ColossalAI/colossalai/shardformer/policies
Wenxuan Tan 8fd25d6e09
[Feature] Split cross-entropy computation in SP (#5959)
* halfway

* fix cross-PP-stage position id length diff bug

* fix typo

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* unified cross entropy func for all shardformer models

* remove redundant lines

* add basic ring attn; debug cross entropy

* fwd bwd logic complete

* fwd bwd logic complete; add experimental triton rescale

* precision tests passed

* precision tests passed

* fix typos and remove misc files

* update softmax_lse shape by new interface

* change tester name

* remove buffer clone; support packed seq layout

* add varlen tests

* fix typo

* all tests passed

* add dkv_group; fix mask

* remove debug statements

* adapt chatglm, command-R, qwen

* debug

* halfway

* fix cross-PP-stage position id length diff bug

* fix typo

* fix typo

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* unified cross entropy func for all shardformer models

* remove redundant lines

* add basic ring attn; debug cross entropy

* fwd bwd logic complete

* fwd bwd logic complete; add experimental triton rescale

* precision tests passed

* precision tests passed

* fix typos and remove misc files

* add sp_mode to benchmark; fix varlen interface

* update softmax_lse shape by new interface

* add varlen tests

* fix typo

* all tests passed

* add dkv_group; fix mask

* remove debug statements

* add comments

* q1 index only once

* remove events to simplify stream sync

* simplify forward/backward logic

* 2d ring forward passed

* 2d ring backward passed

* fixes

* fix ring attn loss

* 2D ring backward + llama passed

* merge

* update logger

* fix typo

* rebase

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix typo

* remove typos

* fixes

* support GPT

---------

Co-authored-by: Edenzzzz <wtan45@wisc.edu>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2024-09-10 12:06:50 +08:00
..
__init__.py [shardformer] init shardformer code structure (#3731) 2023-07-04 16:05:01 +08:00
auto_policy.py [FP8] rebase main (#5963) 2024-08-06 16:29:37 +08:00
base_policy.py [fp8] Merge feature/fp8_comm to main branch of Colossalai (#6016) 2024-08-22 09:21:34 +08:00
bert.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
blip2.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
bloom.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
chatglm2.py [Feature] Split cross-entropy computation in SP (#5959) 2024-09-10 12:06:50 +08:00
command.py [fp8] Merge feature/fp8_comm to main branch of Colossalai (#6016) 2024-08-22 09:21:34 +08:00
deepseek.py [Feature] Zigzag Ring attention (#5905) 2024-08-16 13:56:38 +08:00
falcon.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
gpt2.py [Feature] Split cross-entropy computation in SP (#5959) 2024-09-10 12:06:50 +08:00
gptj.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
llama.py [Feature] Split cross-entropy computation in SP (#5959) 2024-09-10 12:06:50 +08:00
mistral.py [fp8] Merge feature/fp8_comm to main branch of Colossalai (#6016) 2024-08-22 09:21:34 +08:00
mixtral.py [fp8] Merge feature/fp8_comm to main branch of Colossalai (#6016) 2024-08-22 09:21:34 +08:00
opt.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
qwen2.py [Feature] Zigzag Ring attention (#5905) 2024-08-16 13:56:38 +08:00
sam.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
t5.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
vit.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00
whisper.py [fp8] support hybrid parallel plugin (#5982) 2024-08-12 18:17:05 +08:00