pytorch/benchmark

Pull Requests Commits

[Dependabot] Update(deps): Bump transformers from 4.57.3 to 5.6.2

dependabot[bot] committed 22 days ago

Verified 1078e5fd

Support pybind enums (#180631)

anijain2305 committed 23 days ago

ca4fb30b

Improve benchmark installation speed (#2683)

benjaminglass1 committed 23 days ago

240ae9a6

Fix duplicated-word typos in comments and docstrings (#181028)

frgossen committed 24 days ago

41dc19fc

Add batch-invariant accuracy mode for benchmark perf tests (#180610)

karthickai committed 25 days ago

87e5ce35

Remove unused noqa directives in torch/, batch 1 (#180134)

aorenste committed 29 days ago

5c865e03

replace CONSTANT_VARIABLE_* usages with ConstantVariable.create (#179162)

williamwen42 committed 31 days ago

439e797f

fix guard ordering on nn modules (#170493)

williamwen42 committed 34 days ago

32ec5eea

Revert "Make adding new compile option simpler (#179195)"

generatedunixname499836121 committed 37 days ago

1201e5a7

Refactor dynamo benchmark timm models config file (#178576)

generatedunixname499836121 committed 38 days ago

147c18d9

Upgrade pyrefly 0.58 (#178793)

maggiemoss committed 39 days ago

d5d314c0

Make adding new compile option simpler (#179195)

anijain2305 committed 39 days ago

ea94b78d

add unbacked perf testing to indcutor periodic (#177034)

laithsakka committed 45 days ago

aebe8bb9

Refactor exception handling in torch._dynamo with raise_python_observed_exception (#178619)

generatedunixname499836121 committed 46 days ago

c08776a5

Fix incorrect remediation instructions in cudagraph pending backward warning (#176865)

generatedunixname499836121 committed 49 days ago

854f135f

Dynamo/fix ignore logging functions (#178506)

generatedunixname499836121 committed 49 days ago

f6325186

Add deterministic mode for benchmark perf tests (#178233)

karthickai committed 51 days ago

a16842fc

dynamo: support next() on itertools.count inputs (#177876)

bobrenjc93 committed 51 days ago

b8d35ba5

Deprecate inline_inbuilt_nn_modules config (#178205)

anijain2305 committed 52 days ago

5c74cbd0

Use proper grad mode during model validation to avoid OOM (#168176)

generatedunixname499836121 committed 52 days ago

82cd8520

Delete tests that explicitly set inline_inbuilt_nn_modules=False (#177979)

anijain2305 committed 55 days ago

a233587d

Support id() on containers and copy.deepcopy tracing (#177484)

anijain2305 committed 56 days ago

f07f3d9d

Revert "Support id() on containers and copy.deepcopy tracing (#177484)"

generatedunixname499836121 committed 59 days ago

6353eb2b

Remove unused variable ver_regex parameter usage at...

Sameer Khanna committed 59 days ago

5d2e9719

Back out "set cuBLASLt as a default BLAS backend when available (#174594)" (#2674)

atalman committed 59 days ago

7f5bfbc7

Re-land: symbolic shapes guarding_hint_or_throw and optimization_hint (#175497) (#2672)

laithsakka committed 59 days ago

2ce71fad

Replace non-idiomatic None comparison in pytorch/benchmark/bisection....

Sameer Khanna committed 60 days ago

a5f916df

enable torch._functorch.config.guess_tangent_strides_as_outputs for OSS by default

IvanKobzarev committed 60 days ago

8bed41fb

Support id() on containers and copy.deepcopy tracing (#177484)

anijain2305 committed 61 days ago

5c2b0bbf

Fix GuardOnDataDependentSymNode UserError to preserve original exception (#176016)

generatedunixname499836121 committed 63 days ago

97ff2dd1

Older