strisunshinewentingwang/pytorch

Pull Requests Commits

Update hipification logic for all ROCm headers (#85320)

jithunnair-amd committed 1 year ago

90b64e23

[quant][docs] README for FX Graph Mode Quantization (#85070)

jerryzh168 committed 1 year ago

2c285f3e

Move functorch C++ into aten/src/ATen/functorch (#85381)

zou3519 committed 1 year ago

5fa104a7

[FSDP] Add back `forward_prefetch` (#85177)

awgu committed 1 year ago

125e9256

[FSDP] Simplify backward prefetch implementation (#85176)

awgu committed 1 year ago

977f8fce

[fix] rrelu, rrelu_, & RReLU when lower bound > upper bound (#84996)

khushi-411 committed 1 year ago

563b065f

Change Lezcano to lezcano (#85396)

lezcano committed 1 year ago

de0f3c42

[pytorch] cuBLAS addmm malfunction test (#85084)

souravmandal committed 1 year ago

0297c75c

Rework printing tensor aliases in CSAN error message (#85008)

sypneiwski committed 1 year ago

b70c254e

as_strided symbolic support (#85264)

ezyang committed 1 year ago

3eb27229

[quant][core][feature] Implement masked_fill for CUDA tensors (#85108)

jcaip committed 1 year ago

793deeef

Add nvFuser support for transpose (#84629)

IvanYashchuk committed 1 year ago

308b26fe

Ported matmul compositeimplicitautograd impl into core (#85239)

Chillee committed 1 year ago

2f4a517d

Revert "Exposing native _scaled_dot_product_attention to torch.nn (#85044)"

pytorchmergebot committed 1 year ago

a3dc338e

quantization: align observer dtype with reference model spec (#85345)

vkuzo committed 1 year ago

09965957

[vision hash update] update the pinned vision hash (#85380)

pytorchmergebot committed 1 year ago

08f413bd

Address eellison's CR comments on AOTAutograd (#85370)

ezyang committed 1 year ago

75451d3c

[BE] Mark unused range-loop vars with `C10_UNUSED` (#85383)

malfet committed 1 year ago

3c870dad

Revert "Improve and expose cpp_backtrace to python binding (#84896)"

pytorchmergebot committed 1 year ago

3122a96e

Exposing native _scaled_dot_product_attention to torch.nn (#85044)

drisspg committed 1 year ago

9fdd8a8b

Remove TS compatibility transition code (#85003)

mikekgfb committed 1 year ago

d7029fea

Improve and expose cpp_backtrace to python binding (#84896)

SherlockNoMad committed 1 year ago

73fbca1e

Update torch.ops.aten.all ref to be symbolic-trace friendly (#85352)

wconstab committed 1 year ago

52fd7e49

[PyTorch] StorageImpl: cache size_bytes.is_symbolic() (#85309)

swolchok committed 1 year ago

f6a18d3d

Fixed memory issues in linalg_lstsq (#85357)

Chillee committed 1 year ago

90fa744c

fx quant: fix bug in custom module test (#85344)

vkuzo committed 1 year ago

cb8e73bb

Fix indentation in functorch limitations docs (#85346)

mplemay committed 1 year ago

62786a09

Setup fake tensor and symbolic shapes once at beginning of AOTAutograd (#85233)

ezyang committed 1 year ago

e1f63475

Correctly handle duplicate arguments to AOTAutograd (#85301)

ezyang committed 1 year ago

5f623f5c

Added `Tensor.to` overloads to `torch._refs.to` (#84802)

jjsjann123 committed 1 year ago

b9b27f76

Newer Older