cbaa104216
release FP8 news ( #6068 )
...
* add FP8 news
* release FP8 news
* release FP8 news
2024-09-25 11:57:16 +08:00
8501202a35
Merge pull request #6065 from duanjunwen/dev/zero_bubble
...
[Feat] Support zero bubble with shardformer input
2024-09-24 19:17:37 +08:00
7e6f793c51
[fix] fix detach_output_obj clone;
2024-09-24 08:08:32 +00:00
6c1e1550ae
[fix] fix dumb clone;
2024-09-23 06:43:49 +00:00
a875212a42
[fix] fix ci --> oom in 4096 hidden dim;
2024-09-23 05:55:16 +00:00
c114d1429a
[fix] fix detach clone release order;
2024-09-23 04:00:24 +00:00
da3220f48c
[fix] fix pipeline util func deallocate --> release_tensor_data; fix bwd_b loss bwd branch;
2024-09-20 09:48:35 +00:00
1739df423c
[fix] fix fwd branch, fwd pass both micro_batch & internal_inputs'
2024-09-20 07:34:43 +00:00
b6616f544e
[fix] rm comments;
2024-09-20 07:29:41 +00:00
c6d6ee39bd
[fix] use tree_flatten replace dict traverse;
2024-09-20 07:18:49 +00:00
26783776f1
[fix] fix input_tensors buffer append input_obj(dict) --> Tuple (microbatch, input_obj) , and all bwd b related cal logic;
2024-09-20 06:41:19 +00:00
4753bf7add
[fix] fix mem assert;
2024-09-19 08:27:47 +00:00
a115106f8d
[fix] fix bwd w input;
2024-09-19 08:10:05 +00:00
349272c71f
[fix] updatw bwd b&w input; dict --> list[torch.Tensor]
2024-09-19 07:47:01 +00:00
6ee9584b9a
[fix] fix require_grad & deallocate call;
2024-09-19 05:53:03 +00:00
1f5c7258aa
Merge remote-tracking branch 'upstream/feature/zerobubble' into dev/zero_bubble
2024-09-19 03:52:13 +00:00
dabc2e7430
[release] update version ( #6062 )
2024-09-19 10:45:32 +08:00
f9546ba0be
[ColossalEval] support for vllm ( #6056 )
...
* support vllm
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* modify vllm and update readme
* run pre-commit
* remove dupilicated lines and refine code
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* update param name
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* refine code
* update readme
* refine code
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2024-09-18 17:09:45 +08:00
af2c2f8092
[feat] add more test;
2024-09-18 07:51:54 +00:00
3dbad102cf
[fix] fix zerobubble pp for shardformer type input;
2024-09-18 07:14:34 +00:00
4fa6b9509c
[moe] add parallel strategy for shared_expert && fix test for deepseek ( #6063 )
2024-09-18 10:09:01 +08:00
63314ce4e4
Merge pull request #6064 from wangbluo/fix_attn
...
[sp] : fix the attention kernel for sp
2024-09-18 10:08:15 +08:00
10e4f7da72
fix
2024-09-16 13:45:04 +08:00
37e35230ff
Merge pull request #6061 from wangbluo/sp_fix
...
[sp] : fix the attention kernel for sp
2024-09-14 20:54:35 +08:00
827ef3ee9a
fix
2024-09-14 10:40:35 +00:00
bdb125f83f
[doc] FP8 training and communication document ( #6050 )
...
* Add FP8 training and communication document
* add fp8 docstring for plugins
* fix typo
* fix typo
2024-09-14 11:01:05 +08:00
f20b066c59
[fp8] Disable all_gather intranode. Disable Redundant all_gather fp8 ( #6059 )
...
* all_gather only internode, fix pytest
* fix cuda arch <89 compile pytest error
* fix pytest failure
* disable all_gather_into_tensor_flat_fp8
* fix fp8 format
* fix pytest
* fix conversations
* fix chunk tuple to list
2024-09-14 10:40:01 +08:00
b582319273
fix
2024-09-13 10:24:41 +00:00
0ad3129cb9
fix
2024-09-13 09:01:26 +00:00
0b14a5512e
fix
2024-09-13 07:06:14 +00:00
696fced0d7
[fp8] fix missing fp8_comm flag in mixtral ( #6057 )
2024-09-13 14:30:05 +08:00
dc032172c3
fix
2024-09-13 06:00:58 +00:00
f393867cff
fix
2024-09-13 05:24:52 +00:00
6eb8832366
fix
2024-09-13 05:06:56 +00:00
683179cefd
fix
2024-09-13 03:40:56 +00:00
0a01e2a453
fix the attn
2024-09-13 03:38:35 +00:00
216d54e374
[pre-commit.ci] auto fixes from pre-commit.com hooks
...
for more information, see https://pre-commit.ci
2024-09-13 02:38:40 +00:00
fdd84b9087
fix the sp
2024-09-13 02:32:03 +00:00
9bc3b6e220
[feat] moehybrid support zerobubble;
2024-09-12 02:51:46 +00:00
a35a078f08
[doc] update sp doc ( #6055 )
...
