pytorch/benchmark

Pull Requests Commits

Add GPT2ForSequenceClassification to use_larger_multiplier_for_smaller_tensor list (#160001)

generatedunixname499836121 committed 181 days ago

9e418359

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

ppanchalia committed 184 days ago

6da75d11

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

Lucaskabela committed 184 days ago

24586097

Log graph_node_shape to dynamo_compile (#160556)

jovianjaison committed 185 days ago

e11cbaa1

Make yolov3 exportable (#2630)

tugsbayasgalan committed 186 days ago

290c966f

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

ZainRizvi committed 186 days ago

136d13a9

Log stack_trace to dynamo_compile (#160348)

jovianjaison committed 187 days ago

a9216941

Bump transformers pin (#159291)

anijain2305 committed 187 days ago

13ac4f7c

Add nativert benchmark (#159922)

yiming0416 committed 192 days ago

c9ccd888

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

Lucaskabela committed 193 days ago

ac742a2b

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

Lucaskabela committed 194 days ago

5bbf4e8c

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

Lucaskabela committed 194 days ago

903b5003

gc before warming up benchmarking (#159670)

aorenste committed 196 days ago

3c500130

Enable sample nightly PT2 benchmark on B200 (#158011)

huydhn committed 198 days ago

b4dd0684

Add typing annotations to guard and source

Lucaskabela committed 200 days ago

fdea2a14

Support native Inductor as backend for MTIA

andyanwang committed 202 days ago

28b27b9e

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

anijain2305 committed 202 days ago

b377227f

Revert D78458745: Support native Inductor as backend for MTIA

ZainRizvi committed 204 days ago

bed0bfe6

Support native Inductor as backend for MTIA

andyanwang committed 205 days ago

6a1b132d

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

jamesjwu committed 205 days ago

f6902624

Skip slow tests for aarch64-inductor-benchmarks (#158842)

generatedunixname499836121 committed 206 days ago

2b283c94

Revert D78772867: Consistent behavior of int attributes (#157262)

Yashasvi Makin committed 207 days ago

4fd824e4

Consistent behavior of int attributes (#157262)

anijain2305 committed 207 days ago

1dd2203c

Grab bag of (mostly) typing improvements (#158075)

generatedunixname499836121 committed 208 days ago

033e172b

Type trace_rules.py (#158679)

Lucaskabela committed 209 days ago

e3c8aa35

add configurable prefix (#2625)

coconutruben committed 209 days ago

7ddd6181

Use proper sources for constructing dataclass defaults (#157993)

mlazos committed 212 days ago

51ed2299

add a hook for recompile user context (#157961)

BoyuanFeng committed 219 days ago

167ae57b

Fix logging of config_suppress_errors and config_inline_inbuilt_nn_modules (#157947)

Raymo111 committed 220 days ago

b251cae3

Set higher tolerance for squeezenet1_1 with bf16 (#156920)

generatedunixname499836121 committed 222 days ago

12cfa2e0

Newer Older