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

Commit Graph

  • 4f963ff266 [feat] update eval jingyaogong 2025-10-18 00:24:36 +0800
  • 3812a16d20 [feat] update eval jingyaogong 2025-10-18 00:24:36 +0800
  • fd79201bd4 [feat] update requirements jingyaogong 2025-10-17 00:14:31 +0800
  • 0682134e21 [feat] update requirements jingyaogong 2025-10-17 00:14:31 +0800
  • 7497efec18 [fix] update model jingyaogong 2025-10-17 00:11:08 +0800
  • b8adffbe89 [fix] update model jingyaogong 2025-10-17 00:11:08 +0800
  • 8bb98c9790 [fix] update model jingyaogong 2025-10-17 00:09:32 +0800
  • 4e35fb9da8 [fix] update model jingyaogong 2025-10-17 00:09:32 +0800
  • 8bb29e5805 update app jingyaogong 2025-10-16 23:33:35 +0800
  • 36159fb2ab update app jingyaogong 2025-10-16 23:33:35 +0800
  • c002418c10 fix: semgrep_trailofbits.python.pickles-in-pytorch.pickles-in-pytorch_trainer/train_pretrain.py_93 orbisai-sec 2025-10-16 06:06:39 +0000
  • 209c0caaa7 fix cdn jingyaogong 2025-10-12 23:33:42 +0800
  • 941db7a5e6 fix i18n jingyaogong 2025-10-12 22:46:53 +0800
  • 3c5feebf3f Restore i18n suffix structure jingyaogong 2025-10-12 22:38:09 +0800
  • 2c0f4f64d2 Fix i18n configuration for ReadTheDocs jingyaogong 2025-10-12 22:37:23 +0800
  • 8f525fe949 fix en jingyaogong 2025-10-12 22:29:38 +0800
  • edc4580b57 add requirements jingyaogong 2025-10-12 22:22:45 +0800
  • 3639ad0df1 init minimind-docs jingyaogong 2025-10-12 22:14:17 +0800
  • 1e3734dd0d
         Create codeql.yml Tawns-lab 2025-10-10 05:33:15 -0400
  • 808d488653
         Create pyre.yml Tawns-lab 2025-10-10 05:32:56 -0400
  • fd6448ce61 feat: add explanation Zhenyu Wang 2025-08-15 14:45:05 +0800
  • c5d2e61e2e update zhiyuanRen 2025-06-29 00:00:37 +0800
  • 67debd3632 优化模型加载和训练过程，添加torch.compile以提升性能，调整训练参数 yangnianboy 2025-06-27 01:00:46 +0800
  • feba9a930e 删除多余的图像文件并调整训练脚本中的批量大小 yangnianboy 2025-06-27 01:00:33 +0800
  • 9df92b986c
         Add files via upload 紫苏 2025-05-28 17:50:08 +0800
  • ab170f57f3
         Delete train_pretrain.py 紫苏 2025-05-28 17:48:49 +0800
  • 3babaea826
         Delete train_full_sft.py 紫苏 2025-05-28 17:48:38 +0800
  • 89011ebb3a
         Add files via upload 紫苏 2025-05-28 17:42:44 +0800
  • cd10b371c8
         Merge 3cc6d0d074 into 561979c7e3 Lycoris 2025-05-26 21:05:28 +0800
  • 3cc6d0d074
         Update train_lora.py Lycoris 2025-05-26 18:30:46 +0800
  • 4a0975e94e 修改了一些bug和漏洞 math-zhuxy 2025-05-25 19:50:53 +0800
  • 60fa1cc635 done math-zhuxy 2025-05-25 18:46:42 +0800
  • 9e35e5348f new model_chat.py math-zhuxy 2025-05-25 16:14:33 +0800
  • ccdf6c576f Improve training performance with torch.compile and torch.amp Gouryella 2025-05-21 12:16:02 +0800
  • 872ab07b20 升腾NPU适配 z30075425 2025-05-16 15:09:50 +0800
  • 554c6d56e8 chore: update .gitignore 和 trainer 相关脚本，完善文档 0x0059 2025-05-08 10:24:01 +0800
