Skip to content

Actions: deepspeedai/DeepSpeed

Actions

python

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
python.yml
2,288 workflow runs
2,288 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
Add native (DeepNVMe) host-memory pinning backend for accelerators python #16152: Pull request #8211 synchronize by sfc-gh-truwase
2m 7s tjruwase/native-pin-memory
tjruwase/native-pin-memory
2m 7s
Unmanaged gradient accumulation: ZeRO stage 3 support python #16151: Pull request #8217 synchronize by sfc-gh-truwase
1m 39s sfc-gh-truwase/gas_mgmt_zero3
sfc-gh-truwase/gas_mgmt_zero3
1m 39s
Add native (DeepNVMe) host-memory pinning backend for accelerators python #16150: Pull request #8211 synchronize by sfc-gh-truwase
1m 42s tjruwase/native-pin-memory
tjruwase/native-pin-memory
1m 42s
Add an opt-in DeepEP transport for the AutoEP expert all-to-all python #16149: Pull request #8213 synchronize by yh0903
Action required yh0903:autoep-deepep-backend
yh0903:autoep-deepep-backend
Action required
Unmanaged gradient accumulation: ZeRO stage 3 support python #16148: Pull request #8217 opened by sfc-gh-truwase
4m 15s sfc-gh-truwase/gas_mgmt_zero3
sfc-gh-truwase/gas_mgmt_zero3
4m 15s
Do not let one op builder's compatibility probe break importing deepspeed python #16147: Pull request #8216 synchronize by sfc-gh-truwase
Action required adityasingh2400:fix-op-builder-compat-probe
adityasingh2400:fix-op-builder-compat-probe
Action required
Keep parameter dtype through ZeRO-3 weight quantization python #16146: Pull request #8215 synchronize by adityasingh2400
Action required adityasingh2400:fix-zero-quantized-weights-bf16
adityasingh2400:fix-zero-quantized-weights-bf16
Action required
python python #16145: Merge group checks requested
1m 26s
1m 26s
Add native (DeepNVMe) host-memory pinning backend for accelerators python #16144: Pull request #8211 synchronize by sfc-gh-truwase
1m 39s tjruwase/native-pin-memory
tjruwase/native-pin-memory
1m 39s
python python #16143: Merge group checks requested
2m 47s
2m 47s
Add native (DeepNVMe) host-memory pinning backend for accelerators python #16142: Pull request #8211 synchronize by sfc-gh-truwase
1m 47s tjruwase/native-pin-memory
tjruwase/native-pin-memory
1m 47s
Do not let one op builder's compatibility probe break importing deepspeed python #16141: Pull request #8216 opened by adityasingh2400
3m 11s adityasingh2400:fix-op-builder-compat-probe
adityasingh2400:fix-op-builder-compat-probe
3m 11s
Keep parameter dtype through ZeRO-3 weight quantization python #16140: Pull request #8215 opened by adityasingh2400
Action required adityasingh2400:fix-zero-quantized-weights-bf16
adityasingh2400:fix-zero-quantized-weights-bf16
Action required
python python #16139: Merge group checks requested
3m 10s
3m 10s
Unmanaged gradient accumulation: ZeRO stage 2 support python #16138: Pull request #8203 synchronize by sfc-gh-truwase
2m 13s sfc-gh-truwase/gas_mgmt_zero2
sfc-gh-truwase/gas_mgmt_zero2
2m 13s
Unmanaged gradient accumulation: ZeRO stage 2 support python #16137: Pull request #8203 synchronize by sfc-gh-truwase
53s sfc-gh-truwase/gas_mgmt_zero2
sfc-gh-truwase/gas_mgmt_zero2
53s
Unmanaged gradient accumulation: ZeRO stage 2 support python #16136: Pull request #8203 synchronize by sfc-gh-truwase
1m 43s sfc-gh-truwase/gas_mgmt_zero2
sfc-gh-truwase/gas_mgmt_zero2
1m 43s
python python #16135: Merge group checks requested
1m 28s
1m 28s
[DeepCompile] Fix KeyError on frozen parameters in ZeRO-3 python #16134: Pull request #8214 opened by roycho96
Action required roycho96:fix/deepcompile-frozen-param-grad-buffer
roycho96:fix/deepcompile-frozen-param-grad-buffer
Action required
Share DeepNVMe pinned-tensor manager and route swap buffers through I/O handles python #16133: Pull request #8212 synchronize by sfc-gh-truwase
1m 41s tjruwase/deepnvme-shared-pin-manager
tjruwase/deepnvme-shared-pin-manager
1m 41s
Fix ZeRO++ secondary shard copy for small params python #16132: Pull request #8210 synchronize by ZyEng-Art
Action required ZyEng-Art:zengyong/fix-zeropp-small-param-secondary-shard
ZyEng-Art:zengyong/fix-zeropp-small-param-secondary-shard
Action required
Add an opt-in DeepEP transport for the AutoEP expert all-to-all python #16131: Pull request #8213 synchronize by yh0903
Action required yh0903:autoep-deepep-backend
yh0903:autoep-deepep-backend
Action required
Fix ZeRO++ secondary shard copy for small params python #16130: Pull request #8210 synchronize by ZyEng-Art
Action required ZyEng-Art:zengyong/fix-zeropp-small-param-secondary-shard
ZyEng-Art:zengyong/fix-zeropp-small-param-secondary-shard
Action required
Fix ZeRO++ secondary shard copy for small params python #16129: Pull request #8210 synchronize by ZyEng-Art
Action required ZyEng-Art:zengyong/fix-zeropp-small-param-secondary-shard
ZyEng-Art:zengyong/fix-zeropp-small-param-secondary-shard
Action required
[AutoTP] Complete uneven sharding and universal checkpoint support python #16128: Pull request #8185 synchronize by jinyouzhi
Action required jinyouzhi:uneven
jinyouzhi:uneven
Action required