Skip to content

Actions: deepspeedai/DeepSpeed

Actions

Formatting

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
formatting.yml
3,207 workflow runs
3,207 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

fix: When there are tensors registered with register buffer in the weight file, the weights are only loaded on device 0 when loading weights across multiple devices. Formatting #18910: Pull request #7717 opened by KeeProMise
Action required KeeProMise:7709
KeeProMise:7709
Action required
fix: The different lr for Muon doesn't work Formatting #18909: Pull request #7716 opened by KeeProMise
Action required KeeProMise:7713
KeeProMise:7713
Action required
Formatting Formatting #18908: Scheduled
1m 29s master
master
1m 29s
Formatting Formatting #18907: Scheduled
1m 27s master
master
1m 27s
Fix SuperOffloadOptimizer_Stage3 crash due to missing param_names parameter Formatting #18906: Pull request #7715 opened by ImaGoodFella
Action required ImaGoodFella:master
ImaGoodFella:master
Action required
Formatting Formatting #18905: Scheduled
1m 29s master
master
1m 29s
Formatting Formatting #18904: Scheduled
1m 30s master
master
1m 30s
Wall clock timers API Formatting #18903: Pull request #7714 synchronize by sfc-gh-truwase
1m 26s sfc-gh-truwase/wall_clock_timers_api
sfc-gh-truwase/wall_clock_timers_api
1m 26s
Wall clock timers API Formatting #18902: Pull request #7714 synchronize by sfc-gh-truwase
1m 29s sfc-gh-truwase/wall_clock_timers_api
sfc-gh-truwase/wall_clock_timers_api
1m 29s
Wall clock timers API Formatting #18901: Pull request #7714 opened by sfc-gh-truwase
1m 28s sfc-gh-truwase/wall_clock_timers_api
sfc-gh-truwase/wall_clock_timers_api
1m 28s
Enabled compiled autograd for backward pass Formatting #18900: Pull request #7667 synchronize by deepcharm
1m 39s deepcharm:enable-compiled-autograd
deepcharm:enable-compiled-autograd
1m 39s
Low-precision master params/grads/optimizer states Formatting #18898: Pull request #7700 synchronize by tohtana
1m 47s tohtana:tohtana/pure_bf16_config
tohtana:tohtana/pure_bf16_config
1m 47s
Low-precision master params/grads/optimizer states Formatting #18897: Pull request #7700 synchronize by tohtana
33s tohtana:tohtana/pure_bf16_config
tohtana:tohtana/pure_bf16_config
33s
Low-precision master params/grads/optimizer states Formatting #18895: Pull request #7700 synchronize by tohtana
1m 31s tohtana:tohtana/pure_bf16_config
tohtana:tohtana/pure_bf16_config
1m 31s
Formatting Formatting #18894: Scheduled
1m 26s master
master
1m 26s
fix: skip aio wait when swap tensors is empty Formatting #18893: Pull request #7712 opened by xylian86
1m 28s xylian86:fix_aio_assertion
xylian86:fix_aio_assertion
1m 28s
Fix that ds_secondary_tensor may be dirty when loading the model or zero checkpoint for zero++. Formatting #18892: Pull request #7707 synchronize by sfc-gh-truwase
1m 25s zhengchenyu:issue-7606
zhengchenyu:issue-7606
1m 25s
Formatting Formatting #18891: Scheduled
1m 18s master
master
1m 18s
Enabled compiled autograd for backward pass Formatting #18890: Pull request #7667 synchronize by deepcharm
2m 16s deepcharm:enable-compiled-autograd
deepcharm:enable-compiled-autograd
2m 16s
[ROCm] Relax tolerances for FP8 unit test for fp16 and bf16 cases Formatting #18889: Pull request #7655 synchronize by sfc-gh-truwase
1m 45s rraminen:relax_tol_testFP8_ROCm
rraminen:relax_tol_testFP8_ROCm
1m 45s
[ROCm] Relax tolerances for FP8 unit test for fp16 and bf16 cases Formatting #18888: Pull request #7655 synchronize by sfc-gh-truwase
1m 26s rraminen:relax_tol_testFP8_ROCm
rraminen:relax_tol_testFP8_ROCm
1m 26s
Formatting Formatting #18887: Scheduled
1m 30s master
master
1m 30s
Formatting Formatting #18886: Scheduled
1m 26s master
master
1m 26s
Formatting Formatting #18885: Scheduled
1m 27s master
master
1m 27s
Formatting Formatting #18884: Scheduled
1m 22s master
master
1m 22s