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] Avoid concatenation copies for compiled split Linear parameters Documentation #21058: Pull request #3623 synchronize by pggPL
1m 28s pggPL:split_parameters_linear
pggPL:split_parameters_linear
1m 28s
[Docs] Add Mixture of Experts guide Documentation #21057: Pull request #3494 synchronize by pggPL
1m 19s pggPL:docs_moe
pggPL:docs_moe
1m 19s
Fix AArch64 NCCL EP linking with BOLT enabled Documentation #21056: Pull request #3627 opened by aybchan
1m 36s aybchan:fix/ep-linker-flag
aybchan:fix/ep-linker-flag
1m 36s
[Docs] Add Mixture of Experts guide Documentation #21055: Pull request #3494 synchronize by pggPL
1m 41s pggPL:docs_moe
pggPL:docs_moe
1m 41s
[PyTorch] DeepSeekV3Layer: full MoE transformer layer (MLA + DeepSeek MoE) Documentation #21054: Pull request #3471 synchronize by pggPL
1m 35s pggPL:deepseek_v3_layer
pggPL:deepseek_v3_layer
1m 35s
Let the CUDA allocator manage cuBLAS workspaces Documentation #21053: Pull request #3625 synchronize by Vtmpas
Action required Vtmpas:fix/cublas-workspace-per-stream
Vtmpas:fix/cublas-workspace-per-stream
Action required
Let the CUDA allocator manage cuBLAS workspaces Documentation #21052: Pull request #3625 synchronize by Vtmpas
Action required Vtmpas:fix/cublas-workspace-per-stream
Vtmpas:fix/cublas-workspace-per-stream
Action required
[PyTorch] Avoid concatenation copies for compiled split Linear parameters Documentation #21051: Pull request #3623 synchronize by pggPL
1m 29s pggPL:split_parameters_linear
pggPL:split_parameters_linear
1m 29s
Let the CUDA allocator manage cuBLAS workspaces Documentation #21050: Pull request #3625 opened by Vtmpas
Action required Vtmpas:fix/cublas-workspace-per-stream
Vtmpas:fix/cublas-workspace-per-stream
Action required
Cache cuBLAS workspace by CUDA stream Documentation #21049: Pull request #3624 opened by Vtmpas
7m 57s Vtmpas:codex/cublas-workspace-per-stream
Vtmpas:codex/cublas-workspace-per-stream
7m 57s
[PyTorch] Avoid concatenation copies for compiled split Linear parameters Documentation #21048: Pull request #3623 opened by pggPL
1m 36s pggPL:split_parameters_linear
pggPL:split_parameters_linear
1m 36s
[PyTorch] Deprecate grouped linear and MLP environment gates Documentation #21047: Pull request #3584 synchronize by vthumbe1503
1m 33s zhongbozhu:cleanup/grouped-mlp-env-gates
zhongbozhu:cleanup/grouped-mlp-env-gates
1m 33s
TE MoE Return Expert Counts and Custom Routing for DSv4 Support Documentation #21046: Pull request #3596 synchronize by dinodeep
1m 19s dinodeep:dinodeep/te-moe-block-dsv4-support
dinodeep:dinodeep/te-moe-block-dsv4-support
1m 19s
feat(attention): cuDNN FROST attention backend for head_dim in (256, 512], with context parallelism Documentation #21045: Pull request #3527 synchronize by nvegesna-netizen
1m 16s nvegesna-netizen:nvegesna/te-frost-d512-cp
nvegesna-netizen:nvegesna/te-frost-d512-cp
1m 16s
feat(attention): cuDNN FROST attention backend for head_dim in (256, 512], with context parallelism Documentation #21044: Pull request #3527 synchronize by nvegesna-netizen
1m 11s nvegesna-netizen:nvegesna/te-frost-d512-cp
nvegesna-netizen:nvegesna/te-frost-d512-cp
1m 11s
[Common] Add launch bounds to multi-tensor swizzle kernels Documentation #21043: Pull request #3622 opened by ravimajeti
Action required ravimajeti:claude/issue-3621
ravimajeti:claude/issue-3621
Action required
[Common] Fall back from gated TMA kernels when shared memory does not fit Documentation #21042: Pull request #3600 synchronize by ravimajeti
Action required ravimajeti:codex/issue-3299
ravimajeti:codex/issue-3299
Action required
Prototype NVFP4 with FP8 UE5M3 block scales Documentation #21041: Pull request #3325 synchronize by timmoon10
1m 48s timmoon10:nvfp4-ue5m3-prototype
timmoon10:nvfp4-ue5m3-prototype
1m 48s
[Common] Preserve NaN through half-precision MXFP8 amax reductions Documentation #21040: Pull request #3574 synchronize by CodeAlex52
Action required CodeAlex52:fix/mxfp8-nan-amax-reduction
CodeAlex52:fix/mxfp8-nan-amax-reduction
Action required
Grouped Linear CPU Overhead Reductions By Removing Unnecessary Python Logic Documentation #21039: Pull request #3613 synchronize by zhongbozhu
1m 25s zhongbozhu:optimize/grouped-linear-cpu-overhead
zhongbozhu:optimize/grouped-linear-cpu-overhead
1m 25s
Grouped Linear CPU Overhead Reductions By Removing Unnecessary Python Logic Documentation #21038: Pull request #3613 synchronize by zhongbozhu
1m 24s zhongbozhu:optimize/grouped-linear-cpu-overhead
zhongbozhu:optimize/grouped-linear-cpu-overhead
1m 24s
[Common] Remove CuTeDSL first launch protection Documentation #21037: Pull request #3620 synchronize by kainzhong
1m 33s kainzhong:remove_cutedsl_first_launch_protection
kainzhong:remove_cutedsl_first_launch_protection
1m 33s
[Common] Add CuTeDSL grouped MXFP8 quantization Documentation #21036: Pull request #3615 synchronize by kainzhong
1m 25s kainzhong:cutedsl_mxfp8_grouped
kainzhong:cutedsl_mxfp8_grouped
1m 25s
[Common] Add CuTeDSL grouped MXFP8 quantization Documentation #21035: Pull request #3615 synchronize by kainzhong
1m 59s kainzhong:cutedsl_mxfp8_grouped
kainzhong:cutedsl_mxfp8_grouped
1m 59s
[PyTorch] Deprecate grouped linear and MLP environment gates Documentation #21034: Pull request #3584 synchronize by zhongbozhu
1m 36s zhongbozhu:cleanup/grouped-mlp-env-gates
zhongbozhu:cleanup/grouped-mlp-env-gates
1m 36s