058059?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Vixel2006",
"html_url": "https://github.com/Vixel2006",
"followers_url": "https://api.github.com/users/Vixel2006/followers",
"following_url": "https://api.github.com/users/Vixel2006/following{/other_user}",
"gists_url": "https://api.github.com/users/Vixel2006/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Vixel2006/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Vixel2006/subscriptions",
"organizations_url": "https://api.github.com/users/Vixel2006/orgs",
"repos_url": "https://api.github.com/users/Vixel2006/repos",
"events_url": "https://api.github.com/users/Vixel2006/events{/privacy}",
"received_events_url": "https://api.github.com/users/Vixel2006/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-10T17:19:07 | 2025-06-11T14:57:26 | 2025-06-11T14:57:26 | CONTRIBUTOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38737",
"html_url": "https://github.com/huggingface/transformers/pull/38737",
"diff_url": "https://github.com/huggingface/transformers/pull/38737.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38737.patch",
"merged_at": null
} | # What does this PR do?
<!--
Congratulations! You've made it this far! You're not quite done yet though.
Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons---sometimes notifications get lost.
-->
<!-- Remove if not applicable -->
Fixes #26286
Solves the duplicate answers problem when `align_to_words=True` which is default in the `question_answering` pipeline by adding `handle_duplicate_answers=True` by default which handles duplicate answer by keeping track of the answers history and only adding the answer if the `start` and `end` pair not already in the output.
## Who can review?
[@Rocketknight1](https://github.com/Rocketknight1) [@ArthurZucker ](https://github.com/ArthurZucker)
| {
"login": "Vixel2006",
"id": 166058059,
"node_id": "U_kgDOCeXYSw",
"avatar_url": "https://avatars.githubusercontent.com/u/166058059?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/Vixel2006",
"html_url": "https://github.com/Vixel2006",
"followers_url": "https://api.github.com/users/Vixel2006/followers",
"following_url": "https://api.github.com/users/Vixel2006/following{/other_user}",
"gists_url": "https://api.github.com/users/Vixel2006/gists{/gist_id}",
"starred_url": "https://api.github.com/users/Vixel2006/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/Vixel2006/subscriptions",
"organizations_url": "https://api.github.com/users/Vixel2006/orgs",
"repos_url": "https://api.github.com/users/Vixel2006/repos",
"events_url": "https://api.github.com/users/Vixel2006/events{/privacy}",
"received_events_url": "https://api.github.com/users/Vixel2006/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38737/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/38737/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38736 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38736/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38736/comments | https://api.github.com/repos/huggingface/transformers/issues/38736/events | https://github.com/huggingface/transformers/pull/38736 | 3,134,197,293 | PR_kwDOCUB6oc6Z5M2J | 38,736 | Fixed a multiple-devices issue in SmolVLM model | {
"login": "remi-or",
"id": 83456801,
"node_id": "MDQ6VXNlcjgzNDU2ODAx",
"avatar_url": "https://avatars.githubusercontent.com/u/83456801?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/remi-or",
"html_url": "https://github.com/remi-or",
"followers_url": "https://api.github.com/users/remi-or/followers",
"following_url": "https://api.github.com/users/remi-or/following{/other_user}",
"gists_url": "https://api.github.com/users/remi-or/gists{/gist_id}",
"starred_url": "https://api.github.com/users/remi-or/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/remi-or/subscriptions",
"organizations_url": "https://api.github.com/users/remi-or/orgs",
"repos_url": "https://api.github.com/users/remi-or/repos",
"events_url": "https://api.github.com/users/remi-or/events{/privacy}",
"received_events_url": "https://api.github.com/users/remi-or/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-10T16:22:10 | 2025-06-11T08:08:01 | 2025-06-11T08:08:01 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38736",
"html_url": "https://github.com/huggingface/transformers/pull/38736",
"diff_url": "https://github.com/huggingface/transformers/pull/38736.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38736.patch",
"merged_at": "2025-06-11T08:08:01"
} | This PR fixes a multiple-devices issue in SmolVLM model. It was originally fixed in the `amd-hf-ci-branch` but it would be nice to have it in `main`.
Without this PR, the test `tests/models/smolvlm/test_modeling_smolvlm.py::SmolVLMForConditionalGenerationIntegrationTest::test_integration_test` failed on MI3XX machines.
cc. @zucchini-nlp (I think) | {
"login": "zucchini-nlp",
"id": 100715397,
"node_id": "U_kgDOBgDLhQ",
"avatar_url": "https://avatars.githubusercontent.com/u/100715397?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/zucchini-nlp",
"html_url": "https://github.com/zucchini-nlp",
"followers_url": "https://api.github.com/users/zucchini-nlp/followers",
"following_url": "https://api.github.com/users/zucchini-nlp/following{/other_user}",
"gists_url": "https://api.github.com/users/zucchini-nlp/gists{/gist_id}",
"starred_url": "https://api.github.com/users/zucchini-nlp/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/zucchini-nlp/subscriptions",
"organizations_url": "https://api.github.com/users/zucchini-nlp/orgs",
"repos_url": "https://api.github.com/users/zucchini-nlp/repos",
"events_url": "https://api.github.com/users/zucchini-nlp/events{/privacy}",
"received_events_url": "https://api.github.com/users/zucchini-nlp/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38736/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/38736/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38735 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38735/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38735/comments | https://api.github.com/repos/huggingface/transformers/issues/38735/events | https://github.com/huggingface/transformers/pull/38735 | 3,134,066,844 | PR_kwDOCUB6oc6Z4wGJ | 38,735 | Revert "Trigger doc-builder job after style bot" | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-10T15:36:11 | 2025-06-11T12:56:41 | 2025-06-11T12:56:39 | COLLABORATOR | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38735",
"html_url": "https://github.com/huggingface/transformers/pull/38735",
"diff_url": "https://github.com/huggingface/transformers/pull/38735.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38735.patch",
"merged_at": "2025-06-11T12:56:39"
} | Reverts huggingface/transformers#38398
Unfortunately, this is not working.
