Skip to content

Actions: deepspeedai/DeepSpeed

Actions

Build and publish DeepSpeed release

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
release.yml
21 workflow runs
21 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
[bugfix] _DimZeroAllToAll silently fails to send grads under torch.co… Build and publish DeepSpeed release #71: Commit f5af15c pushed by loadams
1m 6s v0.19.7
v0.19.7
1m 6s
Honor adam_w_mode in the CPU multi_tensor_adam binding (#8307) Build and publish DeepSpeed release #70: Commit c7cc64a pushed by loadams
1m 3s v0.19.6
v0.19.6
1m 3s
Unmanaged gradient accumulation: ZeRO offload support (#8225) Build and publish DeepSpeed release #69: Commit cf44300 pushed by loadams
1m 6s v0.19.5
v0.19.5
1m 6s
Update version.txt pre-0.19.4 release (#8218) Build and publish DeepSpeed release #68: Commit c455031 pushed by loadams
1m 43s v0.19.4
v0.19.4
1m 43s
Adopt PTF Code of Conduct (#8165) Build and publish DeepSpeed release #67: Commit 0c36f6d pushed by loadams
1m 1s v0.19.3
v0.19.3
1m 1s
Mixed-precision: per-policy param/buffer dtype cast (preserve fp32 bu… Build and publish DeepSpeed release #66: Commit b919284 pushed by loadams
1m 9s v0.19.2
v0.19.2
1m 9s
Fix DeepCompile AOT kwargs patching for PyTorch >= v2.11 (#8024) Build and publish DeepSpeed release #65: Commit 510ebe5 pushed by sfc-gh-truwase
1m 48s v0.19.1
v0.19.1
1m 48s
Update version.txt before 0.19.0 release (#7995) Build and publish DeepSpeed release #64: Commit 98a2fbd pushed by loadams
6s v0.19.0
v0.19.0
6s
Update version.txt for latest incoming release 0.18.9 (#7935) Build and publish DeepSpeed release #63: Commit 8c93851 pushed by loadams
58s v0.18.9
v0.18.9
58s
Update version (#7903) Build and publish DeepSpeed release #62: Commit 5f7b687 pushed by loadams
59s v0.18.8
v0.18.8
59s
fix(zero3): use current_stream() instead of default_stream() for grad… Build and publish DeepSpeed release #61: Commit be60451 pushed by loadams
48s v0.18.8
v0.18.8
48s
Fix hook count performance regression from v0.18.5 (#7886) Build and publish DeepSpeed release #60: Commit 6c59d54 pushed by loadams
1m 4s v0.18.7
v0.18.7
1m 4s
Replace torch.jit.script with torch.compile (#7835) (#7840) Build and publish DeepSpeed release #59: Commit d2ca6e7 pushed by loadams
1m 5s v0.18.6
v0.18.6
1m 5s
Update PyTorch to v2.9 for modal tests (#7816) Build and publish DeepSpeed release #58: Commit b19987c pushed by loadams
59s v0.18.5
v0.18.5
59s
fix: avoid IndexError in BF16_Optimizer.destroy() when using DummyOpt… Build and publish DeepSpeed release #57: Commit b35d9eb pushed by loadams
1m 2s v0.18.4
v0.18.4
1m 2s
Wall clock timers API (#7714) Build and publish DeepSpeed release #56: Commit 6eb98aa pushed by sfc-gh-truwase
5m 9s v0.18.3
v0.18.3
5m 9s
README refresh (#7668) Build and publish DeepSpeed release #55: Commit e993fea pushed by loadams
1m 29s v0.18.2
v0.18.2
1m 29s
Update version.txt in advance of next release Build and publish DeepSpeed release #54: Commit 3631712 pushed by loadams
1m 18s v0.18.1
v0.18.1
1m 18s
Update email address (#7624) Build and publish DeepSpeed release #53: Commit 79caae1 pushed by sfc-gh-truwase
1m 27s v0.18.0
v0.18.0
1m 27s
[bugfix] fix partition context unpatch (#7566) Build and publish DeepSpeed release #52: Commit e4f6da9 pushed by loadams
1m 8s v0.17.6
v0.17.6
1m 8s
Add index to HPU devices (#7497) Build and publish DeepSpeed release #51: Commit 047a759 pushed by sfc-gh-truwase
50m 45s v0.17.5
v0.17.5
50m 45s