Commits

Commits on Jun 24, 2024

flatten runtime inputs with fast path (#129116)

xmfan
authored and
facebook-github-bot
committedJun 24, 2024
update benchmarks to use cli flags for fullgraph/dynamic (#127960)

xmfan
authored and
facebook-github-bot
committedJun 24, 2024

Commits on Jun 16, 2024

Silently disable default saved tensor hooks during tracing (#123196)

xmfan
authored and
facebook-github-bot
committedJun 16, 2024

Commits on May 2, 2024

Revert "Explicitly call aten.div with CUDA denominator for cudagraphs (#119729)" (#125246)

xmfan
authored and
facebook-github-bot
committedMay 2, 2024

Commits on Apr 26, 2024

add cudagraph_skips counter (#124804)

xmfan
authored and
facebook-github-bot
committedApr 26, 2024
Explicitly call aten.div with CUDA denominator for cudagraphs (#119729)

xmfan
authored and
facebook-github-bot
committedApr 26, 2024

Commits on Apr 18, 2024

Add --use-warm-peak-memory (#124326)

xmfan
authored and
facebook-github-bot
committedApr 18, 2024
Add --snapshot-memory to get memory pickles for eager vs compiled (#119411)

xmfan
authored and
facebook-github-bot
committedApr 18, 2024

Commits on Apr 17, 2024

Create PT2-D MAST sample (#2229 )

xmfan
authored and
facebook-github-bot
committedApr 17, 2024

Commits on Apr 14, 2024

Trim runtime refs for list inputs from dynamo (#122535)

xmfan
authored and
facebook-github-bot
committedApr 14, 2024
Make compiled graph take in boxed inputs (#122353)

xmfan
authored and
facebook-github-bot
committedApr 14, 2024

Commits on Apr 4, 2024

Support frontend graphs with list arguments (#123212)

xmfan
authored and
facebook-github-bot
committedApr 4, 2024

Commits on Jan 25, 2024

Remove dynamo runner's dependency on distributed build (#117903)

xmfan
authored and
facebook-github-bot
committedJan 25, 2024

Commits on Jan 15, 2024

Torchbench Dynamo Runner: Enable DDP for perf test and traces (#113332)

xmfan
authored and
facebook-github-bot
committedJan 15, 2024

Commits on Jan 12, 2024

add --compile-autograd to dynamo benchmarks (#117196)

xmfan
authored and
facebook-github-bot
committedJan 12, 2024

Commits on Oct 27, 2023

yolov3: reduce batch size due to OOM (#111959)

xmfan
authored and
facebook-github-bot
committedOct 27, 2023

Commits on Oct 26, 2023

Dynamo runner: add FSDP handcrafted module wrapping policy (#111505)

xmfan
authored and
facebook-github-bot
committedOct 26, 2023

Commits on Oct 10, 2023

Add new model: simple_gpt_tp_manual (#1969 )

xmfan
authored and
facebook-github-bot
committedOct 10, 2023

Commits on Sep 26, 2023

Fix PyTorch CI HUD dashboard missing perf numbers: hf_Whisper (#1935 )

xmfan
authored and
facebook-github-bot
committedSep 26, 2023

Commits on Sep 21, 2023

Add nanogpt_train to torchbench, and merge with nanogpt_generate into joint folder (#1911 )

xmfan
authored and
facebook-github-bot
committedSep 21, 2023

Commits on Sep 8, 2023

Add DTensor LLaMA inference model: simple_gpt (#1867 )

xmfan
authored and
facebook-github-bot
committedSep 8, 2023