アマデウス b8899e0905 [TP] allow layernorm without bias (#750)

2022-04-14 11:43:56 +08:00

..

layer

[TP] allow layernorm without bias (#750)

2022-04-14 11:43:56 +08:00

loss

[hotfix] Raise messages for indivisible batch sizes with tensor parallelism (#622)

2022-04-02 16:12:04 +08:00

lr_scheduler

Refactored docstring to google style

2022-03-29 17:17:47 +08:00

metric

[hotfix] Raise messages for indivisible batch sizes with tensor parallelism (#622)

2022-04-02 16:12:04 +08:00

model

Develop/experiments (#59)

2021-12-09 15:08:29 +08:00

optimizer

[zero] improve adaptability for not-shard parameters (#708)

2022-04-11 13:38:51 +08:00

__init__.py

Layer integration (#83)

2021-12-27 15:04:32 +08:00

init.py

Refactored docstring to google style

2022-03-29 17:17:47 +08:00