Reza Yazdani 8e891aa568 Transformer kernel/fix layer norm (#1587) ... * fixing the softmax masking when using triangular masking * fix a bug in the the layernorm backward kernels * revert back some changes & remove debug code * change the constants to a macro Co-authored-by: Olatunji Ruwase <olruwase@microsoft.com>		2021-12-01 20:17:22 +00:00
..
adagrad	force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)	2021-11-29 15:41:18 -08:00
adam	force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)	2021-11-29 15:41:18 -08:00
aio	force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)	2021-11-29 15:41:18 -08:00
common	force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)	2021-11-29 15:41:18 -08:00
includes	Transformer kernel/fix layer norm (#1587)	2021-12-01 20:17:22 +00:00
lamb	Quantization + inference release (#1091)	2021-05-24 01:10:39 -07:00
quantization	Fix many typos (#1423)	2021-10-01 19:56:32 -07:00
sparse_attention	Sparse attn + ops/runtime refactor + v0.3.0 (#343)	2020-09-01 18:06:15 -07:00
transformer	Transformer kernel/fix layer norm (#1587)	2021-12-01 20:17:22 +00:00
utils	DeepSpeed JIT op + PyPI support (#496)	2020-11-12 11:51:38 -08:00