* update sp doc
* fix
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* fix
* fix
* fix
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2024-09-11 17:25:14 +08:00
13946c4448
[fp8] hotfix backward hook ( #6053 )
...
* [fp8] hotfix backward hook
* [fp8] hotfix pipeline loss accumulation
2024-09-11 16:11:25 +08:00
11ae6848c6
[zerobubble]Support ZeroBubble Pipeline ( #6034 )
...
* [feat] add zerobubble pp (just a frame now); add POC test for dx_dw; add test for zerobubble;
* [feat] add dw test;
* [fix] fix weight not close;
* [update] update text;
* [feat] add test run_fwd_bwd automatic scheduling;
* [feat] split communication and calculation; fix pop empty send_bwd_buffer error;
* [feat] add test for p & p grad;
* [feat] add comments for ZBV func;
* [fix] rm useless assign and comments;
* [fix] fix ci test; add pytest;
* [feat] add run_fwd_bwd_with_microbatch (replace input) & test; add p&p.grad assert close test & all pass;
* [feat] add apply v_schedule graph; p & p.grad assert err exist;
* [fix] update
* [feat] fix ci; add assert;
* [feat] fix poc format
* [feat] fix func name & ci; add comments;
* [fix] fix poc test; add comments in poc;
* [feat] add optim backward_b_by_grad
* [feat] fix optimizer bwd b & w; support return accum loss & output
* [feat] add fwd_bwd_step, run_fwd_only;
* [fix] fix optim bwd; add license for v_schedule; remove redundant attributes; fix schedule loop "while"--> "for"; add communication dict;
* [fix] fix communication_map;
* [feat] update test; rm comments;
* [fix] rm zbv in hybridplugin
* [fix] fix optim bwd;
* [fix] fix optim bwd;
* [fix] rm output.data after send fwd;
* [fix] fix bwd step if condition; remove useless comments and format info;
* [fix] fix detach output & release output;
* [fix] rm requir_grad for output;
* [fix] fix requir grad position and detach position and input&output local buffer append position;
* [feat] add memory assertation;
* [fix] fix mem check;
* [fix] mem assertation'
* [fix] fix mem assertation
* [fix] fix mem; use a new model shape; only assert mem less and equal than theo;
* [fix] fix model zoo import;
* [fix] fix redundant detach & clone; add buffer assertation in the end;
* [fix] add output_obj_grad assert None at bwd b step; replace input_obj.require_grad_ with treemap;
* [fix] update optim state dict assert (include param group & state); fix mem assert after add optim;
* [fix] add testcase with microbatch 4;
2024-09-10 17:33:09 +08:00
c54c4fcd15
[hotfix] moe hybrid parallelism benchmark & follow-up fix ( #6048 )
...
* [example] pass use_fp8_comm flag to all plugins
* [example] add mixtral benchmark
* [moe] refine assertion and check
* [moe] fix mixtral & add more tests
* [moe] consider checking dp * sp group and moe_dp_group
* [mixtral] remove gate tp & add more tests
* [deepseek] fix tp & sp for deepseek
* [mixtral] minor fix
* [deepseek] add deepseek benchmark
2024-09-10 17:30:53 +08:00
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
b3db1058ec
[release] update version ( #6041 )
...
* [release] update version
* [devops] update comp test
* [devops] update comp test debug
* [devops] debug comp test
* [devops] debug comp test
* [devops] debug comp test
* [devops] debug comp test
* [devops] debug comp test
2024-09-10 10:31:09 +08:00
6c2a120bed
[fix] add testcase with microbatch 4;
2024-09-09 10:16:03 +00:00
8366a7855f
[fix] update optim state dict assert (include param group & state); fix mem assert after add optim;
2024-09-09 09:27:13 +00:00
ce58d8e8bf
[fix] add output_obj_grad assert None at bwd b step; replace input_obj.require_grad_ with treemap;
2024-09-09 08:19:58 +00:00
7568b34626
[fix] fix redundant detach & clone; add buffer assertation in the end;
2024-09-09 08:04:28 +00:00
fed8b1587d
[fix] fix model zoo import;
2024-09-09 06:39:33 +00:00
binmakeswell