strisunshinewentingwang/pytorch

Pull Requests Commits

Preserve layout on like constructors (#110242)

eellison committed 1 year ago

3812f2e4

[4/N] Move remaining c10::variant calls to std::variant (#110382)

cyyever committed 1 year ago

d58a91b2

[inductor] Cast loads from boolean tensors to `tl.int1` (#110388)

peterbell10 committed 1 year ago

01b2f25e

[quant][pt2e][docs] Update main quant doc with pt2 export quantization information (#110260)

jerryzh168 committed 1 year ago

28b3ff79

Revert "[HigherOrderOp] Flatten outputs of `wrap`. (#109433)"

pytorchmergebot committed 1 year ago

cba3f407

Revert "Refactor expect tests on test_higher_order_ops.py. (#110290)"

pytorchmergebot committed 1 year ago

85973351

[FSDP][optim_state_dict] Fuse allgather for optim_state_dict when use_orig_params is True (#108298)

fegin committed 1 year ago

cdde899a

Actually enable typechecking for _inductor/index_propagation.py (#110110)

int3 committed 1 year ago

15dfe7b8

[ATen] Remove ATen.h includes from transformers (#110199)

peterbell10 committed 1 year ago

80b6f072

Fix test_mem_efficient_attention_vs_math_ref_grads tolerance from test_transformers.py (#108094)

Fuzzkatt committed 1 year ago

c28bb464

Benchmark flag to include slowdowns when computing gmean of speedups over eager (#108375)

BowenBao committed 1 year ago

6b2c5227

Fix optimize_for_inference to support modules that don't have a forward method (#110013)

sunghyunjun committed 1 year ago

b5268456

[HigherOrderOp] Flatten outputs of `wrap`. (#109433)

ysiraichi committed 1 year ago

651b198c

Refactor expect tests on test_higher_order_ops.py. (#110290)

ysiraichi committed 1 year ago

d9aecaef

[PyTorch] Add Expanded call stack to nodes [Take 2] (#110229)

RihamSelim committed 1 year ago

92242f59

Add function to port FX minified graph to HLO via StableHLO (#109084)

awskila committed 1 year ago

16e3f158

Revert "Multiprocessing support for NT (#110292)"

pytorchmergebot committed 1 year ago

7e6cf04a

Multiprocessing support for NT (#110292)

jbschlosser committed 1 year ago

881e7304

Increase job timeout limit when running with memory leak check (#110193)

huydhn committed 1 year ago

7827ae28

Revert "Adding Backward Support for NestedTensors and FlashAttention (#97485)"

pytorchmergebot committed 1 year ago

8d647972

[dtensor] skip pytree when not necessary (#110132)

wanchaol committed 1 year ago

26900d21

Add missing CUDA error check (#110368)

cyyever committed 1 year ago

fd6c993e

fix(lint): Fix lint issues on main (#110389)

jon-chuang committed 1 year ago

46d1f9b3

Generalize toAccumulateType() (#108248)

qqaatw committed 1 year ago

a3c1e3c9

Fix override warnings in nvfuser (#110350)

cyyever committed 1 year ago

7853f8f6

[aotinductor] Use dynamic_shape instead of constraints (#110360)

angelayi committed 1 year ago

e47e946b

Added new test sample to interpolate op in OpInfo (#104181)

vfdev-5 committed 1 year ago

87f8bc65

Enable torch.promote_types in Dynamo tracing (#110358)

cdzhan committed 1 year ago

175b6262

Avoid undefined behavior in JIT-generated conversion code (#110212)

Flamefire committed 1 year ago

e0348cec

[inductor][Optimus]Improve logging for Optimus (#110186)

mengluy committed 1 year ago

f7812cdb

Newer Older