Skip to content

Actions: deepspeedai/DeepSpeed

Actions

python

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
python.yml
2,155 workflow runs
2,155 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Enable support for Tutel when k != 1 for shared moe python #15912: Pull request #8174 opened by jinyouzhi
Action required jinyouzhi:tutel
jinyouzhi:tutel
Action required
Stabilize fork-sensitive nightly tests and AutoSP coverage python #15911: Pull request #8172 opened by tohtana
4m 48s tohtana:tohtana/fix-nightly-zero-nvme-hang
tohtana:tohtana/fix-nightly-zero-nvme-hang
4m 48s
Fix WarmupLR collapsing multi-group base LRs to group 0's python #15910: Pull request #8171 opened by ebarkhordar
Action required ebarkhordar:fix/warmuplr-multigroup-max-lr
ebarkhordar:fix/warmuplr-multigroup-max-lr
Action required
[AutoTP] Allow ZeRO stage 3 inference with tensor parallelism python #15909: Pull request #8167 synchronize by delock
1m 41s delock:gma/allow-autotp-with-zero3
delock:gma/allow-autotp-with-zero3
1m 41s
python python #15908: Scheduled
1m 14s master
master
1m 14s
Run pull request code only inside a Modal Sandbox python #15907: Pull request #8170 opened by tohtana
1m 29s tohtana/fix/secure-modal-pr-sandbox
tohtana/fix/secure-modal-pr-sandbox
1m 29s
Fix DeepCompile ZeRO-3 gathered parameter ownership python #15906: Pull request #8157 synchronize by tohtana
1m 28s tohtana:tohtana/deepcompile-z3-gather-ownership
tohtana:tohtana/deepcompile-z3-gather-ownership
1m 28s
Scope DeepCompile compiler state to graph and engine lifecycles python #15905: Pull request #8159 synchronize by tohtana
1m 22s tohtana:tohtana/deepcompile-compiler-lifecycle-standalone
tohtana:tohtana/deepcompile-compiler-lifecycle-standalone
1m 22s
Stabilize DeepCompile ZeRO-3 memory scheduling python #15904: Pull request #8169 opened by tohtana
1m 39s tohtana:tohtana/deepcompile-zero3-memory-stability-scheduler-standalone
tohtana:tohtana/deepcompile-zero3-memory-stability-scheduler-standalone
1m 39s
NVMe write warning python #15903: Pull request #8164 synchronize by loadams
1m 31s sfc-gh-truwase/nvme_warning
sfc-gh-truwase/nvme_warning
1m 31s
[AutoTP] Fix ZeRO-3 checkpoint consolidation to gather across TP and DP python #15902: Pull request #8168 synchronize by delock
1m 32s delock:gma/investigate-autotp-zero3-ucp
delock:gma/investigate-autotp-zero3-ucp
1m 32s
[AutoTP] Fix ZeRO-3 checkpoint consolidation to gather across TP and DP python #15901: Pull request #8168 synchronize by delock
1m 25s delock:gma/investigate-autotp-zero3-ucp
delock:gma/investigate-autotp-zero3-ucp
1m 25s
[AutoTP] Fix ZeRO-3 checkpoint consolidation to gather across TP and DP python #15900: Pull request #8168 opened by delock
1m 30s delock:gma/investigate-autotp-zero3-ucp
delock:gma/investigate-autotp-zero3-ucp
1m 30s
[AutoTP] Enalbe HF colwise_gather_output to support lm_head replace python #15899: Pull request #8146 synchronize by jinyouzhi
1m 35s jinyouzhi:lm_head_colwise
jinyouzhi:lm_head_colwise
1m 35s
[AutoTP] Enalbe HF colwise_gather_output to support lm_head replace python #15898: Pull request #8146 synchronize by jinyouzhi
Action required jinyouzhi:lm_head_colwise
jinyouzhi:lm_head_colwise
Action required
[AutoTP] Enalbe HF colwise_gather_output to support lm_head replace python #15897: Pull request #8146 synchronize by jinyouzhi
-1s jinyouzhi:lm_head_colwise
jinyouzhi:lm_head_colwise
-1s
[AutoTP] Allow ZeRO stage 3 inference with tensor parallelism python #15896: Pull request #8167 synchronize by delock
1m 51s delock:gma/allow-autotp-with-zero3
delock:gma/allow-autotp-with-zero3
1m 51s
[AutoTP] Allow ZeRO stage 3 inference with tensor parallelism python #15895: Pull request #8167 synchronize by delock
1m 27s delock:gma/allow-autotp-with-zero3
delock:gma/allow-autotp-with-zero3
1m 27s
[AutoTP] Allow ZeRO stage 3 inference with tensor parallelism python #15894: Pull request #8167 opened by delock
1m 33s delock:gma/allow-autotp-with-zero3
delock:gma/allow-autotp-with-zero3
1m 33s
Guard LRRangeTest and OneCycle schedulers against zero step sizes python #15893: Pull request #8166 opened by ebarkhordar
Action required ebarkhordar:fix/lr-schedules-zero-step-guards
ebarkhordar:fix/lr-schedules-zero-step-guards
Action required
[Pipeline] Add Ray-backed stage placement and execution with pluggable transportsFeature/ray pipeline parallelism python #15892: Pull request #8161 synchronize by SisPiao
Action required SisPiao:feature/ray-pipeline-parallelism
SisPiao:feature/ray-pipeline-parallelism
Action required
Scope DeepCompile compiler state to graph and engine lifecycles python #15891: Pull request #8159 synchronize by tohtana
1m 33s tohtana:tohtana/deepcompile-compiler-lifecycle-standalone
tohtana:tohtana/deepcompile-compiler-lifecycle-standalone
1m 33s
python python #15890: Merge group checks requested
1m 24s
1m 24s
python python #15889: Scheduled
1m 19s master
master
1m 19s
python python #15888: Merge group checks requested
1m 20s
1m 20s