Commit Graph

18 Commits (f28c0213769dbb2037cd08123ecbf1c8a3f5114b)

Author SHA1 Message Date
ver217 ab8c6b4a0e
[zero] refactor memstats collector (#706)
* refactor memstats collector

* fix disposable

* polish code
2022-04-11 10:46:08 +08:00
アマデウス 54e688b623
moved ensure_path_exists to utils.common (#591) 2022-04-01 09:46:33 +08:00
Liang Bowen ec5086c49c Refactored docstring to google style 2022-03-29 17:17:47 +08:00
HELSON f24b5ed201
[MOE] remove old MoE legacy (#493) 2022-03-22 17:37:16 +08:00
HELSON 7544347145
[MOE] add unitest for MOE experts layout, gradient handler and kernel (#469) 2022-03-21 13:35:04 +08:00
Frank Lee b72b8445c6
optimized context test time consumption (#446) 2022-03-17 14:40:52 +08:00
Jiarui Fang 5a560a060a Feature/zero (#279)
* add zero1 (#209)

* add zero1

* add test zero1

* update zero stage 1 develop (#212)

* Implement naive zero3 (#240)

* naive zero3 works well

* add zero3 param manager

* add TODOs in comments

* add gather full param ctx

* fix sub module streams

* add offload

* fix bugs of hook and add unit tests

* fix bugs of hook and add unit tests (#252)

* add gather full param ctx

* fix sub module streams

* add offload

* fix bugs of hook and add unit tests

* polish code and add state dict hook

* fix bug

* update unit test

* refactor reconstructed zero code

* clip_grad support zero3 and add unit test

* add unit test for Zero3ParameterManager

* [WIP] initialize the shard param class

* [WIP] Yet another sharded model implementation (#274)

* [WIP] initialize the shard param class

* [WIP] Yes another implementation of shardModel. Using a better hook method.

* torch.concat -> torch.cat

* fix test_zero_level_1.py::test_zero_level_1 unitest

* remove deepspeed implementation and refactor for the reconstructed zero module

* polish zero dp unittests

Co-authored-by: ver217 <lhx0217@gmail.com>
Co-authored-by: Frank Lee <somerlee.9@gmail.com>
2022-03-11 15:50:28 +08:00
アマデウス 9ee197d0e9 moved env variables to global variables; (#215)
added branch context;
added vocab parallel layers;
moved split_batch from load_batch to tensor parallel embedding layers;
updated gpt model;
updated unit test cases;
fixed few collective communicator bugs
2022-02-15 11:31:13 +08:00
HELSON 0f8c7f9804
Fixed docstring in colossalai (#171) 2022-01-21 10:44:30 +08:00
Frank Lee e2089c5c15
adapted for sequence parallel (#163) 2022-01-20 13:44:51 +08:00
HELSON dceae85195
Added MoE parallel (#127) 2022-01-07 15:08:36 +08:00
ver217 96780e6ee4
Optimize pipeline schedule (#94)
* add pipeline shared module wrapper and update load batch

* added model parallel process group for amp and clip grad (#86)

* added model parallel process group for amp and clip grad

* update amp and clip with model parallel process group

* remove pipeline_prev/next group (#88)

* micro batch offload

* optimize pipeline gpu memory usage

* pipeline can receive tensor shape (#93)

* optimize pipeline gpu memory usage

* fix grad accumulation step counter

* rename classes and functions

Co-authored-by: Frank Lee <somerlee.9@gmail.com>
2021-12-30 15:56:46 +08:00
アマデウス 01a80cd86d
Hotfix/Colossalai layers (#92)
* optimized 1d layer apis; reorganized nn.layer modules; fixed tests

* fixed 2.5d runtime issue

* reworked split batch, now called in trainer.schedule.load_batch

Co-authored-by: BoxiangW <45734921+BoxiangW@users.noreply.github.com>
2021-12-29 23:32:10 +08:00
ver217 8f02a88db2
add interleaved pipeline, fix naive amp and update pipeline model initializer (#80) 2021-12-20 23:26:19 +08:00
Frank Lee 35813ed3c4
update examples and sphnix docs for the new api (#63) 2021-12-13 22:07:01 +08:00
Frank Lee da01c234e1
Develop/experiments (#59)
* Add gradient accumulation, fix lr scheduler

* fix FP16 optimizer and adapted torch amp with tensor parallel (#18)

* fixed bugs in compatibility between torch amp and tensor parallel and performed some minor fixes

* fixed trainer

* Revert "fixed trainer"

This reverts commit 2e0b0b7699.

