Skip to content

Actions: microsoft/DeepSpeed

Actions

python

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
python.yml
5,204 workflow runs
5,204 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[inf] Add config var to enable keeping module on host python #11189: Pull request #6846 synchronize by oelayan7
December 17, 2024 07:46 2m 7s oelayan7:keep_module_on_host
oelayan7:keep_module_on_host
December 17, 2024 07:46 2m 7s
[inf] Add config var to enable keeping module on host python #11188: Pull request #6846 synchronize by oelayan7
December 17, 2024 07:39 Action required oelayan7:keep_module_on_host
oelayan7:keep_module_on_host
December 17, 2024 07:39 Action required
Fix error caused by all_reduce call in domino python #11187: Pull request #6880 synchronize by hwchen2017
December 17, 2024 01:46 2m 11s hongwei/fix_domino_allreduce
hongwei/fix_domino_allreduce
December 17, 2024 01:46 2m 11s
Add arctic model support by adding w2 to all_reduce python #11186: Pull request #6856 synchronize by tjruwase
December 17, 2024 01:35 2m 12s pi314ever:arctic-enabling-upstream
pi314ever:arctic-enabling-upstream
December 17, 2024 01:35 2m 12s
Adds ignore_index to sequence parallel cross entropy python #11185: Pull request #6882 opened by ronald-d-rogers
December 17, 2024 01:08 2m 11s ronald-d-rogers:add-ignore-index-sp-loss
ronald-d-rogers:add-ignore-index-sp-loss
December 17, 2024 01:08 2m 11s
Fix checkpointable_layers Logic python #11184: Pull request #6881 opened by Quentin-Anthony
December 17, 2024 00:11 2m 24s Quentin-Anthony:qanthony/fix-act-recomp
Quentin-Anthony:qanthony/fix-act-recomp
December 17, 2024 00:11 2m 24s
python python #11183: Scheduled
December 17, 2024 00:10 2m 13s master
master
December 17, 2024 00:10 2m 13s
Fix error caused by all_reduce call in domino python #11182: Pull request #6880 synchronize by hwchen2017
December 16, 2024 23:50 2m 11s hongwei/fix_domino_allreduce
hongwei/fix_domino_allreduce
December 16, 2024 23:50 2m 11s
Fix error caused by all_reduce call in domino python #11181: Pull request #6880 opened by hwchen2017
December 16, 2024 23:47 2m 13s hongwei/fix_domino_allreduce
hongwei/fix_domino_allreduce
December 16, 2024 23:47 2m 13s
Inference ops unit test failures/fixes python #11180: Pull request #6879 opened by loadams
December 16, 2024 23:08 2m 13s loadams/inference-ops-test-repro
loadams/inference-ops-test-repro
December 16, 2024 23:08 2m 13s
Zero2: avoid graph breaks in torch.compile by using param_idx python #11179: Pull request #6803 synchronize by loadams
December 16, 2024 22:52 2m 9s nelyahu:zero2_param_idx
nelyahu:zero2_param_idx
December 16, 2024 22:52 2m 9s
Fix --enable_each_rank_log when used with PDSH multi-node runner python #11178: Pull request #6863 synchronize by loadams
December 16, 2024 22:49 2m 11s akeshet:akeshet/pdsh_rank_log
akeshet:akeshet/pdsh_rank_log
December 16, 2024 22:49 2m 11s
Add the missing view operations from sequence parallel(async). python #11177: Pull request #6750 synchronize by loadams
December 16, 2024 22:49 2m 15s inkcherry:ds_overlap_fix
inkcherry:ds_overlap_fix
December 16, 2024 22:49 2m 15s
Zero2: avoid graph breaks in torch.compile by using param_idx python #11176: Pull request #6803 synchronize by loadams
December 16, 2024 22:15 2m 10s nelyahu:zero2_param_idx
nelyahu:zero2_param_idx
December 16, 2024 22:15 2m 10s
Fix --enable_each_rank_log when used with PDSH multi-node runner python #11175: Pull request #6863 synchronize by loadams
December 16, 2024 21:28 2m 7s akeshet:akeshet/pdsh_rank_log
akeshet:akeshet/pdsh_rank_log
December 16, 2024 21:28 2m 7s
Fix: forbid repeated deepspeed.initialize on training objects python #11174: Pull request #6874 synchronize by traincheck-team
December 16, 2024 21:02 Action required traincheck-team:fix-6848-forbid-repeated-init
traincheck-team:fix-6848-forbid-repeated-init
December 16, 2024 21:02 Action required
Fix: forbid repeated deepspeed.initialize on training objects python #11173: Pull request #6874 synchronize by traincheck-team
December 16, 2024 20:59 Action required traincheck-team:fix-6848-forbid-repeated-init
traincheck-team:fix-6848-forbid-repeated-init
December 16, 2024 20:59 Action required
Support pure meta model lm_head tp python #11172: Pull request #6812 synchronize by loadams
December 16, 2024 19:34 2m 15s Yejing-Lai:lyj/lm_head_replace
Yejing-Lai:lyj/lm_head_replace
December 16, 2024 19:34 2m 15s
Add MLP/lm_head tp grain size setting. python #11171: Pull request #6828 synchronize by loadams
December 16, 2024 19:33 2m 13s Yejing-Lai:lyj/tp_grain_size
Yejing-Lai:lyj/tp_grain_size
December 16, 2024 19:33 2m 13s
Add the missing view operations from sequence parallel(async). python #11170: Pull request #6750 synchronize by loadams
December 16, 2024 19:33 2m 13s inkcherry:ds_overlap_fix
inkcherry:ds_overlap_fix
December 16, 2024 19:33 2m 13s
Fix --enable_each_rank_log when used with PDSH multi-node runner python #11169: Pull request #6863 synchronize by loadams
December 16, 2024 19:06 2m 11s akeshet:akeshet/pdsh_rank_log
akeshet:akeshet/pdsh_rank_log
December 16, 2024 19:06 2m 11s
Remove pin from transformers version and fix Processing/Threading issues in tests python #11168: Pull request #6822 synchronize by loadams
December 16, 2024 17:43 2m 19s loadams/unpin-transformers-download-model
loadams/unpin-transformers-download-model
December 16, 2024 17:43 2m 19s
Remove pin from transformers version and fix Processing/Threading issues in tests python #11167: Pull request #6822 synchronize by loadams
December 16, 2024 17:40 2m 23s loadams/unpin-transformers-download-model
loadams/unpin-transformers-download-model
December 16, 2024 17:40 2m 23s
Fix --enable_each_rank_log when used with PDSH multi-node runner python #11166: Pull request #6863 synchronize by loadams
December 16, 2024 17:16 2m 14s akeshet:akeshet/pdsh_rank_log
akeshet:akeshet/pdsh_rank_log
December 16, 2024 17:16 2m 14s
Add arctic model support by adding w2 to all_reduce python #11165: Pull request #6856 synchronize by tjruwase
December 16, 2024 12:24 2m 11s pi314ever:arctic-enabling-upstream
pi314ever:arctic-enabling-upstream
December 16, 2024 12:24 2m 11s