That PR use `workflow_call` event to call the `PR documentation build` inside the triggered `Style bot` workflow. This would be identified as a different job than `PR documentation build` itself (which is a required check), and PRs could not be merged anyway.
I talk to infra team with using a PAT token, but I don't get a response yet.
This PR is also causing some issues: for the last commit on a PR, if there is any other new PR merged into `main`, and we approve the workflow run for the PR, it will fail to fetch the corresponding commits.
Like https://github.com/huggingface/transformers/actions/runs/15487007776/job/43623404484
| {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38735/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/38735/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38734 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38734/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38734/comments | https://api.github.com/repos/huggingface/transformers/issues/38734/events | https://github.com/huggingface/transformers/pull/38734 | 3,134,045,500 | PR_kwDOCUB6oc6Z4rYL | 38,734 | Add AGENTS.md | {
"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",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-10T15:28:51 | 2025-06-10T16:27:38 | 2025-06-10T16:27:37 | MEMBER | null | null | false | {
"url": "https://api.github.com/repos/huggingface/transformers/pulls/38734",
"html_url": "https://github.com/huggingface/transformers/pull/38734",
"diff_url": "https://github.com/huggingface/transformers/pull/38734.diff",
"patch_url": "https://github.com/huggingface/transformers/pull/38734.patch",
"merged_at": "2025-06-10T16:27:37"
} | This PR adds an `AGENTS.md` file, which contains instructions for code agents interacting with the codebase. We'll keep an eye on code agent PRs and probably iterate on this, depending on which mistakes we see them making! | {
"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",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38734/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/38734/timeline | null | null | null | null | true | true |
https://api.github.com/repos/huggingface/transformers/issues/38733 | https://api.github.com/repos/huggingface/transformers | https://api.github.com/repos/huggingface/transformers/issues/38733/labels{/name} | https://api.github.com/repos/huggingface/transformers/issues/38733/comments | https://api.github.com/repos/huggingface/transformers/issues/38733/events | https://github.com/huggingface/transformers/issues/38733 | 3,133,939,505 | I_kwDOCUB6oc66zB8x | 38,733 | GRPO per_device_eval_batch_size can't be set as 1, when there is only 1 GPU | {
"login": "CasanovaLLL",
"id": 98315436,
"node_id": "U_kgDOBdwsrA",
"avatar_url": "https://avatars.githubusercontent.com/u/98315436?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/CasanovaLLL",
"html_url": "https://github.com/CasanovaLLL",
"followers_url": "https://api.github.com/users/CasanovaLLL/followers",
"following_url": "https://api.github.com/users/CasanovaLLL/following{/other_user}",
"gists_url": "https://api.github.com/users/CasanovaLLL/gists{/gist_id}",
"starred_url": "https://api.github.com/users/CasanovaLLL/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/CasanovaLLL/subscriptions",
"organizations_url": "https://api.github.com/users/CasanovaLLL/orgs",
"repos_url": "https://api.github.com/users/CasanovaLLL/repos",
"events_url": "https://api.github.com/users/CasanovaLLL/events{/privacy}",
"received_events_url": "https://api.github.com/users/CasanovaLLL/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | [] | closed | false | null | [] | null | [] | 2025-06-10T14:58:11 | 2025-06-11T09:45:32 | 2025-06-11T09:45:32 | NONE | null | null | null | null | `eval batch size must be evenly divisible by the number of generations per prompt. ` When I only have one GPU, I cannot set `per_device_eval_batch_size=1` because there will be no reasonable G to choose from. Is it possible to automatically calculate a value similar to the number of gradient accumulation steps to achieve this feature? | {
"login": "CasanovaLLL",
"id": 98315436,
"node_id": "U_kgDOBdwsrA",
"avatar_url": "https://avatars.githubusercontent.com/u/98315436?v=4",
"gravatar_id": "",
"url": "https://api.github.com/users/CasanovaLLL",
"html_url": "https://github.com/CasanovaLLL",
"followers_url": "https://api.github.com/users/CasanovaLLL/followers",
"following_url": "https://api.github.com/users/CasanovaLLL/following{/other_user}",
"gists_url": "https://api.github.com/users/CasanovaLLL/gists{/gist_id}",
"starred_url": "https://api.github.com/users/CasanovaLLL/starred{/owner}{/repo}",
"subscriptions_url": "https://api.github.com/users/CasanovaLLL/subscriptions",
"organizations_url": "https://api.github.com/users/CasanovaLLL/orgs",
"repos_url": "https://api.github.com/users/CasanovaLLL/repos",
"events_url": "https://api.github.com/users/CasanovaLLL/events{/privacy}",
"received_events_url": "https://api.github.com/users/CasanovaLLL/received_events",
"type": "User",
"user_view_type": "public",
"site_admin": false
} | {
"url": "https://api.github.com/repos/huggingface/transformers/issues/38733/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/38733/timeline | null | completed | {
"total": 0,
"completed": 0,
"percent_completed": 0
} | {
"blocked_by": 0,
"total_blocked_by": 0,
"blocking": 0,
"total_blocking": 0
} | false | true |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.