strisunshinewentingwang/pytorch

Pull Requests Commits

Better and more consistent error messages in torch.linalg (#62734)

lezcano committed 2 years ago

f4dd8848

[jit][edge] Add control stack frame to lite interpreter (#65963)

zhxchen17 committed 2 years ago

4dce051c

.github: Migrate linux-xenial-py3.6-gcc7 to GHA (#67072)

zhaoalex committed 2 years ago

ac948f4f

Move the registration of CPython builtin modules to BuiltinRegistry (#67085)

PaliC committed 2 years ago

9de08888

Disable SVE when cross-compiling for M1 (#67114)

malfet committed 2 years ago

d68bb50e

[Static Runtime] Add static_runtime::fused_sigrid_transforms (#66659)

Mike Iovine committed 2 years ago

5d9ff8f3

Use `at::native::is_nonzero` in promoted ops to improve portability (#67097)

pavithranrao committed 2 years ago

8d164a36

[Pytorch][Bootcamp] Add fixes and vanilla testing for Adagrad non-vectorized and vectorized optimizers to handle complex numbers (#66671)

Christopher Gray Howard committed 2 years ago

acb340de

[SR] Remove unused operator() overload (#67001)

Mike Iovine committed 2 years ago

a0495b3c

[SR] Factor operator() implementation into separate function (#67125)

Mike Iovine committed 2 years ago

364645cd

Accept 0-dim channel inputs in convolution layer (#66256)

v0dro committed 2 years ago

edd4d246

OpInfo : nn.functional.embedding (#66997)

kshitij12345 committed 2 years ago

6c985b57

[quant] Fix docs build (#67169)

jerryzh168 committed 2 years ago

adc21f19

[JIT] Freeze allows preservation of submodule attributes (#66102)

Mike Iovine committed 2 years ago

dd81fa90

Set test owners for jit tests (#66808)

janeyx99 committed 2 years ago

09c7771e

[quant] Fix docs error in convert_fx (#67152)

jerryzh168 committed 2 years ago

364c4959

jit trace (#59949)

Krovatkin committed 2 years ago

a7ebf76a

Automated submodule update: kineto (#67133)

facebook-github-bot committed 2 years ago

f1b5f189

[ez] [Docs] Missing import in example for post_local_sgd (#67047)

rohan-varma committed 2 years ago

b5173152

[FSDP] No need for list() in _get_shard (#66957)

rohan-varma committed 2 years ago

0000c88e

[FSDP] Add some comments after reading the code. (#66956)

rohan-varma committed 2 years ago

580efb35

Revert D31514095: Use kernel_func_name from aotCompiler

ngimel committed 2 years ago

b6fa9988

[quant] Fix lint errors (#67138)

jerryzh168 committed 2 years ago

313939c9

Use kernel_func_name from aotCompiler (#66337)

priyaramani committed 2 years ago

7b55dc83

[pt] Add Half precision support for bucketize and searchsorted op (#67077)

jianyuh committed 2 years ago

64c68eda

[quant][graphmode][fx] Remove fbgemm_backend_config_dict for now (#67066)

jerryzh168 committed 2 years ago

2d81d5ab

[quant][fx] Add an option in convert_fx to accept qconfig_dict to skip quantization (#66878)

supriyar committed 2 years ago

8460fa57

[quant][[fx] update observer_fqn to not depend on node.name (#66767)

supriyar committed 2 years ago

d13829e6

Fix common device computation for comparison ops. (#66245)

ysiraichi committed 2 years ago

83f70db9

[quant][graphmode][fx] Use the new convert function instead of the old one in quant-fx2trt tests (#67065)

jerryzh168 committed 2 years ago

3f5adf4f

Newer Older