pytorch/benchmark

Pull Requests Commits

Restore TorchBench PR tests (#2636)

huydhn committed 117 days ago

979cb9fb

Add torchscript jit.trace to benchmark option (#161223)

yiming0416 committed 118 days ago

9644652f

Log exception_stack_trace to dynamo_compile (#161096)

jovianjaison committed 120 days ago

78745baf

Revamp TorchBench nightly image build process (#2635)

huydhn committed 121 days ago

b0efcbe2

Refactor convert_frame._compile_inner to return compiled bytecode + output graph. [3/n] (#160855)

zhxchen17 committed 122 days ago

abd8dd23

Add caching for inside torch.compile.disable function to avoid unnecessary recompilation. (#160934)

generatedunixname499836121 committed 122 days ago

3e377c41

Type coverage for symbolic_convert (#160922)

Lucaskabela committed 122 days ago

73eb12ed

Cleanup unused TorchBench workflows (#2634)

huydhn committed 123 days ago

9a4b981e

Revert "Fix the benchmark CI failure due to Opacus 1.5.4 (#2621)" (#2632)

huydhn committed 123 days ago

22bc29b4

pack linear for FP32 dynamic mode (#157542)

generatedunixname499836121 committed 124 days ago

af93b8be

Add GPT2ForSequenceClassification to use_larger_multiplier_for_smaller_tensor list (#160001)

generatedunixname499836121 committed 125 days ago

9e418359

Add logging for when inbuilt_inline_nn_modules will help with ID_MATCH guard triggered recompiles (#160592)

ppanchalia committed 127 days ago

6da75d11

Type improvements in `_dynamo/utils` to generics (#159824)

Lucaskabela committed 128 days ago

24586097

Log graph_node_shape to dynamo_compile (#160556)

jovianjaison committed 128 days ago

e11cbaa1

Make yolov3 exportable (#2630)

tugsbayasgalan committed 129 days ago

290c966f

Replace `pytorch-labs` with `meta-pytorch` (#2629)

ZainRizvi committed 130 days ago

136d13a9

Log stack_trace to dynamo_compile (#160348)

jovianjaison committed 130 days ago

a9216941

Bump transformers pin (#159291)

anijain2305 committed 131 days ago

13ac4f7c

Add nativert benchmark (#159922)

yiming0416 committed 135 days ago

c9ccd888

Typing `torch/_dynamo/guards.py` (#159315)

Lucaskabela committed 136 days ago

ac742a2b

Type annotation for `torch/_dynamo/output_graph.py` (#159602)

Lucaskabela committed 138 days ago

5bbf4e8c

Type coverage for `torch/_dynamo/utils.py` (#159580)

Lucaskabela committed 138 days ago

903b5003

gc before warming up benchmarking (#159670)

aorenste committed 139 days ago

3c500130

Enable sample nightly PT2 benchmark on B200 (#158011)

huydhn committed 141 days ago

b4dd0684

Add typing annotations to guard and source

Lucaskabela committed 144 days ago

fdea2a14

Support native Inductor as backend for MTIA

andyanwang committed 145 days ago

28b27b9e

Set model name early to keep warmup and main model same (#159231)

anijain2305 committed 145 days ago

b377227f

Revert D78458745: Support native Inductor as backend for MTIA

ZainRizvi committed 148 days ago

bed0bfe6

Support native Inductor as backend for MTIA

andyanwang committed 148 days ago

6a1b132d

Various small bugfixes, add CachingPrecompile to torchbench (#158847)

jamesjwu committed 148 days ago

f6902624

Older