github
/
minimind
mirror of https://github.com/jingyaogong/minimind.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • 288a1d7212 [feat] get params jingyaogong 2025-12-31 20:44:34 +0800
  • e34bd5c90e docs: clarify pretraining data format in README dieu 2025-12-31 13:39:59 +0800
  • 8ace03f068 [feat] update config jingyaogong 2025-12-31 10:29:13 +0800
  • eead9538b2 [feat] update config jingyaogong 2025-12-31 10:29:13 +0800
  • 6ee867fffc [feat] update lr jingyaogong 2025-12-31 10:27:09 +0800
  • 6242980917 [feat] update lr jingyaogong 2025-12-31 10:27:09 +0800
  • cc40ff6f45 [feat] compatible tokenizer jingyaogong 2025-12-31 10:26:46 +0800
  • 936d105e9b [feat] compatible tokenizer jingyaogong 2025-12-31 10:26:46 +0800
  • bc6f6dea99 [feat] stream load data jingyaogong 2025-12-28 16:58:52 +0800
  • 4a5c9f5ece [feat] stream load data jingyaogong 2025-12-28 16:58:52 +0800
  • fc2eeff509 [feat] remove empty_cache jingyaogong 2025-12-27 07:14:36 +0800
  • 7eae14f3ce [feat] remove empty_cache jingyaogong 2025-12-27 07:14:36 +0800
  • daf6cc0c2e
         Apply suggestions from code review 李子浩 2025-12-24 16:14:18 +0800
  • 0afc6d6741 [fix] table flow jingyaogong 2025-12-24 13:43:10 +0800
  • 41ef5fd8b8 perf: merge LoRA weights into base model for inference whitesword 2025-12-23 20:55:52 +0800
  • 3d4f37bf61 start to pr whitesword 2025-12-23 20:14:53 +0800
  • 9faf7a6fea feat: implement merge_lora function whitesword 2025-12-23 20:10:01 +0800
  • 658ddcae3a [feat] explicit left padding jingyaogong 2025-12-23 18:59:48 +0800
  • 11b962da06 [feat] explicit left padding jingyaogong 2025-12-23 18:59:48 +0800
  • 4cbca65883 [fix] lora weight jingyaogong 2025-12-22 21:27:29 +0800
  • a9c56b20e9 [fix] lora weight jingyaogong 2025-12-22 21:27:29 +0800
  • 0664b383e4 Merge pull request #594 from whiteswordLI/fix/lora-load-ddp-weights jingyaogong 2025-12-22 21:19:16 +0800
  • 048d84abc7
         Merge pull request #594 from whiteswordLI/fix/lora-load-ddp-weights jingyaogong 2025-12-22 21:19:16 +0800
  • 65b648be32 Fix: support loading DDP-saved LoRA weights for inference whitesword 2025-12-22 20:50:25 +0800
  • 3a18fdd666 Fix: support loading DDP-saved LoRA weights for inference whitesword 2025-12-22 20:50:25 +0800
  • 3692d609be refactor: optimize tensor wrapping in lm_dataset.py dsq 2025-12-20 15:09:26 +0800
  • b3069d4743
         Delete README_web.md yuyu5333 2025-12-17 16:52:43 +0800
  • d137420466
         Merge pull request #1 from Litmeb/codex/modify-lora-code-for-rsvd Ltimbe 2025-12-16 19:37:14 +0800
  • bae81a2ce9 Add rSVD-based adaptive LoRA rank estimation Ltimbe 2025-12-16 19:36:14 +0800
  • 4fd7b5a65b 添加数据集加载逻辑、网页内容抓取与多卡并行模型推理 DirakeasSea 2025-12-16 01:59:49 +0000
  • e988830b31
         Merge 7de4b9ecd6 into fe24501602 yuyu5333 2025-12-15 15:43:10 +0800
  • dacd4e5121 [feat] adjust seq length jingyaogong 2025-12-14 20:41:58 +0800
