strisunshinewentingwang/pytorch

Pull Requests Commits

add tags for inplace view ops in native_functions.yaml (#65412)

bdhirsh committed 2 years ago

b0a8ca2c

add slice/select/diagonal_scatter variants as primitive ops (#64430)

bdhirsh committed 2 years ago

03f3a033

move some codegen utilities into utils.py (#63094)

bdhirsh committed 2 years ago

665c148e

Back out "Make fb::sigrid_hash_compute_multipler_shift return a std::tuple<int64_t, int64_t>" (#67456)

Mike Iovine committed 2 years ago

b100a9ea

[quant][graphmode][fx][test] Refactor test code for quant-fx2trt unit tests (#67070)

jerryzh168 committed 2 years ago

a8f85300

Add FSDP tests to verify forward overlap and memory usage (#67117)

zhaojuanmao committed 2 years ago

325b1503

Remove process group barrier and all_reduce function calls from tensorpipe agent (#65946)

H-Huang committed 2 years ago

938afa37

Disable linux-xenial-cuda10.2 config (#67344)

malfet committed 2 years ago

0c93c8e3

Document `torch.jit.is_tracing()` (#67326)

kmaehashi committed 2 years ago

6ed68f3f

Fix deadlock when forward and backward AD are used at the same time (#67360)

albanD committed 2 years ago

b27b1ff8

Fix indentation in forward_grad.h (#67359)

albanD committed 2 years ago

d3f03af4

[fbcode] Fix operator_benchmark with jit mode (#67382)

dracifer committed 2 years ago

6900aacf

Add test owners for elastic tests (#67293)

janeyx99 committed 2 years ago

eb8b80b7

[LT] Add ir_util for ComputePostOrder (#67282)

desertfire committed 2 years ago

23669480

update coverage ignore to not skip whole modules (#67395)

albanD committed 2 years ago

6293e0ad

[ONNX] Relax check on Prim::PythonOp nodes for ONNX_FALLTHROUGH (#66172) (#67273)

shubhambhokare1 committed 2 years ago

961fd76a

[ONNX] Support conv-bn fusion in blocks (#66152) (#67272)

BowenBao committed 2 years ago

02a78bdb

[ONNX] Use Reciprocal operator instead of Div(1, x). (#65382) (#67271)

garymm committed 2 years ago

9deb6027

fixed type checking errors in fuse.py (#66799)

Onyiee committed 2 years ago

eea20bfa

[SR] NNC out variant for aten::where (#67255)

Mike Iovine committed 2 years ago

7da9c4ed

[quant][embedding qat][bugfix] Fix and test QAT EmbeddingBag from_float error message (#66989)

b-koopman committed 2 years ago

3aadff65

Replace issue templates with new issue forms (#65917)

NicolasHug committed 2 years ago

62feadd7

[Static Runtime][DI] Fuse list unpack and variadic_grouped_accessor_op (#66585)

Mike Iovine committed 2 years ago

6827d36c

specializeGradSumToSize patch - propagate profile_none through profile_ivalue (#63941)

jjsjann123 committed 2 years ago

90b722c5

[sharded_tensor] easier initialization for Shard (#66351)

wanchaol committed 2 years ago

fc664ac2

[sharded_tensor] simplify init_from_local_shards API (#64481)

wanchaol committed 2 years ago

71a67d0c

[quant][graphmode][fx] Add input_idx_to_dtype and ouptut_idx_to_dtype to backend_config_dict (#67067)

jerryzh168 committed 2 years ago

0117ada4

[iOS][CoreML] Remove shape information from TensorSpec (#67412)

xta0 committed 2 years ago

e332d802

[Core ML] Assign Core ML computationUnit to executor (#67411)

xta0 committed 2 years ago

04aba42e

[Core ML] Fix error messages (#67410)

xta0 committed 2 years ago

7e1a53cd

Newer Older