* improved consistency between trainer, engine and schedule (#23)

Co-authored-by: 1SAA <c2h214748@gmail.com>

* Split conv2d, class token, positional embedding in 2d, Fix random number in ddp
Fix convergence in cifar10, Imagenet1000

* Integrate 1d tensor parallel in Colossal-AI (#39)

* fixed 1D and 2D convergence (#38)

* optimized 2D operations

* fixed 1D ViT convergence problem

* Feature/ddp (#49)

* remove redundancy func in setup (#19) (#20)

* use env to control the language of doc (#24) (#25)

* Support TP-compatible Torch AMP and Update trainer API (#27)

* Add gradient accumulation, fix lr scheduler

* fix FP16 optimizer and adapted torch amp with tensor parallel (#18)

* fixed bugs in compatibility between torch amp and tensor parallel and performed some minor fixes

* fixed trainer

* Revert "fixed trainer"

This reverts commit 2e0b0b7699.

* improved consistency between trainer, engine and schedule (#23)

Co-authored-by: 1SAA <c2h214748@gmail.com>

Co-authored-by: 1SAA <c2h214748@gmail.com>
Co-authored-by: ver217 <lhx0217@gmail.com>

* add an example of ViT-B/16 and remove w_norm clipping in LAMB (#29)

* add explanation for ViT example (#35) (#36)

* support torch ddp

* fix loss accumulation

* add log for ddp

* change seed

* modify timing hook

Co-authored-by: Frank Lee <somerlee.9@gmail.com>
Co-authored-by: 1SAA <c2h214748@gmail.com>
Co-authored-by: binmakeswell <binmakeswell@gmail.com>

* Feature/pipeline (#40)

* remove redundancy func in setup (#19) (#20)

* use env to control the language of doc (#24) (#25)

* Support TP-compatible Torch AMP and Update trainer API (#27)

* Add gradient accumulation, fix lr scheduler

* fix FP16 optimizer and adapted torch amp with tensor parallel (#18)

* fixed bugs in compatibility between torch amp and tensor parallel and performed some minor fixes

* fixed trainer

* Revert "fixed trainer"

This reverts commit 2e0b0b7699.

* improved consistency between trainer, engine and schedule (#23)

Co-authored-by: 1SAA <c2h214748@gmail.com>

Co-authored-by: 1SAA <c2h214748@gmail.com>
Co-authored-by: ver217 <lhx0217@gmail.com>

* add an example of ViT-B/16 and remove w_norm clipping in LAMB (#29)

* add explanation for ViT example (#35) (#36)

* optimize communication of pipeline parallel

* fix grad clip for pipeline

Co-authored-by: Frank Lee <somerlee.9@gmail.com>
Co-authored-by: 1SAA <c2h214748@gmail.com>
Co-authored-by: binmakeswell <binmakeswell@gmail.com>

* optimized 3d layer to fix slow computation ; tested imagenet performance with 3d; reworked lr_scheduler config definition; fixed launch args; fixed some printing issues; simplified apis of 3d layers (#51)

* Update 2.5d layer code to get a similar accuracy on imagenet-1k dataset

* update api for better usability (#58)

update api for better usability

Co-authored-by: 1SAA <c2h214748@gmail.com>
Co-authored-by: ver217 <lhx0217@gmail.com>
Co-authored-by: puck_WCR <46049915+WANG-CR@users.noreply.github.com>
Co-authored-by: binmakeswell <binmakeswell@gmail.com>
Co-authored-by: アマデウス <kurisusnowdeng@users.noreply.github.com>
Co-authored-by: BoxiangW <45734921+BoxiangW@users.noreply.github.com>
2021-12-09 15:08:29 +08:00
Frank Lee 3defa32aee
Support TP-compatible Torch AMP and Update trainer API (#27)
* Add gradient accumulation, fix lr scheduler

* fix FP16 optimizer and adapted torch amp with tensor parallel (#18)

* fixed bugs in compatibility between torch amp and tensor parallel and performed some minor fixes

* fixed trainer

* Revert "fixed trainer"

This reverts commit 2e0b0b7699.

* improved consistency between trainer, engine and schedule (#23)

Co-authored-by: 1SAA <c2h214748@gmail.com>

Co-authored-by: 1SAA <c2h214748@gmail.com>
Co-authored-by: ver217 <lhx0217@gmail.com>
2021-11-18 19:45:06 +08:00
zbian 404ecbdcc6 Migrated project 2021-10-28 18:21:23 +02:00