2930 Commits

Author SHA1 Message Date
hiyouga
09b52a3078 Update wechat.jpg
Former-commit-id: 2a0f1f8398bd771982d41eb53dd38a0dc8087f05
2023-07-14 17:29:43 +08:00
hiyouga
7ea072536b fix eval and pred loss
Former-commit-id: c30db9f1f0db5a6a660cdc60016755241762aae7
2023-07-14 13:11:57 +08:00
hiyouga
32ec440d2e fix pretrain
Former-commit-id: a04115ec278e646347a6fec12f562494abd66e51
2023-07-13 23:41:54 +08:00
hiyouga
2b0fced03b fix Baichuan-13B
Former-commit-id: 08439d29b2031ffbe77fe581c148e6d94e68bfc4
2023-07-13 23:08:45 +08:00
hoshi-hiyouga
cecdf886cb Merge pull request #156 from ZhengJun-AI/main
Support for WebNovel dataset

Former-commit-id: 8cd76ef3c3ac4b92d2ea9b605251a3abbaff90aa
2023-07-12 20:11:19 +08:00
zxbsmk
3b15aacf02 Support for WebNovel dataset
Former-commit-id: 4955dc9eed33a904e4e2b9d5985b3fda87c3674a
2023-07-12 17:29:47 +08:00
hoshi-hiyouga
9c319f3d49 Merge pull request #145 from elicassion/patch-1
Fix typo in common.py

Former-commit-id: 894f13e41f31ff49704ee2f633899f9c88035a9c
2023-07-12 13:50:39 +08:00
Jinghuan Shang
d42d175670 Fix typo in common.py
lastest -> latest

Former-commit-id: dc1e8b7181ba86f028dc22bea9d3c419814c1e19
2023-07-11 18:03:53 -04:00
hiyouga
925a026853 fix sft encode
Former-commit-id: b2f7cb446591e3722b5be8d250ddfe0caa226384
2023-07-11 19:50:33 +08:00
hiyouga
685ba1e02e add baichuan template
Former-commit-id: 1af031c02b3871341cc4d719b5d62cfd02bc126a
2023-07-11 18:57:50 +08:00
hiyouga
507726b2f0 support Baichuan-13B
Former-commit-id: f936a7af0be98cd886ce9a0611a744033818e20f
2023-07-11 16:16:14 +08:00
hiyouga
804cc07db0 Update README.md
Former-commit-id: 8447206bbc2ca579e1fc57d6bda76180c7f89c86
2023-07-10 23:09:11 +08:00
hiyouga
53b8e444b8 Update wechat.jpg
Former-commit-id: 061c324972ac8e20da1445a70332cdd0e347425e
2023-07-10 18:41:53 +08:00
hiyouga
e5c7681146 Update README.md
Former-commit-id: 4182c7aa8b08c80aac0f7a4fc03164c46d8f322c
2023-07-09 14:57:13 +08:00
hiyouga
c812429011 update api to match langchain
Former-commit-id: 84a06318d40fb595f3aa6d1141c107ef7710376c
2023-07-07 20:35:39 +08:00
hiyouga
28ec96bf19 Update README.md
Former-commit-id: 233f20864b69b6483b8fc460df036230c3cbed7d
2023-07-07 12:06:28 +08:00
hiyouga
8b4ef062b7 support InternLM
Former-commit-id: a2f507c56238d7fb2670edbab52d6b275f245e27
2023-07-07 11:02:28 +08:00
hiyouga
50a27509a3 fix rouge score
Former-commit-id: caa00d3ac24c989d56887af9d7ca79244e72714a
2023-07-06 14:28:34 +08:00
hiyouga
28c1381671 update readme
Former-commit-id: 89c623e4bfef760174f013c5de27810cc90e0f6e
2023-07-05 23:03:58 +08:00
hiyouga
22fa516bd5 fix streaming response in API
Former-commit-id: 4abd2485e10e7198b63a396ce5432ac89c9f2d98
2023-07-05 22:42:31 +08:00
hiyouga
bcd7493212 fix freeze tuning
Former-commit-id: e6603977f695b71f6a36640256f6653ff513d783
2023-07-05 21:18:28 +08:00
hiyouga
265dc1b6a0 fix bug in PPO stage
Former-commit-id: a2ba69183b8e72c09242317a34545ab966ea8991
2023-07-05 19:14:10 +08:00
hiyouga
7ba52f5b6e fix compute dtype
Former-commit-id: 8e3540c62d748dcecbcb3a29d57abe665e7b31c8
2023-07-05 15:13:00 +08:00
hiyouga
c87910ada3 support falcon model #72
Former-commit-id: c136f362c1aa75d3374b151188ba4a55d9313a59
2023-07-05 15:00:06 +08:00
hiyouga
827ff46008 Update wechat.jpg
Former-commit-id: 966b5c70fc88936dd05f667975913577cfa7f2f2
2023-07-05 00:22:22 +08:00
hiyouga
9e8854535b fix bleu score
Former-commit-id: cac87fd553500e6fc76430909a469e8b037a46ef
2023-07-05 00:11:21 +08:00
hiyouga
023f9367fc set use_cache before saving model
Former-commit-id: 395ed1cf1b811a21fe61dd6b57040653ce98ba49
2023-07-04 23:18:20 +08:00
hiyouga
b9d56b2ac5 fix seq2seq predictions
Former-commit-id: 65e9ce2cdda3924170774f7b3c6e3fbdbdd87b7f
2023-07-04 22:56:51 +08:00
hoshi-hiyouga
ce622cd93e Merge pull request #119 from codemayq/main
add the pre-built version of bitsandbytes library for windows user

