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 ORPO algorithm #854

Merged
merged 41 commits into from
May 7, 2024
Merged

support ORPO algorithm #854

merged 41 commits into from
May 7, 2024

Conversation

hjh0119
Copy link
Collaborator

@hjh0119 hjh0119 commented Apr 30, 2024

PR type

  • Bug Fix
  • New Feature
  • Document Updates
  • More Models or Datasets Support

PR information

Bug Fix

  • Training and inference for internvl-chat with GPUs that do not support flash attention
  • V100推理internVL-1.5问题 #870
  • Adding timm dependency requirements for several VL models

New Feature

More Models or Datasets Support

@hjh0119 hjh0119 changed the title [WIP] support ORPO algorithm support ORPO algorithm May 6, 2024
@hjh0119 hjh0119 merged commit eb940b1 into modelscope:main May 7, 2024
1 of 2 checks passed
tastelikefeet added a commit to tastelikefeet/swift that referenced this pull request May 7, 2024
…va-llama

* commit 'eb940b1a1e8ac1adcf79f5e566ba3d68b545f069':
  support ORPO algorithm (modelscope#854)
  fix dataset info deepcopy (modelscope#871)
  fix dataset_test_ratio=1 (modelscope#869)

# Conflicts:
#	README.md
#	README_CN.md
tastelikefeet added a commit to tastelikefeet/swift that referenced this pull request May 7, 2024
* main:
  add llava-llama (modelscope#873)
  support ORPO algorithm (modelscope#854)
  fix dataset info deepcopy (modelscope#871)
  fix dataset_test_ratio=1 (modelscope#869)
@hjh0119 hjh0119 deleted the orpo branch May 7, 2024 11:14
tastelikefeet added a commit to tastelikefeet/swift that referenced this pull request May 10, 2024
* main: (24 commits)
  fix pre-commit
  traindataset异常提示 (modelscope#859)
  Feat/pack (modelscope#881)
  fix swift cli exit code if subprocess is failed (modelscope#879)
  support Deepseek-V2-Chat and InternVL-Chat-V1.5-int8 model  (modelscope#876)
  add llava-llama (modelscope#873)
  support ORPO algorithm (modelscope#854)
  fix dataset info deepcopy (modelscope#871)
  fix dataset_test_ratio=1 (modelscope#869)
  Refactor dataset (modelscope#802)
  Feat/loras (modelscope#865)
  Update tuner docs (modelscope#853)
  update docs (modelscope#850)
  Fix code format and docs (modelscope#847)
  update (modelscope#846)
  fix xcomposer device_map (modelscope#844)
  fix merge_lora_dtype (modelscope#842)
  Fix infer default dtype (modelscope#834)
  fix ui (modelscope#830)
  support Internvl-chat-v1.5 model (modelscope#824)
  ...

# Conflicts:
#	docs/source/LLM/自定义与拓展.md
#	docs/source_en/LLM/Customization.md
#	examples/pytorch/llm/custom.py
#	scripts/benchmark/exp_utils.py
#	scripts/utils/run_dataset_info.py
#	swift/aigc/diffusers/train_controlnet.py
#	swift/aigc/diffusers/train_controlnet_sdxl.py
#	swift/aigc/diffusers/train_text_to_image.py
#	swift/aigc/diffusers/train_text_to_image_lora.py
#	swift/aigc/diffusers/train_text_to_image_lora_sdxl.py
#	swift/aigc/diffusers/train_text_to_image_sdxl.py
#	swift/llm/__init__.py
#	swift/llm/deploy.py
#	swift/llm/dpo.py
#	swift/llm/export.py
#	swift/llm/infer.py
#	swift/llm/sft.py
#	swift/llm/tuner.py
#	swift/llm/utils/__init__.py
#	swift/llm/utils/argument.py
#	swift/llm/utils/client_utils.py
#	swift/llm/utils/dataset.py
#	swift/llm/utils/model.py
#	swift/llm/utils/preprocess.py
#	swift/llm/utils/template.py
#	swift/llm/utils/utils.py
#	swift/trainers/trainers.py
#	swift/tuners/base.py
#	swift/ui/llm_infer/llm_infer.py
#	swift/ui/llm_infer/runtime.py
#	swift/ui/llm_train/dataset.py
#	swift/ui/llm_train/llm_train.py
#	tests/llm/test_run.py
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

2 participants