strisunshinewentingwang/pytorch

Pull Requests Commits

janeyx99 committed 2 years ago

a004f135

Added prod op to FX2TRT (#72284)

Kelly Stanton committed 2 years ago

9ac28cbf

Define TORCH_ASSERT_ONLY_METHOD_OPERATORS in ATen/core (#72344)

peterbell10 committed 2 years ago

bebf8dd5

dbr quant function fusion [2/x]: use fusion for observation and inference (#71781)

vkuzo committed 2 years ago

998a5adf

fx quant: add fusion matching for operator.add and torch.relu (#71780)

vkuzo committed 2 years ago

d672bbd0

dbr quant function fusion [1/x]: record matches for functions (#71764)

vkuzo committed 2 years ago

5937c48f

dbr quant: record dag of non-quantizeable ops (#71551)

vkuzo committed 2 years ago

7e2a7728

dbr quant: refactor first_call hooks to be standalone (#71324)

vkuzo committed 2 years ago

15808564

dbr quant: rename seen_op_info to seen_q_op_info (#71312)

vkuzo committed 2 years ago

0ff9cf39

[bench] Adding a cpp benchmark to compare performance of nnc with static and symbolic shapes (#72197)

navahgar committed 2 years ago

4eb277ac

[bench] Fix build issues with TensorExpr cpp benchmarks (#72196)

navahgar committed 2 years ago

237e960e

Automated submodule update: FBGEMM (#72158)

facebook-github-bot committed 2 years ago

8559d39c

[Codemod][FBSourceBlackLinter] Daily `arc lint --take BLACK`

CodemodService Bot committed 2 years ago

1edf6f56

Add num_head param to native multihead attention to evade dim_per_head always 64 bug (#72375)

zrphercule2 committed 2 years ago

b9d4c0c7

enables module modification during repackaging (#71520)

PaliC committed 2 years ago

5c6b8975

[Static Runtime] Fix build issues (#72310)

navahgar committed 2 years ago

745d8a05

[Static Runtime] Fix printing graphs in debug mode during fusion (#72222)

navahgar committed 2 years ago

2e82c1e5

[PyTorch][JIT] use a better hash table in alias analysis (#69854)

swolchok committed 2 years ago

3ad97179

[PyTorch Edge] Add Quantized Matmul Op (Naive Implementation) (#71783)

salilsdesai committed 2 years ago

26b8a009

Back out "free up dispatch key space (in C++)"

JacobSzwejbka committed 2 years ago

791e7df7

Back out "DispatchKeySet perf improvements"

JacobSzwejbka committed 2 years ago

888e3fbc

Fixing few bugs in torch flatbuffer (#72349)

qihqi committed 2 years ago

57f039b4

Add TORCH_CUDA_CU_API to CUDABlas functions (take 2) (#72340)

malfet committed 2 years ago

f2f40ce8

[quant][fx][devs] Delete unused code (#72244)

jerryzh168 committed 2 years ago

f5c7f815

Update test-infra branch for test stats. (#72233)

kit1980 committed 2 years ago

706f2b30

[ONNX] Resolve attribute error in CI (#72350)

BowenBao committed 2 years ago

90c8623c

[Quant][core][improvement] Enabled slicing on per-channel quantized tensors (support is limited to the a contiguous sliced tensor); added correpsonding test case (#71269)

dzdang committed 2 years ago

b33831dc

[Quant][devs] Separated implementations for quantized & non-quantized tensors in empty_like (#71990)

dzdang committed 2 years ago

d6d6d4cc

Revert D31316086: [fx-acc] PassManager

albanD committed 2 years ago

6451e525

[fx-acc] PassManager (#67261)

alexbeloi committed 2 years ago

aa4f048d

Newer Older