Former-commit-id: cb26f7892384d7209b7388f0cb39cc173de84998
2023-07-03 19:51:46 +08:00
codemayq
7925ffd3cd add the pre-built version of bitsandbytes library for windows user
Former-commit-id: d3b30ecde31ef36c9cb1326666996aee838a55d2
2023-07-03 13:58:10 +08:00
hiyouga
8a1cd612bc Update auto_gptq.py
Former-commit-id: 0db9d2911192194878ef4665b2471a5752b64c65
2023-07-02 20:56:11 +08:00
hiyouga
202e8f1e02 add autogptq
Former-commit-id: cf6d57fd3ee213276e805813a987c4d40395a38c
2023-07-02 20:36:37 +08:00
hiyouga
7a0a6ec28d Update wechat.jpg
Former-commit-id: b8e1f09a2e41cf103f7e3d780709a5d9c93caec5
2023-06-30 15:45:20 +08:00
hiyouga
4ec1bff116 fix typo
Former-commit-id: 92fa515e977889bf1003f174785fe600db14f70e
2023-06-30 10:09:59 +08:00
hiyouga
b064d205c6 Update README.md
Former-commit-id: 021b035c1e9b1627c26e80229c841f28edc6ab51
2023-06-29 19:36:22 +08:00
hiyouga
133399e905 rename evaluate.py
Former-commit-id: f14bd729a850795ac45247e728c8b517509674e7
2023-06-29 15:40:39 +08:00
hiyouga
36c409697e Update evaluate.py
Former-commit-id: 23a7266272a79310d5b20c2f9653c5cd16f40a53
2023-06-29 15:40:03 +08:00
hiyouga
457b53ec6c Update README.md
Former-commit-id: 70592035b81f4f6f682986fa550bf2f08ac5ec77
2023-06-29 15:37:19 +08:00
hiyouga
92070c3d7a add open assistant dataset
Former-commit-id: 3154fec979aba48f54b7afde3740c4990d445a41
2023-06-28 23:09:33 +08:00
hiyouga
1ce7b5e0f3 update loading logic
Former-commit-id: 4d0fddba213beaa55146b047a78963d1d18185a1
2023-06-28 12:07:16 +08:00
hiyouga
2ff577810a fix loading best model
Former-commit-id: 0a46313ccaee91b51bec9f9f92e3111a4a04ce2e
2023-06-28 01:55:12 +08:00
hiyouga
c3cd2067b2 fix RM accuracy
Former-commit-id: 7826a8ca7722b138e79b13c42b1070771f6d5994
2023-06-28 01:40:13 +08:00
hiyouga
204541b56c add star history
Former-commit-id: 9cb1af71f3589b8ca18a26b677207c8d2192f696
2023-06-27 23:56:29 +08:00
hiyouga
c5ec4eaef5 tiny fix
Former-commit-id: 450910c1db969533c5268022cb064cbc2c9cb7e6
2023-06-27 23:54:24 +08:00
hiyouga
e877b8d55b fix initializing data arguments
Former-commit-id: 18f87c1b25c7d1bbc06ea2260a1473b7f296e0ff
2023-06-27 22:50:23 +08:00
hiyouga
307f5866e9 support save full model, replace BOS token
Former-commit-id: 2e01abfda5706d8913860f52ce3bab98739eae55
2023-06-27 21:40:11 +08:00
hiyouga
86a7b2dc5d fix decoding in seq2seq
Former-commit-id: 1c732e2537cce56f537c3dcabf96f9245706f4bb
2023-06-27 19:33:08 +08:00
hiyouga
afc98d2691 Update wechat.jpg
Former-commit-id: 33f21415079d80b71478bb0442b91101ad1c959b
2023-06-27 16:41:09 +08:00
hiyouga
7244f1fdcc Update evaluate.py
Former-commit-id: 4f3772b3424de8b980c5b27e027eeff55010b14a
2023-06-26 23:41:33 +08:00
hiyouga
36666b0b77 Create evaluate.py
Former-commit-id: 5a0a9daf74d9bad52927340f70df11bdcda79c3e
2023-06-26 23:30:18 +08:00