Skip to content

Actions: NVIDIA/TransformerEngine

Actions

Documentation

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
docs.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
[PyTorch] Enable head dim 256 for FA4 Documentation #17017: Pull request #2932 synchronize by yaox12
1m 30s yaox12:xiny/headdim256_fa
yaox12:xiny/headdim256_fa
1m 30s
Implement 4over6 NVFP4 recipe Documentation #17016: Pull request #2972 synchronize by ziang-and
Action required zianglih:4over6
zianglih:4over6
Action required
Implement 4over6 NVFP4 recipe Documentation #17015: Pull request #2972 synchronize by ziang-and
Action required zianglih:4over6
zianglih:4over6
Action required
[Common/PyTorch/JAX] make offset of ClampedSwiGLU configurable Documentation #17014: Pull request #2938 synchronize by hxbai
1m 19s hxbai:swiglu_offset
hxbai:swiglu_offset
1m 19s
Implement 4over6 NVFP4 recipe Documentation #17013: Pull request #2972 synchronize by ziang-and
Action required zianglih:4over6
zianglih:4over6
Action required
Implement 4over6 NVFP4 recipe Documentation #17012: 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 Documentation #17011: Pull request #2977 synchronize by jinzex
1m 30s jinzex:jinzex/tp-invariant-numerics-upstream
jinzex:jinzex/tp-invariant-numerics-upstream
1m 30s
Add examples for MoE models - Mixtral in TE Documentation #17010: Pull request #2642 synchronize by faradawn
5m 7s faradawn:add-moe-example
faradawn:add-moe-example
5m 7s
[Pytorch][Bug] DCP Checkpoint Loading Fixes for FSDP2 with QuantizedModelInit Documentation #17009: Pull request #2974 synchronize by vthumbe1503
1m 27s vthumbe1503:fsdp2_dcp_laod_fix
vthumbe1503:fsdp2_dcp_laod_fix
1m 27s
Fix CUDA graph parameter grad lifetime Documentation #17008: Pull request #2937 synchronize by buptzyb
1m 20s buptzyb:fix/cudagraph-wgrad-lifetime
buptzyb:fix/cudagraph-wgrad-lifetime
1m 20s
Fix CUDA graph parameter grad lifetime Documentation #17007: Pull request #2937 synchronize by buptzyb
1m 24s buptzyb:fix/cudagraph-wgrad-lifetime
buptzyb:fix/cudagraph-wgrad-lifetime
1m 24s
[Pytorch][Bug] DCP Checkpoint Loading Fixes for FSDP2 with QuantizedModelInit Documentation #17006: Pull request #2974 synchronize by vthumbe1503
1m 40s vthumbe1503:fsdp2_dcp_laod_fix
vthumbe1503:fsdp2_dcp_laod_fix
1m 40s
[Pytorch][Bug] DCP Checkpoint Loading Fixes for FSDP2 with QuantizedModelInit Documentation #17005: 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 Documentation #17004: Pull request #2974 synchronize by vthumbe1503
1m 21s vthumbe1503:fsdp2_dcp_laod_fix
vthumbe1503:fsdp2_dcp_laod_fix
1m 21s
TP-invariant Training: bitwise-identical training across TP degrees and GPU architecture Documentation #17003: Pull request #2977 synchronize by pre-commit-ci Bot
1m 23s jinzex:jinzex/tp-invariant-numerics-upstream
jinzex:jinzex/tp-invariant-numerics-upstream
1m 23s
[JAX] Support for cuDNN-backed flex attention Documentation #17002: Pull request #2985 synchronize by pre-commit-ci Bot
1m 20s vcherepanov-nv:cudnn-score-mod-jax
vcherepanov-nv:cudnn-score-mod-jax
1m 20s
TP-invariant Training: bitwise-identical training across TP degrees and GPU architecture Documentation #17001: Pull request #2977 synchronize by jinzex
1m 33s jinzex:jinzex/tp-invariant-numerics-upstream
jinzex:jinzex/tp-invariant-numerics-upstream
1m 33s
[PyTorch] Support for cuDNN-backed flex attention Documentation #17000: Pull request #2984 synchronize by pre-commit-ci Bot
1m 38s vcherepanov-nv:cudnn-score-mod-3
vcherepanov-nv:cudnn-score-mod-3
1m 38s
[JAX] Support for cuDNN-backed flex attention Documentation #16999: Pull request #2985 opened by vcherepanov-nv
1m 25s vcherepanov-nv:cudnn-score-mod-jax
vcherepanov-nv:cudnn-score-mod-jax
1m 25s
[PyTorch] Support for cuDNN-backed flex attention Documentation #16998: Pull request #2984 opened by vcherepanov-nv
1m 7s vcherepanov-nv:cudnn-score-mod-3
vcherepanov-nv:cudnn-score-mod-3
1m 7s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Documentation #16997: Pull request #2978 synchronize by pre-commit-ci Bot
1m 50s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
1m 50s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Documentation #16996: Pull request #2978 synchronize by kainzhong
1m 8s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
1m 8s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Documentation #16995: Pull request #2978 synchronize by pre-commit-ci Bot
6m 38s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
6m 38s
[Common, PyTorch] Improve mHC to match DeepSeek's implementation Documentation #16994: Pull request #2978 synchronize by kainzhong
1m 6s kainzhong:feat/mhc_optimization1
kainzhong:feat/mhc_optimization1
1m 6s
[Pyt][Common] Enabling/Guarding sm120 support (non - attention) Documentation #16993: Pull request #2833 synchronize by KshitijLakhani
1m 25s KshitijLakhani:klakhani/maint/sm120-pyt-cpp-support
KshitijLakhani:klakhani/maint/sm120-pyt-cpp-support
1m 25s