537 Commits

Author SHA1 Message Date
hiyouga
eb5a852dd5 fix import bug
Former-commit-id: 35b91ea34caade45dd51813b94da5177b852aa4c
2023-11-16 02:27:03 +08:00
hiyouga
f441932bd1 support full-parameter PPO
Former-commit-id: ce783036001397a20b0b4c5da2fea6d0c03389d2
2023-11-16 02:08:04 +08:00
hiyouga
0c1fab84f1 add demo mode for web UI
Former-commit-id: 8350bcf85d5e59b63da46b540c6ad860e8419d9e
2023-11-15 23:51:26 +08:00
hiyouga
3e0b76650a update readme and constants
Former-commit-id: 1e19cf242a1f843b590feefbe24b2cc0a17712b5
2023-11-15 18:04:37 +08:00
hiyouga
e30290444a support multiple modules in freeze training #1514
Former-commit-id: 4907452d955367ebe987e6deae4fd4213628f2b2
2023-11-15 17:08:18 +08:00
hiyouga
4a0be64ae6 fix imports
Former-commit-id: bbbce1f516840f722247edd37057d16502ea0557
2023-11-15 16:47:45 +08:00
hiyouga
06a4820836 disentangle model from tuner and rename modules
Former-commit-id: 4736344eb1595ee023a50d49e8118f4eee46305f
2023-11-15 16:29:09 +08:00
hiyouga
8ee48a9c9e fix #1507
Former-commit-id: 2f02f688e15db8bf8f0dad3b7284863998c207a3
2023-11-15 16:22:32 +08:00
hiyouga
2ba88c6b08 Update cal_lr.py
Former-commit-id: 829e879e040cae3b49ace981b3d4a8eaf1e0c4ae
2023-11-14 21:14:42 +08:00
hiyouga
426c2fc340 Update cal_lr.py
Former-commit-id: 5619e76dc5aa22bd77da80c33a6e52adb4f517d5
2023-11-14 21:13:01 +08:00
hiyouga
b45fb2b3da Update cal_lr.py
Former-commit-id: fcb2daf7f37af2833404ee8f45b538705a2ce7b3
2023-11-14 21:09:30 +08:00
hiyouga
fffb8ea764 add cal_lr.py
Former-commit-id: 42c8fc4fb970775159a68a123d5c7bedb701c8cf
2023-11-14 20:58:37 +08:00
hiyouga
8387f3011c fix #1494
Former-commit-id: d125ef55358837d4d76943739afeb6c70a901cd7
2023-11-14 18:07:20 +08:00
hiyouga
9176b55fe6 fix #1489
Former-commit-id: 3743b7420b823c5a56aef6100121d89b677ffb0e
2023-11-14 15:27:05 +08:00
hiyouga
5c4ddebde5 support eval remote dataset
Former-commit-id: 2d42be32c1b32b26548ea5af5fc3c810f4d668c1
2023-11-14 02:42:30 +08:00
hiyouga
42bb8b6400 fix dc link
Former-commit-id: 88ab33254e02dce7437786fd2d153411118c4594
2023-11-13 23:22:56 +08:00
hiyouga
4a767e5593 release v0.2.2, fix #1478 #1466
Former-commit-id: 35cc1e28f675889c44f75a0a3194005c7f23631b
v0.2.2
2023-11-13 23:09:05 +08:00
hiyouga
37db26800c fix #424
Former-commit-id: 87390ae3b70f654d520b9aadb335c9650130a42c
2023-11-13 22:42:23 +08:00
hiyouga
125587b187 refactor evaluation, upgrade trl to 074
Former-commit-id: 442aefb925c4ff02b98aa30c49c2e01d04f6496a
2023-11-13 22:20:35 +08:00
hiyouga
d83cf3cbc6 Update wechat.jpg
Former-commit-id: 528d91192ab87588e2a751708b0e87ac4ebd37a4
2023-11-12 22:34:19 +08:00
hiyouga
982e0e79c2 fix flashattn warning
Former-commit-id: 4bd8e3906d09bf6ec4b8f6b553a347fca9db4f80
2023-11-10 18:34:54 +08:00
hiyouga
55e097aaac add todo
Former-commit-id: a0c31c68c4909637b86c90c319c321fd887c4910
2023-11-10 14:38:18 +08:00
hiyouga
0fbaa42752 refactor constants
Former-commit-id: 3697a3dc9a0be8141951dfe65812844f66059517
2023-11-10 14:16:10 +08:00
hiyouga
6ee32cf71c tiny fix
Former-commit-id: 415bca900e5cc3afaddd5b06d35f472d9ead3263
2023-11-09 17:20:49 +08:00
hoshi-hiyouga
9b98790fb3 Merge pull request #1454 from yyq/main
Update finetuning_args.py

Former-commit-id: 462730cbd742e8598bfd3e01839165e80f50c531
2023-11-09 17:12:18 +08:00
Yanqing
fc05fd52cf Update finetuning_args.py
更新 chatglm/falcon/bloom 的 lora_target 的名称

