strisunshinewentingwang/pytorch

Pull Requests Commits

Better error msg for autograd profiler + multi-worker dataloader crash (#31473)

zou3519 committed 4 years ago

dbe2f265

Exclude generated source docs from Google (#31484)

zou3519 committed 4 years ago

e67064a9

Speed up `Tensor::has_names` for unnamed tensors (#31436)

zou3519 committed 4 years ago

8f3c0d54

Added error message to indicate that reduction operations are not supported for dim>=64 (#31476)

anjali411 committed 4 years ago

9d9bc93b

add back in reference to jit_unsupported section (#31486)

Elias Ellison committed 4 years ago

779b1288

Updated documentation for NLLLoss to explain what x, y and w refer to (#31488)

anjali411 committed 4 years ago

49fe7a74

Guard against copying from quantized Tensor to non-quantized Tensor (#29660)

jerryzh168 committed 4 years ago

d6acc87c

Fix is_fundamental template for MSVC (#30959)

peterjc123 committed 4 years ago

c4121ed8

Updating submodules

svcscm committed 4 years ago

6d6a91fb

davidriazati committed 4 years ago

28376e82

Bump numba version in circleCI config to 0.46.0. (#31435)

gchanan committed 4 years ago

540b9da4

fixing a naming issue in creating a residual loop node in a bailout graph (#31400)

Krovatkin committed 4 years ago

fc3103b1

Revert D19054937: Add support for `del`

David Riazati committed 4 years ago

1e116a50

Add TORCH_DCHECK macro that checks only in debug builds (#31240)

bddppq committed 4 years ago

489dd6cb

catch all exceptions in converting default values to ivalues (#31398)

Elias Ellison committed 4 years ago

fb24f7c4

Fix getAttribute (#31011)

jerryzh168 committed 4 years ago

1bb6c514

Avoid sending large unneeded data over wire in process_group_agent. (#31357)

jjlilley committed 4 years ago

dff7b945

Updating submodules

svcscm committed 4 years ago

1bb800cf

Use `default_observer` and `default_weight_observer` in tests (#31424)

jerryzh168 committed 4 years ago

fe707c78

Add support for `del` (#31273)

davidriazati committed 4 years ago

e1509cb4

add a suggested alternative to _get_trace_graph

suo committed 4 years ago

e7d25a3e

Temporary fix to support building pytorch from fbsource (for xplat dependencies) (#31393)

Kaikai Wang committed 4 years ago

d2e66b44

Fix default instantation of dynamic quantized LSTM

James Reed committed 4 years ago

a3cdb7ec

autograd/profiler: make record_function more threadsafe (#31346)

d4l3k committed 4 years ago

1e80ff7a

Add support for builtins as attributes (#31269)

davidriazati committed 4 years ago

148bcd3e

Cleanup after moving language reference (#31146)

davidriazati committed 4 years ago

503a4e90

Move TorchScript language reference to its own page (#31138)

davidriazati committed 4 years ago

ae2487bf

fix zero-batch handling in convtranspose (#24341)

wat3rBro committed 4 years ago

d08250c2

Fix hex literal parsing (#29935)

davidriazati committed 4 years ago

7692494c

Throw a better error for int too big for int64_t

davidriazati committed 4 years ago

1f50cfc2

Newer Older