strisunshinewentingwang/pytorch

Pull Requests Commits

Remove LazyStreamContext (1 out of 2) (#59298)

lw committed 3 years ago

03a5c6ea

Fix CUDA sync when switching streams in RPC tests (#59297)

lw committed 3 years ago

3e7396f9

Fix race condition in TP agent (#58753)

lw committed 3 years ago

8f4cfaa9

Ensure async_execution works with CUDAFuture (#56863)

lw committed 3 years ago

c0acffa6

Avoid re-doing CUDA stream sync in OwnerRRef (#57355)

lw committed 3 years ago

7bcd8f94

[RPC Framework] Separate initialize_from_module_rref method out of RemoteModule constructor (#59292)

Yi Wang committed 3 years ago

d009c9c1

Fix symbolic derivative of hardswish (#59405)

bertmaher committed 3 years ago

c3bf42e0

[nnc] Add hardsigmoid (#59069)

bertmaher committed 3 years ago

9ac95478

[NNC] Add python bindings for Compute2 (#59350)

huiguoo committed 3 years ago

c717ce67

Make JIT not assume that the device is CUDA. (#54238)

chengjunlu committed 3 years ago

db90533b

[NNC] Fix loopnest.cache_accesses for reduce ops (fixed #59002) (#59136)

huiguoo committed 3 years ago

7c4ac9e3

[torch] Remove migration warning for ScriptDict

Meghan Lele committed 3 years ago

d9d7d5e2

[Static Runtime] Fix bug in quantized::linear wrapper (#59407)

Hao Lu committed 3 years ago

6627c00e

[NNC] Fix BufHandle arguments in loopnest python API (#59348)

huiguoo committed 3 years ago

7d38901e

[torch distributed] Implementing reduce_scatter_base (#57567)

liangluofb committed 3 years ago

77de640f

Revert D28859795: [nnc] Enable CPU fusion inside Facebook, take 4

bertmaher committed 3 years ago

46d724c9

Update reviewer list for the distributed package (#59417)

cbalioglu committed 3 years ago

526445df

Prefer accurate reciprocal on ARMv8 (#59361)

malfet committed 3 years ago

aa4f27c1

Automated submodule update: FBGEMM (#59337)

facebook-github-bot committed 3 years ago

3416b8dd

Fix the "tensors to be on the same device" error in HistogramObserver (#59234)

KAI ZHAO committed 3 years ago

1aa14fcb

Support switching RemoteModule between train/eval (#59026)

Lishan Yang committed 3 years ago

2aa463d9

Revert D28538996: Enable torch::deploy GPU tests in sandcastle

H-Huang committed 3 years ago

c1c9774a

Add build support for kineto + rocm (#58401)

mwootton committed 3 years ago

e66015da

[DDP] log usage of torch_distributed_debug (#59351)

rohan-varma committed 3 years ago

332b01e9

Migrate renorm to ATen (CPU and CUDA) (#59250)

peterbell10 committed 3 years ago

6408cbd9

Extract c10d Store tests to dedicated test file (#59271)

awgu committed 3 years ago

2ad4b8e5

Preserve PyObject even when it goes dead (#56017)

ezyang committed 3 years ago

f05d5bec

[quant] Fix use after free (#59267)

jerryzh168 committed 3 years ago

fa72d9a3

[nnc] Enable CPU fusion inside Facebook, take 4

bertmaher committed 3 years ago

6baa66ec

Revert D28856713: [PyTorch Edge] Add proper error message when loading incompatible model with lite interpreter

Mike Ruberry committed 3 years ago

57e452ff

Newer Older