strisunshinewentingwang/pytorch

Pull Requests Commits

Simple Custom Operator API, V0 (#98440)

zou3519 committed 1 year ago

0157b2d7

make ATen/native/cuda/LegacyThrustHelpers.cu data_ptr-correct (#99273)

mikey dagitses committed 1 year ago

503104ce

make gemv calls data_ptr_correct (#99161)

mikey dagitses committed 1 year ago

e9fef4a7

make ATen/native/cuda/Blas.cpp data_ptr-correct (#99274)

mikey dagitses committed 1 year ago

46c0912c

[SPMD] Implement split_fused_optimizer to split one fused_optimizer node to two (#98784)

fegin committed 1 year ago

148d4926

make vol2col data_ptr-correct (#99152)

mikey dagitses committed 1 year ago

2fc7f984

make radix_sort_pairs data_ptr-correct (#99153)

mikey dagitses committed 1 year ago

ecf4400b

make remaining gemm calls data_ptr-correct (#99156)

mikey dagitses committed 1 year ago

98b62f7c

make ATen/native/cuda/AdaptiveMaxPooling2d.cu data_ptr-correct (#99164)

mikey dagitses committed 1 year ago

306594b2

make ATen/native/cuda/SegmentReduce.cu data_ptr-correct (#99163)

mikey dagitses committed 1 year ago

314cba96

make ATen/native/cuda/Unique.cu data_ptr-correct (#99240)

mikey dagitses committed 1 year ago

777a666a

make ATen/native/cuda/MultinomialKernel.cu data_ptr-correct (#99241)

mikey dagitses committed 1 year ago

31393ea4

fix mul/div overflow issue on CPU float16 (#98820)

mingfeima committed 1 year ago

a8429342

[pt2] add meta function for `linalg.cross` (#99279)

nkaretnikov committed 1 year ago

efc3887e

[pt2] add `SymInt` support for `column_stack` (#99276)

nkaretnikov committed 1 year ago

775dd869

Rename default branch to `main` (#99210)

malfet committed 1 year ago

Verified 2418b945

[PT2E][Quantization] Refactor Quantizer and QNNPACKQuantizer (#99063)

kimishpatel committed 1 year ago

31f311a8

fix fake tensor propagation for cross_entropy with smoothing (#99255)

ngimel committed 1 year ago

888c65b6

simplify codegen for integer min/max since they don't need to propaga… (#99249)

ngimel committed 1 year ago

fa502ab0

Remove obsolete HAS_PRIMS_REFS (#99252)

jansel committed 1 year ago

9ab5fdff

Revert "[quant][fix] Compare resnet with quantizer api with the prepare_fx and decomposed convert flow (#98905)"

pytorchmergebot committed 1 year ago

20a17881

make untemplated gemm calls data_ptr-correct (#99184)

mikey dagitses committed 1 year ago

be0b12ec

make ATen/native/cuda/Embedding.cu data_ptr-correct (#99183)

mikey dagitses committed 1 year ago

29ff5a0c

Add parameter for pin memory of storage to support other devices. (#98692)

wbigat committed 1 year ago

b08c3841

Revert "Expandable blocks in allocator (#96995)"

pytorchmergebot committed 1 year ago

851e89c8

[vmap] Register decomposition for huber_loss_backward (#99236)

qqaatw committed 1 year ago

6f181aae

Further improve symbolic shapes logging (#99159)

ezyang committed 1 year ago

e2923b52

Functionalization of torch.rand/rand_like ops (#97377)

anijain2305 committed 1 year ago

fdbc8625

[inductor] Refactor pre-grad passes into inductor.fx_passes (#99130)

jansel committed 1 year ago

6e1e27fc

[inductor][quant]Enable inductor vec codegen for quantization (#98489)

leslie-fang-intel committed 1 year ago

91279f94

Newer Older