strisunshinewentingwang/pytorch

Pull Requests Commits

[Re-open 90267] [inductor] weight prepack for single conv_transpose2d (#91956)

chunyuan-w committed 1 year ago

f1030dcc

Use symbolic tracing_mode for aot repro with dynamic_shapes (#93393)

ezyang committed 1 year ago

66fd99cc

use aten parallel on lu factor (#93037)

zhuhaozhe committed 1 year ago

298075e1

cherry-picking autodiff support for gather/index_select (#93333)

jjsjann123 committed 1 year ago

bdca5fcd

_sparse_coo_tensor_with_dims_and_tensors backward: simplify and optimize (#91704)

nikitaved committed 1 year ago

b484d17c

[jit] jit._drop fun modifier to allow in jit class non-jit decl funs (#93012)

IvanKobzarev committed 1 year ago

6a2838ee

sparse_mask: extend lhs to sparse COO tensors (#92248)

nikitaved committed 1 year ago

994f85d6

Introduce core_aten_decompositions (#93131)

SherlockNoMad committed 1 year ago

6a7d6cc3

[Quant] X86 qengine always uses fbgemm kernels on OS other than Linux (#93218)

Xia-Weiwen committed 1 year ago

f77f88fb

Fix test_file_system_checkpoint_cpu.py temp directory usage (#93302)

huydhn committed 1 year ago

776079b5

[Quant][ONEDNN] Fix weight reorder issue for grouped convolution (#91934)

Xia-Weiwen committed 1 year ago

eea752f8

[Dynamo][Easy] Remove duplicated code in builder.py (#93809)

yanboliang committed 1 year ago

2457d0ef

[jit][await] Apply review comments (#93284)

IvanKobzarev committed 1 year ago

9daca46d

Partial revert of autogen view_copy ops which return lists (#93411)

JacobSzwejbka committed 1 year ago

feb6c9ae

[ONNX] Fix Gather replacement in RNN peephole (#93120)

titaiwangms committed 1 year ago

9d1263a8

[inductor] Don't skip realize heuristics with dynamic shapes (#93814)

jansel committed 1 year ago

2cd8cb02

Refactor dynamo distributed test helpers to be reusable (#93187)

wconstab committed 1 year ago

ac791bdd

[dtensor][6/N] change to a better/safer op registration (#90735)

wanchaol committed 1 year ago

60e503d4

Inductor cpp wrapper: cache the loading of the kernel (#89742)

chunyuan-w committed 1 year ago

42633cf5

[dtensor][5/N] add cached propagator for TP (#90734)

wanchaol committed 1 year ago

9a56997f

[dtensor][4/N] refactor dispatching logic and add propagator (#90733)

wanchaol committed 1 year ago

b0722451

[Reland] Add sym_size/stride/numel/storage_offset to native_function.yaml (#91… (#92402)

SherlockNoMad committed 1 year ago

965f4ea3

[vision hash update] update the pinned vision hash (#93323)

pytorchmergebot committed 1 year ago

79db5bcc

Re-enable xla workflow (#93334)

yeounoh committed 1 year ago

e752ec6d

Make dynamo tests work under pytest (#93251)

jansel committed 1 year ago

10910758

Configurable repro_tolerance for same_two_models (#93398)

ezyang committed 1 year ago

08041c52

Typofix (#93402)

ezyang committed 1 year ago

3bae5484

[Quant][FX] Lower QConvAddReLU2d for onednn backend (#91155)

leslie-fang-intel committed 1 year ago

0f802eed

[Quant] Add fused ConvAddReLU2d module for onednn backend (#91154)

leslie-fang-intel committed 1 year ago

e77f28a0

[Quant][FX] Lower QConvAdd2d for onednn backend (#91153)

leslie-fang-intel committed 1 year ago

ef4118e4

Newer Older