strisunshinewentingwang/pytorch

Pull Requests Commits

Remove CUTLASS extensions merged upstream (#107612)

alexsamardzic committed 1 year ago

cfd98d3c

fixing bug with non-contiguous mixed_mm [inductor] (#107495)

HDCharles committed 1 year ago

6981bcbc

Manually enable `capture_func_transforms` for testing (#107122)

kshitij12345 committed 1 year ago

977a77ca

Implement autograd support for sparse compressed tensor constructors (#107384)

pearu committed 1 year ago

a816aa78

Run check api rate limit on ephemeral runner (#107621)

clee2000 committed 1 year ago

04a7915d

Update persons_of_interest.rst (#107592)

mthrok committed 1 year ago

a250cc9b

Set crow_indices outputs as non-differentiable. (#107447)

pearu committed 1 year ago

d7c0c5de

[ONNX] Update xfail reasons in fx runtime tests (#107257)

titaiwangms committed 1 year ago

a4eae433

Guard numpy imports in the dynamo folder (#107299)

lezcano committed 1 year ago

612c8a8c

[BE]: Add PYI files to ruff lintrunner (#107524)

Skylion007 committed 1 year ago

79d35bfc

[dynamo][eval frame] Make CacheEntry a PyObject (#107405)

anijain2305 committed 1 year ago

e201e3ff

Use default build env and test config for test times (#107325)

clee2000 committed 1 year ago

3b2c5d47

Print per-tensor guard messages for TENSOR_MATCH (#107562)

ezyang committed 1 year ago

ad07a4bc

Adding allocated and reserved memory values to memory timline view. (#107056)

anupambhatnagar committed 1 year ago

3336aa19

[2d] remove ShardedTensor from fsdp extension (#107472)

wanchaol committed 1 year ago

da765995

Revert "Add scalar conversion using avx instructions for half (#102140)"

pytorchmergebot committed 1 year ago

e0f1fe10

[dynamo+aten] Enable embedding_bag_byte_rowwise_offsets + meta kernel impl (#106105)

jfix71 committed 1 year ago

df16b1ed

[inductor] Revert inductor changes in #105977 (#107468)

angelayi committed 1 year ago

d5b8c711

[export] Serialize constrain_as_size ops (#107386)

angelayi committed 1 year ago

a5efb5eb

[torch.compile x autograd.Function] More test cases (#107467)

zou3519 committed 1 year ago

5f56c4fb

[HigherOrderOp] stop erroring out on non-Tensor returns (#107461)

zou3519 committed 1 year ago

72de9b2e

[HigherOrderOps] Saner error message (#107459)

zou3519 committed 1 year ago

c5c41f96

Single source of truth for guard logging (#107532)

ezyang committed 1 year ago

796ce672

Add frame/recompile counter to all log messages in tracing context (#107530)

ezyang committed 1 year ago

8316affc

Optimize load inline via pch (#106696)

xuhancn committed 1 year ago

5ed60477

Fix RenamePlanner documentation (#107535)

b-chu committed 1 year ago

24968383

[FSDP][state_dict] Expose optimizer state_dict config (#105949)

fegin committed 1 year ago

7ba513b6

[export] Add schema version to serializer/deserializer (#107420)

angelayi committed 1 year ago

63e9b548

Don't use thrust::log(complex) in CUDA as it takes a FOREVER to compile (#107559)

lezcano committed 1 year ago

6dea9927

remove unnecessary import introduced in PR 106535 (#107440)

XilunWu committed 1 year ago

5ce88e7e

Newer Older