minimind

Commit Graph

Select branches

Hide Pull Requests

master

website

#132

#136

#144

#162

#189

#2

#204

#205

#214

#214

#237

#240

#240

#262

#29

#315

#317

#323

#331

#34

#349

#354

#376

#377

#383

#385

#394

#409

#412

#42

#421

#421

#424

#43

#44

#442

#443

#452

#478

#478

#491

#493

#497

#497

#503

#507

#507

#509

#513

#518

#518

#521

#521

#523

#523

#524

#524

#528

#534

#536

#539

#54

#543

#569

#569

#570

#571

#577

#579

#586

#59

#591

#594

#598

#599

#599

#609

#621

#63

#643

#653

#658

#663

#663

#664

#666

#666

#671

#694

#696

#698

#700

#708

#708

#710

#720

#725

#731

#738

#753

#756

#757

#759

#762

#762

#763

#769

#77

#770

#770

#778

#778

#780

#780

#784

#784

#785

#789

#789

#791

#792

#798

#802

#803

#807

#807

#808

#809

#809

#811

#811

#812

#813

#814

#815

#817

#818

#818

#821

#825

#829

#831

#831

#832

#833

#833

#836

#837

#837

#840

#843

#843

#844

#844

#847

#851

#852

#852

#854

#855

#857

#859

#86

#860

#861

#862

#863

#864

#865

#866

#866

#867

#868

#869

#869

#870

#870

#871

#871

#873

#874

#876

#876

#877

#879

#881

#881

#882

#882

#883

#883

#886

#886

#888

#888

#889

#889

#891

#891

#893

#894

#894

#90

#95

minimind-v1

v2

5d645e8326 full_sft leoz9 2025-01-10 16:27:59 +0800
e274974822 add scripts leoz9 2025-01-10 16:09:12 +0800
e6e79a2f06 update readme gongjy 2024-12-13 20:08:59 +0800
6e9cd28ef9 update readme gongjy 2024-12-13 20:08:59 +0800
ba4c7b85df

💻 Code: Remove unnecessary code. Mode switching is unnecessary; the model's weight values will not change due to mode switching. No forward computation is performed; it only saves the model parameters. yym68686 2024-12-03 17:31:49 +0800
4486170ff1

Update 5-dpo_train.py LeiZhang 2024-11-14 19:47:26 +0800
141bff3ef1 update lora-sft gongjy 2024-11-10 20:40:49 +0800
f7127e4310 update lora-sft gongjy 2024-11-10 20:40:49 +0800
8669d9c10f 修正了训练tokenizer中的chat_template中的逻辑,以及修正了tokenizer_config.json相应部分 matianhao 2024-11-08 01:05:51 +0800
1fb0df458b update train_tokenizer gongjy 2024-11-06 17:48:33 +0800
1240829c89 update train_tokenizer gongjy 2024-11-06 17:48:33 +0800
14d16f9176 update fast_inference gongjy 2024-10-30 15:26:28 +0800
7c67ba0b92 update fast_inference gongjy 2024-10-30 15:26:28 +0800
81da3f2f5b update readme gongjy 2024-10-30 10:39:29 +0800
db39571493 update readme gongjy 2024-10-30 10:39:29 +0800
8a2523b345 update eval-chat gongjy 2024-10-28 12:57:25 +0800
0bce9e5a31 update eval-chat gongjy 2024-10-28 12:57:25 +0800
b4e84fe6bd

Update requirements.txt Xialie Zhuang 2024-10-28 11:28:40 +0800
207294115c update readme gongjy 2024-10-27 21:25:55 +0800
6d7a988365 update readme gongjy 2024-10-27 21:25:55 +0800
eccbeb3618 update process gongjy 2024-10-27 17:21:29 +0800
0d10efeeed update process gongjy 2024-10-27 17:21:29 +0800
e5bc100157 update readme gongjy 2024-10-26 14:35:22 +0800
517e0a4e7c update readme gongjy 2024-10-26 14:35:22 +0800
403d00dadf update readme gongjy 2024-10-26 14:32:55 +0800
d63c525d04 update readme gongjy 2024-10-26 14:32:55 +0800
8425ec9eeb update sft-lr gongjy 2024-10-24 08:58:31 +0800
a9e65969c8 update sft-lr gongjy 2024-10-24 08:58:31 +0800
aca94aa137 update data_process gongjy 2024-10-23 12:25:45 +0800
42bd06e55d update data_process gongjy 2024-10-23 12:25:45 +0800
62bed77a70 update data_process gongjy 2024-10-23 12:02:28 +0800
69bcb8dc90 update data_process gongjy 2024-10-23 12:02:28 +0800
7840e6f270 update model gongjy 2024-10-20 15:13:58 +0800
3ff66f7221 update model gongjy 2024-10-20 15:13:58 +0800
17f6021d32 update num_workers gongjy 2024-10-19 06:31:15 +0800
00e09efb5b update num_workers gongjy 2024-10-19 06:31:15 +0800
4e7bf23414 update sft gongjy 2024-10-16 22:58:06 +0800
f16991d7ec update sft gongjy 2024-10-16 22:58:06 +0800
4556cd5386 update rlhf gongjy 2024-10-15 15:20:31 +0800
6861d1af56 update rlhf gongjy 2024-10-15 15:20:31 +0800
829689cb55 update rlhf gongjy 2024-10-15 15:04:38 +0800
c59b8b3e26 update rlhf gongjy 2024-10-15 15:04:38 +0800
25e505ccff update rlhf gongjy 2024-10-15 10:30:32 +0800
11d5cadb9c update rlhf gongjy 2024-10-15 10:30:32 +0800
3e424076fb update trl gongjy 2024-10-13 22:44:28 +0800
02adb7bc0d update trl gongjy 2024-10-13 22:44:28 +0800
42c36f4916 update inference gongjy 2024-10-13 13:11:41 +0800
5c4b34bbe3 update inference gongjy 2024-10-13 13:11:41 +0800
f54130aa8e update inference gongjy 2024-10-13 13:09:52 +0800
ad6fc92399 update inference gongjy 2024-10-13 13:09:52 +0800
7d720f5ea2 update data_process gongjy 2024-10-12 19:46:08 +0800
135421690e update data_process gongjy 2024-10-12 19:46:08 +0800
0fbeba5e5d update data_process gongjy 2024-10-12 18:47:08 +0800
2698f6b57d update data_process gongjy 2024-10-12 18:47:08 +0800
fed0ccc8c2 update lora-sft gongjy 2024-10-11 17:43:52 +0800
36fadc7ef1 update lora-sft gongjy 2024-10-11 17:43:52 +0800
3090a438a4

