Skip to content

Actions: NVIDIA/TransformerEngine

Actions

Build

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
build.yml
2,500+ workflow runs
2,500+ 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
Implement 4over6 NVFP4 recipe Build #17593: Pull request #2972 synchronize by ziang-and
Action required zianglih:4over6
zianglih:4over6
Action required
Implement 4over6 NVFP4 recipe Build #17592: Pull request #2972 synchronize by ziang-and
Action required zianglih:4over6
zianglih:4over6
Action required
[Common/PyTorch/JAX] make offset of ClampedSwiGLU configurable Build #17591: Pull request #2938 synchronize by hxbai
In progress hxbai:swiglu_offset
hxbai:swiglu_offset
In progress
Implement 4over6 NVFP4 recipe Build #17590: Pull request #2972 synchronize by ziang-and
Action required zianglih:4over6
zianglih:4over6
Action required
Implement 4over6 NVFP4 recipe Build #17589: Pull request #2972 synchronize by ziang-and
Action required zianglih:4over6
zianglih:4over6
Action required
TP-invariant Training: bitwise-identical training across TP degrees and GPU architecture Build #17588: Pull request #2977 synchronize by jinzex
2h 15m 46s jinzex:jinzex/tp-invariant-numerics-upstream
jinzex:jinzex/tp-invariant-numerics-upstream
2h 15m 46s
Add examples for MoE models - Mixtral in TE Build #17587: Pull request #2642 synchronize by faradawn
In progress faradawn:add-moe-example
faradawn:add-moe-example
In progress
[Pytorch][Bug] DCP Checkpoint Loading Fixes for FSDP2 with QuantizedModelInit Build #17586: Pull request #2974 synchronize by vthumbe1503
4h 33m 2s vthumbe1503:fsdp2_dcp_laod_fix
vthumbe1503:fsdp2_dcp_laod_fix
4h 33m 2s
Fix CUDA graph parameter grad lifetime Build #17585: Pull request #2937 synchronize by buptzyb
4h 31m 10s buptzyb:fix/cudagraph-wgrad-lifetime
buptzyb:fix/cudagraph-wgrad-lifetime
4h 31m 10s
Fix CUDA graph parameter grad lifetime Build #17584: Pull request #2937 synchronize by buptzyb
3m 55s buptzyb:fix/cudagraph-wgrad-lifetime
buptzyb:fix/cudagraph-wgrad-lifetime
3m 55s
[Pytorch][Bug] DCP Checkpoint Loading Fixes for FSDP2 with QuantizedModelInit Build #17583: Pull request #2974 synchronize by vthumbe1503
22m 3s vthumbe1503:fsdp2_dcp_laod_fix
vthumbe1503:fsdp2_dcp_laod_fix
22m 3s
[Pytorch][Bug] DCP Checkpoint Loading Fixes for FSDP2 with QuantizedModelInit Build #17582: Pull request #2974 synchronize by pre-commit-ci Bot
48s vthumbe1503:fsdp2_dcp_laod_fix
vthumbe1503:fsdp2_dcp_laod_fix
48s
[Pytorch][Bug] DCP Checkpoint Loading Fixes for FSDP2 with QuantizedModelInit Build #17581: Pull request #2974 synchronize by vthumbe1503
1m 42s vthumbe1503:fsdp2_dcp_laod_fix
vthumbe1503:fsdp2_dcp_laod_fix
1m 42s
TP-invariant Training: bitwise-identical training across TP degrees and GPU architecture Build #17580: Pull request #2977 synchronize by pre-commit-ci Bot
3h 14m 34s jinzex:jinzex/tp-invariant-numerics-upstream
jinzex:jinzex/tp-invariant-numerics-upstream
3h 14m 34s
[JAX] Support for cuDNN-backed flex attention Build #17579: Pull request #2985 synchronize by pre-commit-ci Bot
4h 27m 52s vcherepanov-nv:cudnn-score-mod-jax
vcherepanov-nv:cudnn-score-mod-jax
4h 27m 52s
TP-invariant Training: bitwise-identical training across TP degrees and GPU architecture Build #17578: Pull request #2977 synchronize by jinzex
2m 31s jinzex:jinzex/tp-invariant-numerics-upstream
jinzex:jinzex/tp-invariant-numerics-upstream
2m 31s
[PyTorch] Support for cuDNN-backed flex attention Build #17577: Pull request #2984 synchronize by pre-commit-ci Bot
4h 32m 24s vcherepanov-nv:cudnn-score-mod-3
vcherepanov-nv:cudnn-score-mod-3
4h 32m 24s
[JAX] Support for cuDNN-backed flex attention Build #17576: Pull request #2985 opened by vcherepanov-nv
1m 53s vcherepanov-nv:cudnn-score-mod-jax
vcherepanov-nv:cudnn-score-mod-jax
1m 53s
[PyTorch] Support for cuDNN-backed flex attention Build #17575: Pull request #2984 opened by vcherepanov-nv
1m 3s vcherepanov-nv:cudnn-score-mod-3
vcherepanov-nv:cudnn-score-mod-3
1m 3s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Build #17574: Pull request #2978 synchronize by pre-commit-ci Bot
2h 52m 21s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
2h 52m 21s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Build #17573: Pull request #2978 synchronize by kainzhong
55s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
55s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Build #17572: Pull request #2978 synchronize by pre-commit-ci Bot
24m 33s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
24m 33s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Build #17571: Pull request #2978 synchronize by kainzhong
50s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
50s
[Pyt][Common] Enabling/Guarding sm120 support (non - attention) Build #17570: Pull request #2833 synchronize by KshitijLakhani
4h 29m 12s KshitijLakhani:klakhani/maint/sm120-pyt-cpp-support
KshitijLakhani:klakhani/maint/sm120-pyt-cpp-support
4h 29m 12s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Build #17569: Pull request #2978 synchronize by pre-commit-ci Bot
41m 55s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
41m 55s