Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Support agent training, etc. #352

Merged
merged 54 commits into from
Feb 1, 2024

Conversation

tastelikefeet
Copy link
Collaborator

@tastelikefeet tastelikefeet commented Jan 30, 2024

  1. Support sft_beta arg to DPOTrainer, which may stabilize the model generation.
  2. Support Adalora and IA3
  3. Support Agent training
  4. Support dataset mixture
  5. Support two new datasets: ms-agent and ms-bench
  6. Support AnimateDiff merge lora
  7. Fix some bugs in web-ui

swift/llm/agent/models/xverse.py Outdated Show resolved Hide resolved
swift/llm/utils/argument.py Show resolved Hide resolved
swift/llm/utils/argument.py Show resolved Hide resolved
swift/llm/utils/preprocess.py Outdated Show resolved Hide resolved
swift/llm/utils/template.py Show resolved Hide resolved
…t_module

* commit '12e731a713c3f2d7f8c02038bf477e593e5f1240':
  Removing eos_token when doing inference. (modelscope#351)
  Update internlm2 math (modelscope#349)
  Fix template encode bug (modelscope#348)
  fix yi-vl finetune error (modelscope#347)
  Support yi vl (modelscope#345)
  update codefuse series (modelscope#343)
  Update orion 14b (modelscope#341)
  update default_lr; fix do_sample in vllm (modelscope#336)
  Fix ui (modelscope#335)

# Conflicts:
#	swift/llm/infer.py
#	swift/llm/utils/argument.py
#	swift/llm/utils/template.py
#	swift/llm/utils/utils.py
#	swift/trainers/dpo_trainers.py
@yingdachen
Copy link
Contributor

plus also update the pr title and commit msg

@tastelikefeet tastelikefeet changed the title [WIP] many features Support agent training, etc. Feb 1, 2024
…t_module

* commit '545f17e1d6585fbc0dd0ea45adde06b6c4ade2d0':
  support openbmb minicpm (modelscope#364)
  support dpo cli and add examples controlnet and dreambooth (modelscope#344)
  Fix openbmb model name (modelscope#362)

# Conflicts:
#	README.md
#	README_CN.md
@tastelikefeet tastelikefeet merged commit 50eb82b into modelscope:main Feb 1, 2024
2 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.

None yet

3 participants