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
Use cuDNN's deterministic dprob in the fused grouped MLP Documentation #19778: Pull request #3407 synchronize by ZhiyuLi-Nvidia
1m 40s ZhiyuLi-Nvidia:zhiyul/cudnn-deterministic-dprob
ZhiyuLi-Nvidia:zhiyul/cudnn-deterministic-dprob
1m 40s
[PyTorch] Fix Test's Reference Implementation of Clamped Swiglu Documentation #19777: Pull request #3424 opened by vthumbe1503
1m 29s vthumbe1503:clamped_swiglu_fix_v2
vthumbe1503:clamped_swiglu_fix_v2
1m 29s
Support paged stashing for GroupedLinear activations Documentation #19776: Pull request #3423 opened by lhb8125
1m 35s lhb8125:denliu/paged-stash-grouped-tensor
lhb8125:denliu/paged-stash-grouped-tensor
1m 35s
[Common] Default EP comm kernels to 32 SMs; keep shuffle/preprocess on all SMs Documentation #19775: Pull request #3422 opened by phu0ngng
1m 20s phu0ngng:ep-sm-defaults
phu0ngng:ep-sm-defaults
1m 20s
Prototype NVFP4 with FP8 UE5M3 block scales Documentation #19774: Pull request #3325 synchronize by pre-commit-ci Bot
1m 15s timmoon10:nvfp4-ue5m3-prototype
timmoon10:nvfp4-ue5m3-prototype
1m 15s
Prototype NVFP4 with FP8 UE5M3 block scales Documentation #19773: Pull request #3325 synchronize by timmoon10
1m 15s timmoon10:nvfp4-ue5m3-prototype
timmoon10:nvfp4-ue5m3-prototype
1m 15s
fix: guard backward_dw() bias-grad backfill against a legitimately-empty bgrad Documentation #19772: Pull request #3400 synchronize by nvegesna-netizen
1m 22s nvegesna-netizen:nvegesna/fix-backward-dw-bgrad-none
nvegesna-netizen:nvegesna/fix-backward-dw-bgrad-none
1m 22s
Thread tanh logit softcapping through FlashAttention (FA2, opt-in FA3) Documentation #19771: Pull request #3391 synchronize by nvegesna-netizen
1m 40s nvegesna-netizen:nvegesna/gemma2-softcap-core
nvegesna-netizen:nvegesna/gemma2-softcap-core
1m 40s
[JAX] Optimize MoE block Documentation #19770: Pull request #3354 synchronize by jberchtold-nvidia
1m 43s jberchtold-nvidia:jberchtold/moeblock-debug
jberchtold-nvidia:jberchtold/moeblock-debug
1m 43s
[JAX] Optimize MoE block Documentation #19769: Pull request #3354 synchronize by jberchtold-nvidia
1m 31s jberchtold-nvidia:jberchtold/moeblock-debug
jberchtold-nvidia:jberchtold/moeblock-debug
1m 31s
[JAX] Optimize MoE block Documentation #19768: Pull request #3354 synchronize by jberchtold-nvidia
1m 17s jberchtold-nvidia:jberchtold/moeblock-debug
jberchtold-nvidia:jberchtold/moeblock-debug
1m 17s
[JAX] Optimize MoE block Documentation #19767: Pull request #3354 synchronize by jberchtold-nvidia
1m 28s jberchtold-nvidia:jberchtold/moeblock-debug
jberchtold-nvidia:jberchtold/moeblock-debug
1m 28s
[JAX] Return max_logit from TE fused attn Documentation #19766: Pull request #3112 synchronize by pre-commit-ci Bot
1m 40s KshitijLakhani:klakhani/feat/return-softmax-stats
KshitijLakhani:klakhani/feat/return-softmax-stats
1m 40s
[JAX] Return max_logit from TE fused attn Documentation #19765: Pull request #3112 synchronize by KshitijLakhani
1m 26s KshitijLakhani:klakhani/feat/return-softmax-stats
KshitijLakhani:klakhani/feat/return-softmax-stats
1m 26s
[JAX] Optimize MoE block Documentation #19764: Pull request #3354 synchronize by pre-commit-ci Bot
1m 42s jberchtold-nvidia:jberchtold/moeblock-debug
jberchtold-nvidia:jberchtold/moeblock-debug
1m 42s
[JAX] Optimize MoE block Documentation #19763: Pull request #3354 synchronize by jberchtold-nvidia
1m 34s jberchtold-nvidia:jberchtold/moeblock-debug
jberchtold-nvidia:jberchtold/moeblock-debug
1m 34s
Fix: fused QUproj + RoPE + Quant Documentation #19762: Pull request #3421 opened by ghadiaravi13
1m 33s ghadiaravi13:rghadia/fused_mlperf_final
ghadiaravi13:rghadia/fused_mlperf_final
1m 33s
[JAX] Return max_logit from TE fused attn Documentation #19761: Pull request #3112 synchronize by pre-commit-ci Bot
1m 32s KshitijLakhani:klakhani/feat/return-softmax-stats
KshitijLakhani:klakhani/feat/return-softmax-stats
1m 32s
[JAX] Return max_logit from TE fused attn Documentation #19760: Pull request #3112 synchronize by KshitijLakhani
1m 21s KshitijLakhani:klakhani/feat/return-softmax-stats
KshitijLakhani:klakhani/feat/return-softmax-stats
1m 21s
[PyTorch] Reduce CPU overhead on eager EP dispatch and combine Documentation #19759: Pull request #3380 synchronize by phu0ngng
1m 25s phu0ngng:ep-pyt
phu0ngng:ep-pyt
1m 25s
[Pytorch] MOE Sequential Block Documentation #19758: Pull request #3350 synchronize by vthumbe1503
1m 38s vthumbe1503:dispatch_combine_op
vthumbe1503:dispatch_combine_op
1m 38s
[PyTorch] Reduce CPU overhead on eager EP dispatch and combine Documentation #19757: Pull request #3380 synchronize by phu0ngng
1m 21s phu0ngng:ep-pyt
phu0ngng:ep-pyt
1m 21s
[PyTorch] Reduce CPU overhead on eager EP dispatch and combine Documentation #19756: Pull request #3380 synchronize by phu0ngng
1m 28s phu0ngng:ep-pyt
phu0ngng:ep-pyt
1m 28s
[All] Refactor fused attention APIs with cuDNN-frontend support checks and opaque config/params handles Documentation #19755: Pull request #2964 synchronize by pre-commit-ci Bot
2m 2s cyanguwa:fe_check_support
cyanguwa:fe_check_support
2m 2s
[All] Refactor fused attention APIs with cuDNN-frontend support checks and opaque config/params handles Documentation #19754: Pull request #2964 synchronize by cyanguwa
1m 38s cyanguwa:fe_check_support
cyanguwa:fe_check_support
1m 38s