strisunshinewentingwang/pytorch

Pull Requests Commits

remove dynamo fake param/buf check (#103574)

unlimblue committed 1 year ago

df814484

[memory_viz] fix javascript url (#103741)

zdevito committed 1 year ago

ae78e801

allow cpu scalar to be moved to xpu in masked_fill (#103645)

guangyey committed 1 year ago

ad4ee297

[ONNX] Support aten::hstack and aten::vstack (#102872)

titaiwangms committed 1 year ago

d3971f2d

[export] Make pass base composable (#103701)

angelayi committed 1 year ago

f889c886

[ONNX] Support aten::atleast_1d and aten::atleast_2d and aten::atleast_3d (#103061)

titaiwangms committed 1 year ago

0411fc6a

[Reland][Dynamo] VariableTracker.recursively_contains should be updated correctly when mutation happens (#103564) (#103717)

yanboliang committed 1 year ago

703875e3

inductor: make the vec_transpose's tiling stride doesn't depend on out_idx and tiling_idex (#103651)

XiaobingSuper committed 1 year ago

b287cb81

[memory_viz] Unified viewer (#103565)

zdevito committed 1 year ago

19b3e07f

[memory_viz] profile_plot generates snapshot objects (#103497)

zdevito committed 1 year ago

346feb6b

Move memory viz templates into separate javascript files (#103474)

zdevito committed 1 year ago

efc3bcce

Cast computation_node_input_size to int (#103677)

ezyang committed 1 year ago

69969e52

[vision hash update] update the pinned vision hash (#103721)

pytorchupdatebot committed 1 year ago

bcf2beca

Don't apply automatic_dynamic_shapes if we force tensor to be static (#103673)

ezyang committed 1 year ago

d1effcd4

Allow for sympy.Expr in tensor lowering in inductor (#103678)

ezyang committed 1 year ago

39ba2e62

[quant][pt2] Fix no conv bias in convert QAT (#103298)

andrewor14 committed 1 year ago

dad29f90

[export][serde] Add validator to compare deserializer opset version with model opset version (#103691)

larryliu0820 committed 1 year ago

a52b6f08

[reland][inductor] Make clone_graph copy node name as well (#103688)

desertfire committed 1 year ago

1f5ee39c

update README.md to reflect current build from source status on master (#92729)

mikebrow committed 1 year ago

806a642e

Add some missing disabled functions (#103662)

mlazos committed 1 year ago

38f35b4f

Silence CUDA graph tree cuda warning (#103636)

msaroufim committed 1 year ago

ecf4ce7a

Ensure ncclCommAbort can abort stuck ncclCommInitRank (#103264)

pritamdamania87 committed 1 year ago

03881b0c

[inductor] Fix tags for inductor random ops (#103648)

anijain2305 committed 1 year ago

1985c490

[inductor] Fix allow_buffer_reuse=False (#103630)

jansel committed 1 year ago

8c54cd43

[Easy] Dont truncate cudagraph error msg (#103693)

eellison committed 1 year ago

7c152376

[inductor] let coordinate descent tuning respect max block size (#103660)

shunting314 committed 1 year ago

5f979d40

Implement meta functions for rshift and lshift (#103637)

drisspg committed 1 year ago

155691a7

Add symbolic tracing support to torch._dynamo.export (fake input + weights) (#100017)

Thiago Crepaldi committed 1 year ago

6f655d41

[BE][Functorch] Use nested namespace (#103685)

malfet committed 1 year ago

f61b248d

[BE] Remove unused `dim_plane` from `reflection_pad2d_backward_out_template` (#103680)

malfet committed 1 year ago

def01eaf

Newer Older