  • fe24501602 [feat] adjust seq length jingyaogong 2025-12-14 20:41:58 +0800
  • 05616a8bd1 Feat: Introduce Gated Attention to improve model performance wanghao 2025-12-12 21:32:40 +0800
  • 66263101b9 feat: Add LoRA alpha scaling factor for improved parameter efficiency and stability wanghao 2025-12-12 11:30:47 +0800
  • 02ae87113e
         Merge 717d465689 into fa82707c9c VinodHatti_ 2025-12-11 23:17:51 +0530
  • b813b562e2 [feat] update readme jingyaogong 2025-12-11 15:45:50 +0800
  • fa82707c9c [feat] update readme jingyaogong 2025-12-11 15:45:50 +0800
  • 37300329c5 [fix] dtype & lr jingyaogong 2025-12-09 13:01:38 +0800
  • 5129f0e2a2 [fix] dtype & lr jingyaogong 2025-12-09 13:01:38 +0800
  • b6b0c8f675 Merge pull request #571 from dyhuachi/dyhuachi-patch-1 jingyaogong 2025-12-09 12:59:11 +0800
  • aa7dc0f61e
         Merge pull request #571 from dyhuachi/dyhuachi-patch-1 jingyaogong 2025-12-09 12:59:11 +0800
  • 511b317511
         Merge b7f58ce95e into ecd1ae1563 dyhuachi 2025-12-08 16:30:15 +0800
  • b7f58ce95e
         Merge branch 'jingyaogong:master' into feat-merge-lora-model dyhuachi 2025-12-08 16:28:34 +0800
  • c37d924b47 [fix]:Fixed the gradient explosion problem that occurred during pre-training of the MOE module. dieu 2025-12-08 15:25:34 +0800
  • 8fcc81696d [fix] Refactor get_lr function to include min_lr calculation dyhuachi 2025-12-06 17:09:51 +0800
  • bf3878ace8
         [fix] Refactor get_lr function to include min_lr calculation dyhuachi 2025-12-06 17:09:51 +0800
  • 7614a3bf90 [fix] reduce aux_loss_alpha jingyaogong 2025-12-05 23:08:29 +0800
  • ecd1ae1563 [fix] reduce aux_loss_alpha jingyaogong 2025-12-05 23:08:29 +0800
  • be979ec9e7
         Modify LoRA application conditions in model_lora.py Ltimbe 2025-12-05 21:13:42 +0800
  • cc29d9a351 feat: add merge_lora.py to support merging LoRA weights into base model dieu 2025-12-05 13:55:20 +0800
  • 95bcb7ba19 [fix] cuda memory #559 jingyaogong 2025-12-01 16:17:43 +0800
  • 5e1447b913 [fix] cuda memory #559 jingyaogong 2025-12-01 16:17:43 +0800
  • 295b9ad443 [feat] update yarn jingyaogong 2025-12-01 16:15:05 +0800
  • 151fdf7e76 [feat] update yarn jingyaogong 2025-12-01 16:15:05 +0800
  • cafc3b22df [feat] release memory jingyaogong 2025-11-27 19:39:49 +0800
  • 6b86ea399a [feat] release memory jingyaogong 2025-11-27 19:39:49 +0800
  • 0ebf835006 update start bash yuyu5333 2025-11-14 07:40:47 +0000
  • aee1913d3d update para from_resume yuyu5333 2025-11-25 03:28:56 +0000
  • d9dddccd71 update process step wangyuzhan 2025-11-21 18:43:41 +0800
  • 826a1fbe1e add process step wangyuzhan 2025-11-21 17:44:24 +0800
  • b336bc6008 remove useless file yuyu5333 2025-11-21 17:22:33 +0800
  • 4b1bd00eaa 第一次尝试 wzyandyzw 2025-11-21 17:14:12 +0800
  • 5e5b1bedb0 update web ui yuyu5333 2025-11-21 15:34:01 +0800
