strisunshinewentingwang/pytorch

Pull Requests Commits

[fx] Arg normalization not save output node in the node_map (#58058)

842974287 committed 3 years ago

bfd0a461

Trigger Windows multi gpu tests on master (#57817)

skyline75489 committed 3 years ago

3603ba24

Update CI images for rocm4.2 (#58017)

jithunnair-amd committed 3 years ago

8f83bfeb

[ROCm] ubuntu version check in install_rocm.sh (#57751)

jeffdaily committed 3 years ago

94bb1150

test experiment script (#57925)

gcramer23 committed 3 years ago

16d617c3

Enable `BFloat16` for `nan_to_num` on CUDA (#58063)

imaginary-person committed 3 years ago

d212bf18

[wip] enhance DDPSink to work for general outputs (#57073)

rohan-varma committed 3 years ago

c52700db

Remove printout from distributed tests (#58095)

rohan-varma committed 3 years ago

4faa4273

Reimplement torch::flip based on advanced indexing (#56713)

andfoy committed 3 years ago

30f26c58

Grammatically updated the tech docs (complex_numbers.rst) (#57540)

atsigkas committed 3 years ago

5ea87f9c

OpInfo: reshape, reshape_as and minor clean-up (#57460)

kshitij12345 committed 3 years ago

ff982ef7

Revert D28291041: enable deterministic path for index_copy_cuda with index_put

Mike Ruberry committed 3 years ago

c911c305

Remove beta warning for use_deterministic_algorithms (#58074)

kurtamohler committed 3 years ago

c7fb0a0e

std/var: Return real results for complex input (#58066)

peterbell10 committed 3 years ago

e1078d42

Deprecate symeig (#57732)

lezcano committed 3 years ago

db13119f

[profiler][small] Add skip_first parameter to the default schedule (#58025)

Ilia Cherniavskii committed 3 years ago

e18f5f1d

[profiler][small] Speed up postprocessing (#58021)

Ilia Cherniavskii committed 3 years ago

cdf161c3

[profiler][small] Handle empty trace (#58013)

Ilia Cherniavskii committed 3 years ago

bf2ebfc9

[profiler][resend] Add cuda memory and distributed metadata (#58010)

Ilia Cherniavskii committed 3 years ago

f1defeae

enable deterministic path for index_copy_cuda with index_put (#57870)

yuguo68 committed 3 years ago

14badd99

enable deterministic path for index_put with accumulate=False on CPU and CUDA (#57839)

yuguo68 committed 3 years ago

a07a0190

Add a disclaimer about limited CUDA support in RPC (#58023)

lw committed 3 years ago

d623fb7e

[ATen] Use expect_contiguous in layer_norm (#58067)

Hao Lu committed 3 years ago

c3d40fdf

ATen lu_unpack. Required for making `torch.lu_solve` differentiable. (#46913)

nikitaved committed 3 years ago

c790fd2b

[Static Runtime] Fix bug in aten::clone (#58100)

Hao Lu committed 3 years ago

32acc96f

port topk to structured (#57790)

eqy committed 3 years ago

8c91acc1

[Static Runtime] Add schema check to aten::repeat and fb::fast_gather (#58106)

Hao Lu committed 3 years ago

e9e12547

Split `test_testing.py::TestAsserts` for multiple devices (#56365)

pmeier committed 3 years ago

8824f49e

To implement gradient for Pytorch (#54617)

iramazanli committed 3 years ago

8b816e90

Let submodules be collected as args/kwargs (#57840)

Ansley Ussery committed 3 years ago

0d4dc6cb

Newer Older