strisunshinewentingwang/pytorch

Pull Requests Commits

model_dump: Use DumpUnpickler.load instead of .dump (#57659)

dreiss committed 3 years ago

ab1fdbef

model_dump: Add a section that summarizes tensor memory usage (#57658)

dreiss committed 3 years ago

53078924

model_dump: Handle dict rendering (#57657)

dreiss committed 3 years ago

ef4e6036

model_dump: Handle torch.device objects (#57656)

dreiss committed 3 years ago

72ff3163

model_dump: Refactor renderTensor into a helper method (#57655)

dreiss committed 3 years ago

a380575f

model_dump: Implement "Hider" properly (#57654)

dreiss committed 3 years ago

3ff76af2

move code to Blas.cpp, clean up THC magma (#58526)

ngimel committed 3 years ago

3f0b0816

[JIT] improve documentation (#57991)

garymm committed 3 years ago

703cfdc9

s/foward/forward/g (#58497)

Chillee committed 3 years ago

79a258f4

Added OperatorMap for mapping Operator to any template <T> (#58060)

Edvard Ghazaryan committed 3 years ago

ccad77aa

[Reducer] Remove some unused variables (#58524)

rohan-varma committed 3 years ago

1ba05efd

Fix issues regarding binary_checkout (#58558)

seemethere committed 3 years ago

4cf9b110

Split CUDA SpectralOp (#58459)

malfet committed 3 years ago

baf05c3f

[lint] Fix uninitialized variable lint error in `Module.cpp` (#58499)

1ntEgr8 committed 3 years ago

029bec45

Automated submodule update: tensorpipe (#58477)

facebook-github-bot committed 3 years ago

b45a105a

[Quant] Add out variant for int8 quantized::linear (#58282)

Hao Lu committed 3 years ago

4d7abdbd

[nnc] Enable CPU fusion inside Facebook, take 2 (#58347)

bertmaher committed 3 years ago

c76405d3

VaryingShape<Strides>::isComplete() needs to consider whether each Stride is complete (#58510)

bertmaher committed 3 years ago

dcfc2050

[nnc] Do not fuse unsqueeze with variable dim (#58346)

bertmaher committed 3 years ago

3d20ddfe

[nnc] Make the pretty printer prettier (#57874)

bertmaher committed 3 years ago

2ddd8416

[jit] Add a utility class SourceRef to represent Source as keys (#57396)

zhxchen17 committed 3 years ago

3a3959d2

[BE] Use __func__ as checkAllSameGPU() 1st arg (#58502)

malfet committed 3 years ago

0362b753

move unused parameters to end of bucket orders when rebuild buckets for static graph (#58097)

zhaojuanmao committed 3 years ago

ea0f7c47

[PyTorch Edge] bytecode version bump to v5 and enable share constant table (#57888)

cccclai committed 3 years ago

a7b62abe

[nnc][scripts] Add a script for bisecting the TE fuser pass (#58357)

bertmaher committed 3 years ago

9eee782c

removing old comment (#56430)

Krovatkin committed 3 years ago

7d78d72d

Comment why render_test_results is its own step (#58505)

samestep committed 3 years ago

a07cd22e

Add long tensor type to AddFakeFp16 Op (#58504)

venkatacrc committed 3 years ago

8efaab1b

[NNC] Do not optimize conditionals when the corresponding loop is not normalized (#57675)

navahgar committed 3 years ago

4b859cbc

[NNC] Add a method to check if a loop is normalized (#57674)

navahgar committed 3 years ago

a71b99b5

Newer Older