Commit Graph
14 Commits
Author SHA1 Message Date
rlsu9 01282b7622 debug ac; disitll; move fsdp_util 2024-12-12 19:07:43 +00:00
rlsu9 b14cc034e4 update 2024-12-12 18:11:04 +00:00
Peiyuan Zhang 87913376be add training 2024-12-12 05:59:11 +00:00
Peiyuan Zhang f1dba33df7 update 2024-12-11 23:56:13 +00:00
Peiyuan Zhang e8ce4290ff update 2024-12-11 20:27:18 +00:00
Peiyuan Zhang fb1cc737e3 the same input output format for transformer 2024-12-11 19:56:57 +00:00
Peiyuan Zhang f5074a5c31 refactor single card attention to no pad 2024-12-11 19:00:10 +00:00
Peiyuan Zhang 0d4ae825c9 add no pad; support batch size > 1 2024-12-11 18:39:30 +00:00
Peiyuan Zhang da66e0fcea move rope freq inside transformer 2024-12-11 18:18:44 +00:00
foreverpiano dd2cf8ad69 update 2024-12-11 13:04:48 +00:00
foreverpiano f5f5903aa4 update sp 2024-12-10 13:59:33 +00:00
foreverpiano ea0332f62a inference hunyuan ok 2024-12-09 10:10:59 +00:00
Zhang Peiyuan de1e8d868e Cleanup (#75) 2024-12-06 20:56:06 -08:00
Zhang Peiyuanandforeverpiano 8d41d505fe [cleanup] (#72)
Co-authored-by: foreverpiano <pianoqwz@gmail.com>
2024-12-05 21:02:10 -08:00