Skip to content

Actions: vasunvidia/TransformerEngine

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
22 workflow runs
22 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[PyTorch] Fix garbage initialized permuted_scale (#2547) Deploy nightly docs #68: Commit c988548 pushed by vasunvidia
1m 21s main
main
1m 21s
[PyTorch] fix test_current_device test (#2398) Deploy nightly docs #67: Commit 41fb9bc pushed by vasunvidia
1m 43s main
main
1m 43s
Change to print softmax correction factor Deploy nightly docs #66: Commit 73eb7c6 pushed by vasunvidia
1m 53s main
main
1m 53s
[PyTorch] fix test_current_device test (#2398) Deploy nightly docs #65: Commit 41fb9bc pushed by vasunvidia
1m 47s main
main
1m 47s
[JAX] Remove import jax.extend.ffi (#2193) Deploy nightly docs #64: Commit 57b4d7b pushed by vasunvidia
3m 25s main
main
3m 25s
Fix bug when enabling --overlap-grad-reduce in mcore (#2142) Deploy nightly docs #63: Commit 11e9d66 pushed by vasunvidia
2m 5s main
main
2m 5s
Dropout with 8-bit RNG (#2014) Deploy nightly docs #62: Commit e0e3d12 pushed by vasunvidia
2m 18s main
main
2m 18s
[PyTorch][MOE] Tentative Fix For Replacing from_blob with empty for e… Deploy nightly docs #61: Commit c449c6c pushed by vasunvidia
2m 21s main
main
2m 21s
[PyTorch Debug] Skip log test on device if it does not support fp8. (… Deploy nightly docs #60: Commit 1398fa5 pushed by vasunvidia
1m 20s main
main
1m 20s
[JAX] Fix for TE GEMM - Always AllGather RHS non-contracting dims wit… Deploy nightly docs #59: Commit 3fc1e4b pushed by vasunvidia
1m 31s main
main
1m 31s
[PyTorch] Refactor C++ quantizer infrastructure (#1952) Deploy nightly docs #58: Commit cb5013b pushed by vasunvidia
2m 30s main
main
2m 30s
Update README to list JAX NGC container (#1756) Deploy nightly docs #57: Commit 320b81d pushed by vasunvidia
1m 21s main
main
1m 21s
Support FP8 primary weight in FSDP training (#1630) Deploy nightly docs #56: Commit c84d170 pushed by vasunvidia
1m 21s main
main
1m 21s
[JAX] Backward compatible Fixes (#1631) Deploy nightly docs #55: Commit 160be21 pushed by vasunvidia
1m 20s main
main
1m 20s
Update FE to 1.11 (#1580) Deploy nightly docs #54: Commit dc40f9f pushed by vasunvidia
1m 15s main
main
1m 15s
Support tensors with only column-wise data (#1505) Deploy nightly docs #53: Commit 8a20d66 pushed by vasunvidia
1m 11s main
main
1m 11s
[JAX] FFI API compatibility with both 0.4 and 0.5 (#1562) Deploy nightly docs #52: Commit 0e13788 pushed by vasunvidia
1m 24s main
main
1m 24s
[PyTorch] Remove Megatron-LM convergence test (#1521) Deploy nightly docs #51: Commit f090551 pushed by vasunvidia
1m 20s main
main
1m 20s
Add NVTX ranges to FP8 amax AR and grad output preprocessing (#1530) Deploy nightly docs #50: Commit de06a34 pushed by vasunvidia
1m 12s main
main
1m 12s
Delete extra tensor objects after restoring float8 tensors (#1500) Deploy nightly docs #49: Commit d3efaeb pushed by vasunvidia
1m 22s main
main
1m 22s
[PyTorch] Skip context parallelism tests if not enough GPUs (#1508) Deploy nightly docs #48: Commit 2834e4a pushed by vasunvidia
1m 28s main
main
1m 28s
Minor fixes for attention (#1504) Deploy nightly docs #47: Commit 8744188 pushed by vasunvidia
1m 28s main
main
1m 28s