Skip to content

合并master#3

Merged
iugoood merged 35 commits intogemma3from
master
May 23, 2025
Merged

合并master#3
iugoood merged 35 commits intogemma3from
master

Conversation

@iugoood
Copy link
Copy Markdown
Owner

@iugoood iugoood commented May 23, 2025

What does this PR do?

Fixes # (issue)

Adds # (feature)

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?
  • Did you make sure to update the documentation with your changes? E.g. record bug fixes or new features in What's New. Here are the
    documentation guidelines
  • Did you build and run the code without any errors?
  • Did you report the running environment (NPU type/MS version) and performance in the doc? (better record it for data loading, model inference, or training tasks)
  • 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.

@xxx

wcrzlh and others added 30 commits May 14, 2025 20:42
feat(transformers): upgrade modeling_utils/RoPE/optimization/cache_utils etc to v4.50
Add beam_search & Add ForCausalLMLoss
* examples/codellama

* test llama

* update transformers version to 4.50

* update transformers version 4.51

* download checkpoints readme

* correct load name

* move files

* revert pyproject
iugoood and others added 5 commits May 22, 2025 11:42
* feat(transformers): upgrade modeling_utils/RoPE/optimization/cache_utils etc to v4.50

* feat(transformers): upgrade modeling_utils/RoPE/optimization/cache_utils etc to v4.50

* feat(transformers): upgrade modeling_utils/RoPE/optimization/cache_utils etc to v4.50

* feat(transformers): upgrade modeling_utils/RoPE/optimization/cache_utils etc to v4.50

* precommit

* fix cache_util

* update generation_util, logit_processor, stopping_criteria, candidate_generator

* fix cache_utils

* Add beam_search&For ForCausalLMLoss loss

* update modeling_outputs

* update beam_search

* pre-commit check

* add support for dynamic input

* add beam search py

* fix tie_weight dtype mismatch in pynative mode for albert

* fix beam_search

* fix bug of dynamic input

* add _supports_dynamic_input to PretrainedModel

* fix pynative synchronize bug for albert[temporal]

* support multimodal for init_static_cache[hack implementation]

* fix cache_utils bugs

* fix cumsum does not support int64

---------

Co-authored-by: Chaoran Wei <77485245+wcrzlh@users.noreply.github.com>
Co-authored-by: Mark-ZhouWX <zhouwuxing000@gmail.com>
Co-authored-by: liuchuting <liuchuting1@huawei.com>
@iugoood iugoood merged commit dca6379 into gemma3 May 23, 2025
3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

7 participants