fix 5-dpo_train.py bugs Zhao Jiahao 2024-10-11 12:29:23 +0800
c19d111cea update readme gongjy 2024-10-10 22:13:10 +0800
3a034a47c8 update readme gongjy 2024-10-10 22:13:10 +0800
707e598b68 update index.html gongjy 2024-10-09 09:59:06 +0800
73e509b95d update readme gongjy 2024-10-09 09:17:22 +0800
a775e8ddd8 update readme gongjy 2024-10-09 09:12:55 +0800
5dbd6174b3 update readme gongjy 2024-10-09 09:12:55 +0800
70c60f5655 update readme gongjy 2024-10-09 09:12:24 +0800
ed0880ebae update readme gongjy 2024-10-09 09:00:06 +0800
251296b916 update readme gongjy 2024-10-09 08:51:41 +0800
2b9dd44f23 update readme gongjy 2024-10-09 08:51:13 +0800
89f7b18035 update readme gongjy 2024-10-09 00:03:07 +0800
4542ecf858 update readme gongjy 2024-10-09 00:03:07 +0800
e232ed82bb update readme gongjy 2024-10-08 23:40:29 +0800
772834148e update readme gongjy 2024-10-08 23:40:29 +0800
93853be90a update readme gongjy 2024-10-08 23:16:29 +0800
000b0a496b update readme gongjy 2024-10-08 23:16:29 +0800
0fe2351882 update readme gongjy 2024-10-07 18:42:53 +0800
5929d0f8b1 update readme gongjy 2024-10-07 18:42:53 +0800
36e70d30e2 update readme gongjy 2024-10-05 22:59:00 +0800
e4b8789d8c update readme gongjy 2024-10-05 22:59:00 +0800
acea505af0 update readme gongjy 2024-10-05 00:35:54 +0800
eb875da306 update readme gongjy 2024-10-05 00:35:54 +0800
5c06e0175e Auto tokenizer name path fix krmst 2024-10-03 22:40:47 +0900
571fec6fdb update requirements.txt gongjy 2024-10-03 14:45:58 +0800
1b864453fa update requirements.txt gongjy 2024-10-03 14:45:58 +0800
0ee1000392 Update requirements krmst 2024-10-01 11:54:59 +0900
5e67759d1a update model (fix loss bug) gongjy 2024-09-29 16:58:48 +0800
a87f628400 update model (fix loss bug) gongjy 2024-09-29 16:58:48 +0800
1119025ac5 fix deepspeed local_rank bug gongjy 2024-09-27 23:25:56 +0800
4ef9c41563 fix deepspeed local_rank bug gongjy 2024-09-27 23:25:56 +0800
43a7652134 Update data preprocessing methods gongjy 2024-09-27 22:34:30 +0800
2981d3ea86 Update data preprocessing methods gongjy 2024-09-27 22:34:30 +0800
682003cec6 Update data preprocessing methods gongjy 2024-09-27 17:19:03 +0800
75753ea765 Update data preprocessing methods gongjy 2024-09-27 17:19:03 +0800
a80e8e5f52 update readme info gongjy 2024-09-27 16:38:18 +0800
1cc73836d4 update readme info gongjy 2024-09-27 16:38:18 +0800
f351a61c8a Update data preprocessing methods gongjy 2024-09-27 16:19:30 +0800
a8ae342775 Update data preprocessing methods gongjy 2024-09-27 16:19:30 +0800
89b91ceaa0 update batchsize gongjy 2024-09-25 12:35:29 +0800
d57037624b update batchsize gongjy 2024-09-25 12:35:29 +0800
fc1400909a fix dtype bug gongjy 2024-09-25 10:07:30 +0800
89d260145f fix dtype bug gongjy 2024-09-25 10:07:30 +0800
a0081c1931 Merge pull request #44 from iomgaa-ycz/wandb jingyaogong 2024-09-24 14:10:05 +0800