Skip to content

Actions: microsoft/DeepSpeed

Actions

python

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
python.yml
5,008 workflow runs
5,008 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Use default value of initial_scale_power if FP16 scaling params not provided python #11495: Pull request #4986 synchronize by loadams
January 22, 2025 17:28 2m 37s ShukantPal:shukant/fix-initial-scale
ShukantPal:shukant/fix-initial-scale
January 22, 2025 17:28 2m 37s
Update sharded_moe.py to support top2 gate with Tutel python #11494: Pull request #6948 synchronize by loadams
January 22, 2025 17:16 2m 22s xenshinu:patch-1
xenshinu:patch-1
January 22, 2025 17:16 2m 22s
Precisely track nvme optimizer offload python #11493: Pull request #6963 synchronize by tjruwase
January 22, 2025 15:54 2m 28s olruwase/ds_4998
olruwase/ds_4998
January 22, 2025 15:54 2m 28s
Autotp training python #11492: Pull request #6922 synchronize by inkcherry
January 22, 2025 05:40 2m 18s inkcherry:autotp_training
inkcherry:autotp_training
January 22, 2025 05:40 2m 18s
Enabled configurable auto Tensor Parallelism (TP) for the inference of diverse models python #11491: Pull request #6553 synchronize by gyou2021
January 22, 2025 03:03 Action required gyou2021:configurable_autoTP
gyou2021:configurable_autoTP
January 22, 2025 03:03 Action required
python python #11490: Scheduled
January 22, 2025 00:09 2m 12s master
master
January 22, 2025 00:09 2m 12s
Explicitly use the linalg.vector_norm call in comm/ python #11489: Pull request #6960 synchronize by loadams
January 21, 2025 22:35 2m 11s loadams/fix-torch-linalg-norm
loadams/fix-torch-linalg-norm
January 21, 2025 22:35 2m 11s
generalize deepspeed linear and implement it for non cuda systems python #11488: Pull request #6932 synchronize by loadams
January 21, 2025 22:34 2m 15s oelayan7:linear
oelayan7:linear
January 21, 2025 22:34 2m 15s
Update version.txt after 0.16.3 release python #11487: Pull request #6965 opened by loadams
January 21, 2025 22:31 2m 11s AutoPR/0.16.3
AutoPR/0.16.3
January 21, 2025 22:31 2m 11s
generalize deepspeed linear and implement it for non cuda systems python #11485: Pull request #6932 synchronize by loadams
January 21, 2025 21:54 Action required oelayan7:linear
oelayan7:linear
January 21, 2025 21:54 Action required
Add Cache to Comm Group python #11484: Pull request #4849 synchronize by loadams
January 21, 2025 20:42 2m 9s cholmes/comm-group-cache
cholmes/comm-group-cache
January 21, 2025 20:42 2m 9s
Explicitly use the linalg.vector_norm call in comm/ python #11483: Pull request #6960 synchronize by loadams
January 21, 2025 20:35 2m 10s loadams/fix-torch-linalg-norm
loadams/fix-torch-linalg-norm
January 21, 2025 20:35 2m 10s
Support the parallel conversion from ZeRO checkpoints to FP32/FP16/BF16 param weight python #11482: Pull request #6655 synchronize by xylian86
January 21, 2025 20:33 2m 7s xylian86:parallel_zero_to_fp32_conversion
xylian86:parallel_zero_to_fp32_conversion
January 21, 2025 20:33 2m 7s
Support the parallel conversion from ZeRO checkpoints to FP32/FP16/BF16 param weight python #11481: Pull request #6655 synchronize by xylian86
January 21, 2025 20:17 Action required xylian86:parallel_zero_to_fp32_conversion
xylian86:parallel_zero_to_fp32_conversion
January 21, 2025 20:17 Action required
Fix: forbid repeated deepspeed.initialize on training objects python #11480: Pull request #6874 synchronize by loadams
January 21, 2025 19:49 Action required traincheck-team:fix-6848-forbid-repeated-init
traincheck-team:fix-6848-forbid-repeated-init
January 21, 2025 19:49 Action required
Enabled configurable auto Tensor Parallelism (TP) for the inference of diverse models python #11479: Pull request #6553 synchronize by loadams
January 21, 2025 19:49 2m 25s gyou2021:configurable_autoTP
gyou2021:configurable_autoTP
January 21, 2025 19:49 2m 25s
Enabled high-performance Automatic Tensor Parallelism (auto TP) for the Qwen2-MoE and DeepSeek-V2 models on multiple GPUs/HPUs python #11478: Pull request #6964 synchronize by loadams
January 21, 2025 19:02 Action required gyou2021:autoTP_Qwen2Moe_DeepSeekv2
gyou2021:autoTP_Qwen2Moe_DeepSeekv2
January 21, 2025 19:02 Action required
fix: RuntimeError for UCP large DP python #11477: Pull request #6918 synchronize by loadams
January 21, 2025 18:51 2m 11s saforem2/ucp-bug
saforem2/ucp-bug
January 21, 2025 18:51 2m 11s
python python #11476: Merge group checks requested
January 21, 2025 18:48 2m 14s
January 21, 2025 18:48 2m 14s
Explicitly use the linalg.vector_norm call in comm/ python #11475: Pull request #6960 synchronize by loadams
January 21, 2025 18:29 2m 10s loadams/fix-torch-linalg-norm
loadams/fix-torch-linalg-norm
January 21, 2025 18:29 2m 10s
python python #11474: Merge group checks requested
January 21, 2025 16:50 2m 10s
January 21, 2025 16:50 2m 10s
python python #11473: Merge group checks requested
January 21, 2025 16:49 2m 18s
January 21, 2025 16:49 2m 18s
Enabled high-performance Automatic Tensor Parallelism (auto TP) for the Qwen2-MoE and DeepSeek-V2 models on multiple GPUs/HPUs python #11472: Pull request #6964 synchronize by gyou2021
January 21, 2025 10:12 Action required gyou2021:autoTP_Qwen2Moe_DeepSeekv2
gyou2021:autoTP_Qwen2Moe_DeepSeekv2
January 21, 2025 10:12 Action required
Enabled high-performance Automatic Tensor Parallelism (auto TP) for the Qwen2-MoE and DeepSeek-V2 models on multiple GPUs/HPUs python #11471: Pull request #6964 synchronize by gyou2021
January 21, 2025 09:03 Action required gyou2021:autoTP_Qwen2Moe_DeepSeekv2
gyou2021:autoTP_Qwen2Moe_DeepSeekv2
January 21, 2025 09:03 Action required
Enabled configurable auto Tensor Parallelism (TP) for the inference of diverse models python #11470: Pull request #6553 synchronize by gyou2021
January 21, 2025 08:59 Action required gyou2021:configurable_autoTP
gyou2021:configurable_autoTP
January 21, 2025 08:59 Action required