Skip to content

Actions: microsoft/DeepSpeed

Actions

hpu-gaudi2

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
hpu-gaudi2.yml
1,334 workflow runs
1,334 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Support pure meta model lm_head tp hpu-gaudi2 #1480: Pull request #6812 synchronize by loadams
January 10, 2025 18:40 1h 46m 34s Yejing-Lai:lyj/lm_head_replace
Yejing-Lai:lyj/lm_head_replace
January 10, 2025 18:40 1h 46m 34s
Set shuffle=True by default in data_sampler hpu-gaudi2 #1479: Pull request #6531 synchronize by loadams
January 10, 2025 18:34 59m 50s ranzhejiang:zhejiang/fix_runtime_dataloader_shuffle
ranzhejiang:zhejiang/fix_runtime_dataloader_shuffle
January 10, 2025 18:34 59m 50s
Update torch.norm to torch.linalg.norm and torch.linalg.vector_norm hpu-gaudi2 #1478: Pull request #6931 synchronize by loadams
January 10, 2025 15:56 59m 13s loadams/fix-torch-issues
loadams/fix-torch-issues
January 10, 2025 15:56 59m 13s
hpu-gaudi2 hpu-gaudi2 #1477: Scheduled
January 10, 2025 00:11 2h 1m 35s master
master
January 10, 2025 00:11 2h 1m 35s
Add deepseek autotp hpu-gaudi2 #1475: Pull request #6937 opened by Yejing-Lai
January 9, 2025 15:37 58m 6s Yejing-Lai:lyj/deepseek
Yejing-Lai:lyj/deepseek
January 9, 2025 15:37 58m 6s
generalize deepspeed linear and implement it for non cuda systems hpu-gaudi2 #1474: Pull request #6932 synchronize by oelayan7
January 9, 2025 06:46 58m 23s oelayan7:linear
oelayan7:linear
January 9, 2025 06:46 58m 23s
hpu-gaudi2 hpu-gaudi2 #1471: Scheduled
January 9, 2025 00:11 2h 0m 29s master
master
January 9, 2025 00:11 2h 0m 29s
Update torch.norm to torch.linalg.norm and torch.linalg.vector_norm hpu-gaudi2 #1470: Pull request #6931 synchronize by loadams
January 8, 2025 22:29 58m 39s loadams/fix-torch-issues
loadams/fix-torch-issues
January 8, 2025 22:29 58m 39s
Update torch.norm to torch.linalg.norm and torch.linalg.vector_norm hpu-gaudi2 #1469: Pull request #6931 synchronize by loadams
January 8, 2025 22:29 32s loadams/fix-torch-issues
loadams/fix-torch-issues
January 8, 2025 22:29 32s
Enabled configurable auto Tensor Parallelism (TP) for the inference of diverse models hpu-gaudi2 #1468: Pull request #6553 synchronize by loadams
January 8, 2025 19:14 Action required gyou2021:configurable_autoTP
gyou2021:configurable_autoTP
January 8, 2025 19:14 Action required
Update torch.norm to torch.linalg.norm and torch.linalg.vector_norm hpu-gaudi2 #1467: Pull request #6931 synchronize by loadams
January 8, 2025 18:59 57m 22s loadams/fix-torch-issues
loadams/fix-torch-issues
January 8, 2025 18:59 57m 22s
Update torch.norm to torch.linalg.norm and torch.linalg.vector_norm hpu-gaudi2 #1466: Pull request #6931 synchronize by loadams
January 8, 2025 18:32 26m 38s loadams/fix-torch-issues
loadams/fix-torch-issues
January 8, 2025 18:32 26m 38s
Fix building on Windows with presence of Triton hpu-gaudi2 #1465: Pull request #6749 synchronize by loadams
January 8, 2025 16:46 56m 33s woct0rdho:triton-windows
woct0rdho:triton-windows
January 8, 2025 16:46 56m 33s
generalize deepspeed linear and implement it for non cuda systems hpu-gaudi2 #1464: Pull request #6932 opened by oelayan7
January 8, 2025 09:22 59m 5s oelayan7:linear
oelayan7:linear
January 8, 2025 09:22 59m 5s
Support pure meta model lm_head tp hpu-gaudi2 #1463: Pull request #6812 synchronize by loadams
January 8, 2025 05:14 56m 44s Yejing-Lai:lyj/lm_head_replace
Yejing-Lai:lyj/lm_head_replace
January 8, 2025 05:14 56m 44s
hpu-gaudi2 hpu-gaudi2 #1461: Scheduled
January 8, 2025 00:11 1h 59m 21s master
master
January 8, 2025 00:11 1h 59m 21s
inference: remove unused _validate_args function hpu-gaudi2 #1460: Pull request #5505 synchronize by loadams
January 7, 2025 16:42 59m 0s nelyahu:remove_validate_args
nelyahu:remove_validate_args
January 7, 2025 16:42 59m 0s
[inf] Add config var to enable keeping module on host hpu-gaudi2 #1459: Pull request #6846 synchronize by oelayan7
January 7, 2025 09:27 57m 46s oelayan7:keep_module_on_host
oelayan7:keep_module_on_host
January 7, 2025 09:27 57m 46s
Support pure meta model lm_head tp hpu-gaudi2 #1457: Pull request #6812 synchronize by loadams
January 7, 2025 01:26 6h 32m 44s Yejing-Lai:lyj/lm_head_replace
Yejing-Lai:lyj/lm_head_replace
January 7, 2025 01:26 6h 32m 44s
hpu-gaudi2 hpu-gaudi2 #1456: Scheduled
January 7, 2025 00:11 6h 50m 43s master
master
January 7, 2025 00:11 6h 50m 43s
Autotp training hpu-gaudi2 #1455: Pull request #6922 synchronize by loadams
January 6, 2025 22:33 58m 0s inkcherry:autotp_training
inkcherry:autotp_training
January 6, 2025 22:33 58m 0s
Autotp training hpu-gaudi2 #1453: Pull request #6922 synchronize by tjruwase
January 6, 2025 17:26 1h 52m 53s inkcherry:autotp_training
inkcherry:autotp_training
January 6, 2025 17:26 1h 52m 53s
Set shuffle=True by default in data_sampler hpu-gaudi2 #1452: Pull request #6531 synchronize by loadams
January 6, 2025 16:56 1h 55m 27s ranzhejiang:zhejiang/fix_runtime_dataloader_shuffle
ranzhejiang:zhejiang/fix_runtime_dataloader_shuffle
January 6, 2025 16:56 1h 55m 27s
Reduce the device bubble introduced by heavy loop synchronization in coalesced fetch/release(z3_leaf_module) hpu-gaudi2 #1451: Pull request #6694 synchronize by loadams
January 6, 2025 16:47 1h 19m 30s inkcherry:reduce_coalesced_fetch_bubble
inkcherry:reduce_coalesced_fetch_bubble
January 6, 2025 16:47 1h 19m 30s
Use ds-specific module id to avoid conflicts hpu-gaudi2 #1450: Pull request #6847 synchronize by loadams
January 6, 2025 16:12 59m 34s olruwase/pr_6772
olruwase/pr_6772
January 6, 2025 16:12 59m 34s