  • 80f75a252b update log flash wangyuzhan 2025-11-21 15:11:11 +0800
  • dfe1e5c6c7 update web dataset file wangyuzhan 2025-11-21 14:11:37 +0800
  • 8845d9170c update web dataset file wangyuzhan 2025-11-21 12:33:04 +0800
  • 249d1c08fe update web dataset file wangyuzhan 2025-11-21 12:09:05 +0800
  • a0013eae8f update web dataset file wangyuzhan 2025-11-21 10:41:05 +0800
  • 4a741b3d3b update web dataset file wangyuzhan 2025-11-20 21:55:48 +0800
  • fc1f07bf74 remove sdk wangyuzhan 2025-11-20 21:19:28 +0800
  • 102a7c0199 update web ui wangyuzhan 2025-11-20 20:47:00 +0800
  • a95568644b update import web wangyuzhan 2025-11-20 17:21:10 +0800
  • a794898a18 update check web server health wangyuzhan 2025-11-20 15:58:31 +0800
  • 25cf74eec1 update sdk wangyuzhan 2025-11-20 15:54:08 +0800
  • 3382c90246 [fix] ppo mask jingyaogong 2025-11-19 23:39:02 +0800
  • d7f4f4eab8 [fix] ppo mask jingyaogong 2025-11-19 23:39:02 +0800
  • 21cadea5f5 [fix] model attn_mask jingyaogong 2025-11-19 22:26:53 +0800
  • f5374dc87f [fix] model attn_mask jingyaogong 2025-11-19 22:26:53 +0800
  • 63d50b8778 fix: resolve high vulnerability trailofbits.python.pickles-in-pytorch.pickles-in-pytorch orbisai0security 2025-11-19 06:08:46 +0000
  • 66dcb40db7 init sdk wangyuzhan 2025-11-18 21:35:26 +0800
  • 52b7b88ec2 update web wangyuzhan 2025-11-18 21:21:13 +0800
  • 7aef898647 [fix] update model jingyaogong 2025-11-18 13:07:20 +0800
  • a044578d73 [fix] update model jingyaogong 2025-11-18 13:07:20 +0800
  • 22f82be681 Merge pull request #536 from yuyu5333/fix/attn_forward jingyaogong 2025-11-18 13:02:46 +0800
  • ce9394670b
         Merge pull request #536 from yuyu5333/fix/attn_forward jingyaogong 2025-11-18 13:02:46 +0800
  • 9c29046d2e fix: attn_forwad when is_causal=True assert attn_mask is None yuyu5333 2025-11-18 03:17:17 +0000
  • 7d02ce673c fix: attn_forwad when is_causal=True assert attn_mask is None yuyu5333 2025-11-18 03:17:17 +0000
  • 49bd32f168 support rl-grpo, rl-spo yuyu5333 2025-11-17 17:46:19 +0800
  • 58c39a020c [feat] Support Minimind retrieval-augmented generation (RAG) ztzhu1 2025-11-17 00:08:37 +0800
  • c8727733f3 [fix] prompt length calculate jingyaogong 2025-11-15 18:25:37 +0800
  • 9c98cabc9a [fix] prompt length calculate jingyaogong 2025-11-15 18:25:37 +0800
  • 544c889533 rewrite web yuyu5333 2025-11-14 16:08:30 +0800
  • ebf1a853ab update bash start yuyu5333 2025-11-14 15:45:36 +0800
  • c0b39e2396 rewrite web wangyuzhan 2025-11-14 14:21:14 +0800
  • 7c947e59c1 update http && process logs yuyu5333 2025-11-13 07:01:05 +0000
  • 3a03cedd00 support Swanlab check yuyu5333 2025-11-11 13:26:34 +0000
  • 04477b72f2 update safe web yuyu5333 2025-11-11 07:43:17 +0000
  • d66a7945db support ppo and Training web code refactoring yuyu5333 2025-11-11 07:40:38 +0000
  • 3f316cb2c5
         Merge d4f2339c1b into f3441b0078 Wenjie Dong 2025-11-10 22:13:40 +0530