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
Thread tanh logit softcapping through FlashAttention (FA2, opt-in FA3) Documentation #19476: Pull request #3391 synchronize by nvegesna-netizen
In progress nvegesna-netizen:nvegesna/gemma2-softcap-core
nvegesna-netizen:nvegesna/gemma2-softcap-core
In progress
Thread tanh logit softcapping through FlashAttention (FA2, opt-in FA3) Documentation #19475: Pull request #3391 synchronize by nvegesna-netizen
47s nvegesna-netizen:nvegesna/gemma2-softcap-core
nvegesna-netizen:nvegesna/gemma2-softcap-core
47s
Thread tanh logit softcapping through FlashAttention (FA2, opt-in FA3) Documentation #19474: Pull request #3391 synchronize by pre-commit-ci Bot
2m 32s nvegesna-netizen:nvegesna/gemma2-softcap-core
nvegesna-netizen:nvegesna/gemma2-softcap-core
2m 32s
Thread tanh logit softcapping through FlashAttention (FA2, opt-in FA3) Documentation #19473: Pull request #3391 synchronize by nvegesna-netizen
1m 10s nvegesna-netizen:nvegesna/gemma2-softcap-core
nvegesna-netizen:nvegesna/gemma2-softcap-core
1m 10s
fix: scan-ring env var may leak across tests on pytest.skip Documentation #19472: Pull request #3388 synchronize by pre-commit-ci Bot
Action required andrewwhitecdw:codequality/test-distributed-fused-attn-scan-ring-env-var-may-leak-across
andrewwhitecdw:codequality/test-distributed-fused-attn-scan-ring-env-var-may-leak-across
Action required
fix: duplicate pad_between_seqs skips and wrong architecture direction Documentation #19471: Pull request #3390 opened by andrewwhitecdw
Action required andrewwhitecdw:bugfix/test-attention-with-cp-duplicate-pad-between-seqs-skips-and
andrewwhitecdw:bugfix/test-attention-with-cp-duplicate-pad-between-seqs-skips-and
Action required
fix: guard softmax_offset.grad.zero_() by is_training Documentation #19470: Pull request #3389 opened by andrewwhitecdw
Action required andrewwhitecdw:bugfix/run-attention-with-cp-guard-softmax-offset-grad-zero-by-is
andrewwhitecdw:bugfix/run-attention-with-cp-guard-softmax-offset-grad-zero-by-is
Action required
fix: scan-ring env var may leak across tests on pytest.skip Documentation #19469: Pull request #3388 opened by andrewwhitecdw
Action required andrewwhitecdw:codequality/test-distributed-fused-attn-scan-ring-env-var-may-leak-across
andrewwhitecdw:codequality/test-distributed-fused-attn-scan-ring-env-var-may-leak-across
Action required
fix: reorder test uses original shape after swapping SBHD axes Documentation #19468: Pull request #3387 opened by andrewwhitecdw
Action required andrewwhitecdw:bugfix/test-distributed-fused-attn-reorder-test-uses-original-shape-after
andrewwhitecdw:bugfix/test-distributed-fused-attn-reorder-test-uses-original-shape-after
Action required
Change VERSION to 2.20.0.dev0 Documentation #19467: Pull request #3386 opened by fheinecke
1m 32s fheinecke:fred/version-2.20.0.dev0
fheinecke:fred/version-2.20.0.dev0
1m 32s
[Common] Improved performance of Group MXFP8 kernels Documentation #19466: Pull request #3337 synchronize by Oleg-Goncharov
1m 34s Oleg-Goncharov:pr_mxfp8_group_kernels_optimization
Oleg-Goncharov:pr_mxfp8_group_kernels_optimization
1m 34s
Fix duplicate FP8 state updates during activation checkpoint recomputation Documentation #19465: Pull request #3213 synchronize by pre-commit-ci Bot
Action required AlbertYang514:fix/checkpoint-fp8-update-bookkeeping-main
AlbertYang514:fix/checkpoint-fp8-update-bookkeeping-main
Action required
[Common] Improved performance of Group MXFP8 kernels Documentation #19464: Pull request #3337 synchronize by Oleg-Goncharov
1m 55s Oleg-Goncharov:pr_mxfp8_group_kernels_optimization
Oleg-Goncharov:pr_mxfp8_group_kernels_optimization
1m 55s
[PyTorch] Fix MXFP8 master weight cast when a rank owns no shard. Documentation #19463: Pull request #3348 synchronize by rapatel
1m 33s rapatel:ripatel/handle_empty_shards
rapatel:ripatel/handle_empty_shards
1m 33s
[PyTorch] Fix MXFP8 master weight cast when a rank owns no shard. Documentation #19462: Pull request #3348 synchronize by rapatel
1m 29s rapatel:ripatel/handle_empty_shards
rapatel:ripatel/handle_empty_shards
1m 29s
[Pytorch] MOE Sequential Block Documentation #19461: Pull request #3350 synchronize by vthumbe1503
1m 25s vthumbe1503:dispatch_combine_op
vthumbe1503:dispatch_combine_op
1m 25s
[Pytorch] MOE Sequential Block Documentation #19460: Pull request #3350 synchronize by pre-commit-ci Bot
1m 35s vthumbe1503:dispatch_combine_op
vthumbe1503:dispatch_combine_op
1m 35s
[Pytorch] MOE Sequential Block Documentation #19459: Pull request #3350 synchronize by vthumbe1503
46s vthumbe1503:dispatch_combine_op
vthumbe1503:dispatch_combine_op
46s
[Pytorch] MOE Sequential Block Documentation #19458: Pull request #3350 synchronize by vthumbe1503
1m 11s vthumbe1503:dispatch_combine_op
vthumbe1503:dispatch_combine_op
1m 11s
Fix duplicate FP8 state updates during activation checkpoint recomputation Documentation #19457: Pull request #3213 synchronize by AlbertYang514
Action required AlbertYang514:fix/checkpoint-fp8-update-bookkeeping-main
AlbertYang514:fix/checkpoint-fp8-update-bookkeeping-main
Action required
[PyTorch] Document heterogeneous quantization recipes Documentation #19456: Pull request #3336 synchronize by negvet
1m 32s negvet:fine_grained_recipe_docs
negvet:fine_grained_recipe_docs
1m 32s
[PyTorch][torch.compile] Support for DotProductAttention on flash and unfused backends Documentation #19455: Pull request #3286 synchronize by pggPL
1m 35s pggPL:dpa_torch_compile
pggPL:dpa_torch_compile
1m 35s
[PyTorch] Share MXFP8 data tensor between rowwise and columnwise for 2D quantization Documentation #19454: Pull request #3385 synchronize by TangChangcheng
Action required TangChangcheng:mxfp8-2d-share-data
TangChangcheng:mxfp8-2d-share-data
Action required
[PyTorch] Share MXFP8 data tensor between rowwise and columnwise for 2D quantization Documentation #19453: Pull request #3385 opened by TangChangcheng
Action required TangChangcheng:mxfp8-2d-share-data
TangChangcheng:mxfp8-2d-share-data
Action required
fix: add missing NVTE_BHSD_BHSD_BHSD to layout to_string Documentation #19452: Pull request #3361 reopened by andrewwhitecdw
Action required andrewwhitecdw:bugfix/fused-attn-add-missing-nvte-bhsd-bhsd-bhsd-to
andrewwhitecdw:bugfix/fused-attn-add-missing-nvte-bhsd-bhsd-bhsd-to
Action required