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.

Sign up for GitHub

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

Jump to bottom

Add FasterCodeGen/FasterGPTJ #3017

Merged
merged 28 commits into from
Sep 2, 2022
Merged

Add FasterCodeGen/FasterGPTJ #3017

merged 28 commits into from
Sep 2, 2022

Conversation

gongel
Copy link
Member

@gongel gongel commented Aug 11, 2022 •
edited
Loading

PR types

New features

PR changes

Models

Description

  • Add GPTJ model
  • Add FasterGPTJ
  • Add FasterCodeGen
  • Add codegen_server for Github Copilot
  • Add perf script and doc
  • Add finetune script
  • Add doc for Github Copilot and finetune

@gongel gongel requested a review from guoshengCS August 11, 2022 07:10
@gongel gongel added the faster Make training or inference faster label Aug 11, 2022
guoshengCS
guoshengCS reviewed Aug 16, 2022
View reviewed changes
examples/code_generation/codegen/codegen_server.py Outdated Show resolved Hide resolved
paddlenlp/ops/faster_transformer/transformer/decoding.py Show resolved Hide resolved
paddlenlp/ops/faster_transformer/transformer/decoding.py Show resolved Hide resolved
paddlenlp/ops/patches/FasterTransformer/fastertransformer/cuda/masked_multihead_attention.cu Outdated Show resolved Hide resolved
paddlenlp/transformers/gptj/tokenizer.py Outdated Show resolved Hide resolved
paddlenlp/transformers/gptj/modeling.py Show resolved Hide resolved
paddlenlp/transformers/codegen/modeling.py Outdated Show resolved Hide resolved
paddlenlp/ops/ext_utils.py Outdated Show resolved Hide resolved
paddlenlp/ops/faster_transformer/transformer/decoding.py Outdated Show resolved Hide resolved
guoshengCS
guoshengCS reviewed Aug 16, 2022
View reviewed changes
Copy link
Contributor

@guoshengCS guoshengCS left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

另外要注意验证对其他模型是否有影响

guoshengCS
guoshengCS reviewed Aug 16, 2022
View reviewed changes
Copy link
Contributor

@guoshengCS guoshengCS left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

READM补充更新

@gongel
Copy link
Member Author

gongel commented Aug 17, 2022

另外要注意验证对其他模型是否有影响

经验证，sample下模型没问题

FrostML
FrostML reviewed Aug 26, 2022
View reviewed changes
paddlenlp/ops/faster_transformer/src/fusion_gptj_op.cc Outdated Show resolved Hide resolved
paddlenlp/ops/faster_transformer/src/fusion_gptj_op.cc Show resolved Hide resolved
paddlenlp/ops/faster_transformer/src/fusion_gptj_op.cu Show resolved Hide resolved
paddlenlp/ops/faster_transformer/src/fusion_gptj_op.cu Outdated Show resolved Hide resolved
paddlenlp/ops/patches/FasterTransformer/fastertransformer/cuda/attention_kernels.cu Outdated Show resolved Hide resolved
paddlenlp/ops/patches/FasterTransformer/fastertransformer/cuda/decoding_kernels.cu Outdated Show resolved Hide resolved
paddlenlp/ops/patches/FasterTransformer/fastertransformer/cuda/masked_multihead_attention.h Show resolved Hide resolved
...enlp/ops/patches/FasterTransformer/fastertransformer/cuda/masked_multihead_attention_utils.h Show resolved Hide resolved
paddlenlp/ops/patches/FasterTransformer/fastertransformer/cuda/topk_kernels.cu Outdated Show resolved Hide resolved
paddlenlp/ops/patches/FasterTransformer/fastertransformer/utils/allocator.h Outdated Show resolved Hide resolved
guoshengCS
guoshengCS previously approved these changes Sep 2, 2022
View reviewed changes
Copy link
Contributor

@guoshengCS guoshengCS left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

只剩下TODO没有其他问题的话就尽快合入提测吧

paddlenlp/ops/patches/FasterTransformer/fastertransformer/open_decoder.h Show resolved Hide resolved
@gongel gongel merged commit 53bdc2d into PaddlePaddle:develop Sep 2, 2022
This was referenced Sep 3, 2022
Closed
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@FrostML FrostML FrostML approved these changes

@guoshengCS guoshengCS guoshengCS left review comments

Assignees
No one assigned
Labels
faster Make training or inference faster
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@gongel @FrostML @guoshengCS