Former-commit-id: 3684dffa14ca0551d51027467c0134b884ed1c59
2023-11-09 17:04:40 +08:00
hiyouga
4dbb52750f fix #1452
Former-commit-id: 0e86527d7fae9c9fe0df89d6fbd89035c9d83fe3
2023-11-09 16:41:32 +08:00
hiyouga
164559d01d update readme
Former-commit-id: b3572659f562756375f879b0a97e42d4e2129f54
v0.2.1
2023-11-09 16:00:24 +08:00
hiyouga
c5b202d5c6 release v0.2.1
Former-commit-id: 1db59832fd4239e8f70cff2cab8376589550a9df
2023-11-09 15:54:16 +08:00
hiyouga
38755bced7 add template, modify datasets
Former-commit-id: 386f590209e466b51c17a7ac8cee55fc3ce928d7
2023-11-09 15:53:23 +08:00
hoshi-hiyouga
28a9176784 Merge pull request #1436 from lvzii/main
fix tokenizer config changed after pretrain

Former-commit-id: 7ca32d8e69d5a2790b8ac323f6e6d0aea42600e7
2023-11-09 14:30:50 +08:00
hiyouga
f9ebc718d0 support parquet format #1446
Former-commit-id: 3df90b988bd987daa4e3a991c32ae53446481dcd
2023-11-09 14:17:40 +08:00
hiyouga
b9f42172dd fix #1438 #1439
Former-commit-id: 33422e1feff706bb7cfb235b9e40bd2e124222be
2023-11-09 13:45:10 +08:00
lvzi
13eb365eb7 fix tokenizer config changed after pretrain
Changing tokenizer's attribute at preprocessing stage will result in saving a wrong tokenizer.
for example, baichuan2

Former-commit-id: 043c316ac8913e10b2274867033f194ea92bfcd6
2023-11-08 15:50:46 +08:00
hiyouga
91f406cc99 fix ppo train and dpo eval
Former-commit-id: 01260d975477ebb8570933a1bd7f547b4dba607f
2023-11-07 22:48:51 +08:00
hiyouga
100dc4c458 fix #1422
Former-commit-id: 11c1e1e1570d3712109dd4dce831674a98841bd5
2023-11-07 19:42:01 +08:00
hiyouga
ed584b9f52 fix reward model loading
Former-commit-id: c52336d14435bc3bd98b6070cc1309b5e7d706c4
2023-11-07 17:20:51 +08:00
hiyouga
b446582bfd fix args
Former-commit-id: d92f112951a8d8b28b180c3e2f504a094a9885dd
2023-11-07 16:36:06 +08:00
hiyouga
53fcc531b5 update info
Former-commit-id: 17c64a05796cac70fc76ed728705cd60efa41cae
2023-11-07 16:28:21 +08:00
hiyouga
1f2c56bff9 delete file
Former-commit-id: 479d0af2dc4ab8282b9d55aba1b03ab3a54f400b
2023-11-07 16:20:12 +08:00
hiyouga
d843efc413 fix #1418
Former-commit-id: 7ebd63a609d4bc4f8a645c5c60be77842ebac825
2023-11-07 16:17:22 +08:00
hiyouga
3d40bdb600 upgrade peft, fix #1088 #1411
Former-commit-id: b2a60905f384ada92618bf21301fe96dac1c10bf
2023-11-07 16:13:36 +08:00
hiyouga
936297aeac update requirements
Former-commit-id: 66a91e1fe39483b83c7636c8199c8a87cf6a599e
2023-11-06 19:01:21 +08:00
hiyouga
5c19786f7c use seed in evaluate.py
Former-commit-id: de95b6928293e79c7e204be307c1784ce146c1b1
2023-11-06 18:17:51 +08:00
hiyouga
2627f95ac3 update readme (list in alphabetical order)
Former-commit-id: e1e04cb1f1a0d37cac47c228bd3d5bbcda912aff
2023-11-06 17:18:12 +08:00
hiyouga
a919b6a478 update templates
Former-commit-id: a7eeb8e17c2f23f16732f5a5d767b39bcc1ac517
2023-11-06 12:25:47 +08:00
hiyouga
ecdea0036c fix #1383
Former-commit-id: 2e77a5718a06f013685e673a19a201653c4cad03
2023-11-06 11:42:23 +08:00
hiyouga
034b658348 fix deepseek template
Former-commit-id: d08f5e8a147f1929567d42b6bed8bc998c2a866d
2023-11-05 13:08:46 +08:00
hiyouga
04107b7af6 support deepseek coder #1378
Former-commit-id: 2a8a25819524e84a5e6e907923c47693f8b7a48d
2023-11-05 12:51:03 +08:00
hiyouga
0d488085b8 fix #1365
Former-commit-id: 63ff909310eec1db64bd4bcc1ce013dbd96f0f99
2023-11-05 12:21:07 +08:00