strisunshinewentingwang/pytorch

Pull Requests Commits

[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

[functorch] Revert TensorWrapper changes to make things more sane

zou3519 committed 2 years ago

a9327ea8

[functorch] update binary_pointwise_batch_rule

zou3519 committed 2 years ago

fd7e524b

[functorch] Quick grab bag of batching rules

zou3519 committed 2 years ago

4ce294d2

[functorch] Update EXISTING_BDIM_BATCH_RULE

zou3519 committed 2 years ago

0a7954b9

[functorch] Updated variadic_bdims_batch_rule

zou3519 committed 2 years ago

3a1a59a7

[functorch] replace basic_unary_batch_rule with BASIC_UNARY_BATCH_RULE

zou3519 committed 2 years ago

745c6336

[functorch] Updated squeeze and squeeze.dim batching rules to new style and added scalar handling (pytorch/functorch#81)

laurencer committed 2 years ago

7e1d730a

[functorch] remove old python key implementation :)

Chillee committed 2 years ago

53155d3b

[functorch] Added vjpfull

Chillee committed 2 years ago

707c26d3

[functorch] updated nnc_compile to work with new python key

Chillee committed 2 years ago

a92ca843

[functorch] pythonkey refactor

Chillee committed 2 years ago

a11bdcc4

[functorch] Tensor printing

zou3519 committed 2 years ago

8107b13b

[functorch] Fix reshape failure by adding _reshape_alias batch rule

zou3519 committed 2 years ago

b3e39c19

[functorch] Add interpolate/upsample batching rules

Chillee committed 2 years ago

f6bb9acd

Newer Older