王晶
f6b12f4631
!760 支持Mixtral 8x7B MOE模型
...
Merge pull request !760 from 王晶/modellink
2024-03-04 10:38:57 +00:00
guhangsong
9b4153e4cf
!480 支持指令微调功能
...
Merge pull request !480 from guhangsong/sft
2024-01-30 02:13:31 +00:00
liuyanghan
76e9d607e1
modellink基于megatron实现分布式lora微调
2024-01-08 19:51:37 +08:00
yangcheng
96b56395e4
llama2_7b
2024-01-05 10:27:02 +08:00
fengliangjun
15963528c8
rename ascendspeed to modellink
2023-12-21 11:02:07 +08:00
mhh001
ad6f867d94
transformer框架升级
2023-11-14 20:44:46 +08:00
g00841271
981428f3b0
(Part Ⅳ ) replace assert with check_*
2023-10-27 16:25:13 +08:00
Zhihong Chen
5bf57ecb87
change dict[] to dict.get()
2023-10-19 10:23:32 +08:00
mhh001
e3aeab62e7
(第二部分)tensor parallel v1到v3升级,目前提交的代码会整体将ascendspeed框架迁移到新版的tp上面,删除原有的mpu目录,涉及到引用的修改,还有一些新版本代码适配的修改。
2023-10-13 10:45:41 +08:00
g00841271
3483bddb1c
(PartⅠ) replace assert with check_*
2023-10-11 15:25:19 +08:00
fengliangjun
7a21f0bf58
up
2023-09-26 14:30:47 +08:00
fengliangjun
260e8eea8f
create megatron core
2023-07-24 15:00:57 +08:00
wiyr
6304cab765
remove useless code
2023-06-20 16:54:12 +08:00
machangjun
2d8c6fee9d
add bloom st and adapt new data load method
...
modify bloom st run
modify bloom st run
modify times
add new pretrain_bloom.py
add new pretrain_bloom.py
add new pretrain_bloom.py
add new pretrain_bloom.py
add new pretrain_bloom.py
add new pretrain_bloom.py
add new pretrain_bloom.py
add new pretrain_bloom.py
add new pretrain_bloom.py
add st
2023-06-17 17:36:17 +08:00
wiyr
2f826f7351
can run with bloom7b and pass ci
2023-06-12 14:42:29 +08:00
chenzomi
37cc0b949d
change megatron to ascendspeed
2023-06-10 21:26:01 +08:00
fengliangjun
106a415556
inital AscendSpeed
2023-06-09 16:15:23 +08:00
wangyixian
d55d341fe1
Adapt the bloom 7.1b model to the AscendSpeed framework, which is jointly completed by liulinfeng and wangyixian
2023-06-06 22:30:19 +08:00
chenzomi
e4a120a662
fork megatron-deepspeed code.
2023-05-25 14:49:59 +08:00