strisunshinewentingwang/pytorch

Pull Requests Commits

[DCP] Fix test_file_system_checkpoint.py and test_file_system_checkpoint_cpu.py (#94069)

wz337 committed 1 year ago

16387bee

[decomp] Decompose std/std_mean into aten.var/var_mean (#94072)

peterbell10 committed 1 year ago

819990f5

Optimize ConvTransposed2D with mkldnn float32 and bfloat16 on CPU (#92530)

mingfeima committed 1 year ago

26cba842

enable bf16 emb (#94163)

zhuhaozhe committed 1 year ago

f3bf46e8

fix inductor clamp decomp to correctly type promote and avoid wrappin… (#94157)

ngimel committed 1 year ago

ea4cda52

Support neg calls to dyn shapes (#94068)

voznesenskym committed 1 year ago

9350bcf6

Add missing move to torch_dispatch_mode.h (#94154)

Skylion007 committed 1 year ago

7b6e9488

[MPS] Fix `cumsum` for negative indexes (#94119)

malfet committed 1 year ago

10a1efb4

Small refactor of shape guards to allow for 1:1 code_parts (#93894)

voznesenskym committed 1 year ago

60a3b742

[MPS] Fix `index_select` for empty input (#94117)

malfet committed 1 year ago

8a88852d

fix upsampling decompositions to have integer output sizes (#94123)

ngimel committed 1 year ago

8ecda196

[Dynamo] Put torch.cuda.stream into Dynamo FX graph (#93808)

yanboliang committed 1 year ago

2362b5fc

dont graph break on list[SymInt] comparisons (#94054)

voznesenskym committed 1 year ago

25c0737a

Report graph breaks separately from graph count (#94143)

ezyang committed 1 year ago

1d53123f

Add graphs/ops to parse_logs.py (#94138)

ezyang committed 1 year ago

a2db70b3

To vectorize long datatype as mask index (#91076)

EikanWang committed 1 year ago

9895c19a

Hack SymInt.__iadd__ to be working. (#94136)

ezyang committed 1 year ago

834e8f04

Update dynamic benchmark skips (#94114)

ezyang committed 1 year ago

c1da35af

perf: fix missing noexcepts on minpybind in functorch (#94135)

Skylion007 committed 1 year ago

3693039b

[c10d] Update get_backend() in exception_handler (#94063)

wz337 committed 1 year ago

f54fd6fb

Add lowerings for all symbolic shape operators (#94121)

ezyang committed 1 year ago

8c26ed5f

Simplify OpenMP detection in CMake (#91576)

cyyever committed 1 year ago

afd7b581

tools: Add lint for CONSTEXPR (#94089)

seemethere committed 1 year ago

d4a93ead

Fix Win+CUDA builds using VS2017 (#94091)

malfet committed 1 year ago

996cc1c0

Clean-up removed TH from BUCK (#94022)

kit1980 committed 1 year ago

2064fa9f

Revert "trymerge to ignore certain failures (#91134)"

pytorchmergebot committed 1 year ago

7fb2ac2b

Enable Python dispatcher on inference-only aot_dispatch_base (#94118)

ezyang committed 1 year ago

170a3e02

`FusedAdam(W)` should take `OptState` into account before unscaling grads (#94060)

crcrpar committed 1 year ago

4207d3c3

[dynamo 3.11] update instruction sizes (#93984)

williamwen42 committed 1 year ago

adde6fd2

[inductor] fix cpu implement of torch.neg (#94035)

Valentine233 committed 1 year ago

11de3994

Newer Older