strisunshinewentingwang/pytorch

Pull Requests Commits

[functorch] Fix nll_loss_batch_rule (the decomposition is a negative gather)

zou3519 committed 2 years ago

d9c33511

[functorch] gather_backward batch rule

zou3519 committed 2 years ago

5672317b

[functorch] gather batch_rule, nll_loss batch rule improvement

zou3519 committed 2 years ago

07079d8a

[functorch] masked_fill_.Scalar batch rule

zou3519 committed 2 years ago

f17fb50a

[functorch] Add true_divide/floor_divide/cholesky batching rules

Chillee committed 2 years ago

897f0e94

[functorch] Some upsample backward batch rules

zou3519 committed 2 years ago

1201fc28

[functorch] Quick fix

zou3519 committed 2 years ago

c4e5ce3d

[functorch] fix log_softmax_backward batch rule

zou3519 committed 2 years ago

f904936b

[functorch] softmax_backward batch rule

zou3519 committed 2 years ago

14f02a94

[functorch] Added Pythonkey Opinfo tests for tracing a single op

Chillee committed 2 years ago

d80ce0b6

[functorch] Some easy backward batching rules

zou3519 committed 2 years ago

08057335

[functorch] Batch rule for mse_loss

zou3519 committed 2 years ago

6cea76a2

[functorch] Fix vmapvjp testing; it used to not do anything

zou3519 committed 2 years ago

f08e17f2

[functorch] remove nn.functional.linear from test_vmap

Chillee committed 2 years ago

2f490672

[functorch] added rest of scalar batching rules (xlogy, float_power, special_xlog1py/xlogy/zeta)

Chillee committed 2 years ago

99079159

[functorch] Add xlogy batching rule + UNARY_SCALAR macro

Chillee committed 2 years ago

2c4b5968

[functorch] Fixed some more vjp failures

zou3519 committed 2 years ago

1b5cd8e2

[functorch] Mention that users should install ninja to speed up the build

zou3519 committed 2 years ago

5de1f0c0

[functorch] Fix bug in softmax/log_softmax related to scalars + refactored code to be nicer

Chillee committed 2 years ago

227d0c9c

[functorch] Fix op db + nnc_compile failing tests

Chillee committed 2 years ago

e1a342f0

[functorch] Added nn.functional.mse_loss batching rule

Chillee committed 2 years ago

a3eb2ef9

[functorch] Update eager_fusion.py

Chillee committed 2 years ago

5466ac5d

[functorch] Update README.md

zou3519 committed 2 years ago

02713fd1

[functorch] _to_copy batch rule

zou3519 committed 2 years ago

93fb561a

[functorch] vmapvjpvjp testing

zou3519 committed 2 years ago

d4ada7c7

[functorch] fix example to work well :)

Chillee committed 2 years ago

e8a5b372

[functorch] Moved to Keops example

Chillee committed 2 years ago

1a2e5385

[functorch] Added eager-mode fusion prototype

Chillee committed 2 years ago

9ca5ae86

[functorch] Added initial NNC example of eager-mode fusion

Chillee committed 2 years ago

e632a46f

[functorch] Fix "vmap: inplace arithmetic(self, *extra_args) is not possible" for linear

zou3519 committed 2 years ago

7c405f44

Newer Older