strisunshinewentingwang/pytorch

Pull Requests Commits

Merge branch 'erichan1/bt-seqlen-1024-fix' of https://github.com/pytorch/pytorch into erichan1/bt-seqlen-1024-fix

erichan1 committed 2 years ago

02020155

remove torchscript test until compatibility window reached

erichan1 committed 2 years ago

b3c4d3bd

add test comparing fast and slow path of transformer encoder. fix src_key_padding_mask bug with seqlen > 1024

erichan1 committed 2 years ago

c78d514f

erichan1 committed 2 years ago

646612cb

update test to catch seqlen 1024 fastpath bug

erichan1 committed 2 years ago

925ea134

Merge branch 'master' of https://github.com/pytorch/pytorch into erichan1/bt-seqlen-1024-fix

erichan1 committed 2 years ago

c08ae6a6

[fx] Make get_isolated_graphmodule accept tracing mode. (#84238)

zhxchen17 committed 2 years ago

df98c529

[functorch] fix multinomial (#83838)

samdow committed 2 years ago

399b1eb8

[reland] Make allreduce compatible with make_fx (#84221)

mrshenli committed 2 years ago

34e5b099

[fx] Make wrapped_fn also work for non-mutating passes. (#84232)

zhxchen17 committed 2 years ago

a402e100

Fix typo (#83802)

hhy3 committed 2 years ago

8aba2535

Add Lazy backend type string (#84228)

antoniojkim committed 2 years ago

7371761d

Give better error message when merge fails to find any rules (#84160)

huydhn committed 2 years ago

adc54dc2

[vulkan] Add vulkan_api_test as an instrumentation test (#83978)

SS-JIA committed 2 years ago

54d86612

Fix typos in docs (#80602)

apeltop committed 2 years ago

e7635c06

Update start_index and end_index for adaptive pooling (#84010)

Aidyn-A committed 2 years ago

372a19d2

Make mergebot failure messages more readable (#84214)

ZainRizvi committed 2 years ago

95863f2c

Link to instructions on submitting an RFC (#83990)

ZainRizvi committed 2 years ago

d62a6ca5

[ci] move XLA pin update to weekly (#84208)

suo committed 2 years ago

724b63d6

[ONNX][Reland] Export node and value with scope name (#82040)

BowenBao committed 2 years ago

80687851

[Quant][fx] Remove WEIGHT_INDEX_DICT and BIAS_INDEX_DICT (Part 2) (#83853)

jcaip committed 2 years ago

d1445945

Back out "Revert D38984222: Don't introduce new overload for SymInt (#83628)" (#84173)

ezyang committed 2 years ago

ad44670f

[Pytorch][Ondevice quantization] Add device side API to convert model (#83807)

kimishpatel committed 2 years ago

cfd18e10

[Pytorch][quantization][ondevice] Add a wrapper API for server side prep (#83742)

kimishpatel committed 2 years ago

eebdcb5a

[pytorch][on device quant] Finalize method for ondevice quant (#83571)

kimishpatel committed 2 years ago

5c7e801c

[On Device Quantization][pytorch]Make insert_quant_dequant support ondevice ptq (#83570)

kimishpatel committed 2 years ago

446afb5f

Replace "_scalar_type" string with constant (#83569)

kimishpatel committed 2 years ago

6a5d9f1b

[Quantization][Pytorch] On device quantization support part 1 (#83568)

kimishpatel committed 2 years ago

9189edb3

[FSDP] Print exec order only in debug mode (#83868)

rohan-varma committed 2 years ago

8acc92eb

[fx][pass] Fix type of exception (#84094)

angelayi committed 2 years ago

352da6de

Older