strisunshinewentingwang/pytorch

Pull Requests Commits

feat(optim): Add `RAdam` support for complex, with `has_complex` shortcut (#110635)

jon-chuang committed 1 year ago

347ea3fe

[export] Update ArgumentSpec definition. (#110612)

zhxchen17 committed 1 year ago

be5dc3a0

[aotinductor] Fix benchmarks with self.autocast (#110490)

angelayi committed 1 year ago

83061ee1

[ez] Remove print in heuristics aggregation (#110621)

clee2000 committed 1 year ago

8a09fe4a

Revert "Multiprocessing support for NT (#110292)"

pytorchmergebot committed 1 year ago

dac895c1

Added a UserWarning when using torch.{std,var,std_mean,std_var} with dof<=0 (#109824)

tringwald committed 1 year ago

555c83d0

Revert "pin_memory support for NT (#110404)"

pytorchmergebot committed 1 year ago

81ce5d57

[Reland2] Remove calls of c10::either (#110487)

cyyever committed 1 year ago

11b3210a

Revert "Update singleton int to error when inequality relation is undefined (#110044)"

pytorchmergebot committed 1 year ago

330db827

Revert "Support SingletonSymNode mul with coefficient (#110369)"

pytorchmergebot committed 1 year ago

1c3fae46

Revert "Update custom Function preserve torch function when inputs returned as-is (#109825)"

pytorchmergebot committed 1 year ago

236afe73

Revert "Add symbolic singleton int (#110370)"

pytorchmergebot committed 1 year ago

fdf6055e

Revert "Symintify guards.cpp (#110371)"

pytorchmergebot committed 1 year ago

585e2bd8

Revert "Use is_symbolic instead of testing isinstance in some place (#110372)"

pytorchmergebot committed 1 year ago

bcd44dac

Replace enforce_dtype with dtype in ShardedTensor.gather (#110561)

yingufan committed 1 year ago

5d963474

Add functional collective all_to_all_single and support it in Inductor (#110195)

ezyang committed 1 year ago

f274c7b3

perf(inductor): use for loop with shortcut in `Optimizer`s to speedup against list comprehensions (e.g. complex conversion) (#110613)

jon-chuang committed 1 year ago

df7d01ae

[quant][pt2e] Refactor conv related annotation for XNNPACKQuantizer (#110308)

jerryzh168 committed 1 year ago

7b604211

[BE] Get rid of code duplication (#110619)

malfet committed 1 year ago

be021037

[BE] Use nested namespaces in autograd/templates (#110618)

malfet committed 1 year ago

82e353ff

Set _diffThreshold on our TestCase (#110603)

albanD committed 1 year ago

cae53712

[BE]: Enable some basic pytest style rules (#110362)

Skylion007 committed 1 year ago

668eb554

[dtensor] add grad placements kwarg to to_local API (#110629)

wanchaol committed 1 year ago

c95cf4b4

Add option to flop counter formula registration to get raw values (#110591)

Chillee committed 1 year ago

ada65508

[torch] easy missing move in aoti_runtime/model.h (#110469)

swolchok committed 1 year ago

9e72c9cc

[dynamo, logging] Report name of defining class along side function name in Dynamo logs (#110190)

williamwen42 committed 1 year ago

71beca48

fix(optim): adagrad sparse multitensor incorrect early exit (#110454)

jon-chuang committed 1 year ago

c99de9f3

Back out "[Inductor] Break the loop fusion when node2 depends on node1 mutations (#109172)" (#110622)

ckluk2 committed 1 year ago

ecdd1bcf

[state_dict][1/N] Implement the basic functions of distributed.checkpoint._state_dict (#105902)

fegin committed 1 year ago

88616349

[aotinductor] Avoid generating redundant kernel loading code (#110510)

desertfire committed 1 year ago

298f01d9

Newer Older