  • 8d36134e24 docs: 为多个脚本添加详细的中文注释 0x0059 2025-05-05 22:25:02 +0800
  • f74117e5d9 docs(model): 为MiniMind模型添加详细代码注释 0x0059 2025-05-05 21:59:08 +0800
  • ec25af222f update torch version jingyaogong 2025-04-30 16:47:26 +0800
  • 561979c7e3 update torch version jingyaogong 2025-04-30 16:47:26 +0800
  • b19fa8099d [fix] 修改 serve_openai_api.py 的默认参数 screnwei 2025-04-30 15:27:39 +0800
  • de01fcc02d [add] 保持跟远程目录一致 screnwei 2025-04-30 15:25:26 +0800
  • a71a64e608
         Merge branch 'jingyaogong:master' into master kevin.ren 2025-04-30 11:33:32 +0800
  • 71baefa98a fix bugs jingyaogong 2025-04-29 20:45:39 +0800
  • caae54a89e fix bugs jingyaogong 2025-04-29 20:45:39 +0800
  • 375c22ec42 [fix] fix issues/382 修复scaled_dot_product_attention方法不支持cpu、mps的问题 screnwei 2025-04-29 17:36:30 +0800
  • 4a1a9bec8a
         Merge branch 'jingyaogong:master' into master kevin.ren 2025-04-29 10:52:18 +0800
  • 408babf307 [add] merge from jingyaogong screnwei 2025-04-28 18:04:55 +0800
  • 5343923e0f update lora jingyaogong 2025-04-27 15:45:06 +0800
  • 5ffde04b7c update lora jingyaogong 2025-04-27 15:45:06 +0800
  • 8739fc9839 update convert script jingyaogong 2025-04-27 15:44:40 +0800
  • 2e118e9d3d update convert script jingyaogong 2025-04-27 15:44:40 +0800
  • 7d81a0a6f0 fix bugs jingyaogong 2025-04-27 09:56:49 +0800
  • 29454c31af fix bugs jingyaogong 2025-04-27 09:56:49 +0800
  • 5f5bf34400
         Update eval_model.py Howard Su 2025-04-27 07:45:09 +0800
  • c8972f7e89 sft should use pretrain model Wenyu Zhang 2025-04-26 11:50:23 -0400
  • cd342a4249 update readme jingyaogong 2025-04-26 10:37:46 +0800
  • 765c79c7a2 update readme jingyaogong 2025-04-26 10:37:46 +0800
  • f0c2923fab update readme jingyaogong 2025-04-26 10:21:34 +0800
  • 00d145c481 update readme jingyaogong 2025-04-26 10:21:34 +0800
  • be00d302ee 250426 jingyaogong 2025-04-26 10:07:55 +0800
  • 274483cb1b 250426 jingyaogong 2025-04-26 10:07:55 +0800
  • e74843bd09 250426 jingyaogong 2025-04-26 10:05:47 +0800
  • a62faf34bd 250426 jingyaogong 2025-04-26 10:05:47 +0800
  • 220204b713 完善 README 中关于加载已有模型的说明 李学斌 2025-04-20 10:18:54 +0800
  • 3acda20524 update chat-openai-api jingyaogong 2025-04-18 12:43:57 +0800
  • 7da201a944 update chat-openai-api jingyaogong 2025-04-18 12:43:57 +0800
  • 126f6b036e 123 skaura 2025-04-17 22:04:51 +0800
  • 8495189bbd [add] 添加 gitignore文件 screnwei 2025-04-17 15:13:27 +0800
  • 8081504f05 update moe note jingyaogong 2025-04-09 17:38:31 +0800
  • d9453ed9a3 update moe note jingyaogong 2025-04-09 17:38:31 +0800
  • 85b26b433c update eval jingyaogong 2025-04-09 16:56:57 +0800
  • d503093ec4 update eval jingyaogong 2025-04-09 16:56:57 +0800
  • d2f969f306 fix top_p float bug jingyaogong 2025-04-09 16:52:20 +0800
  • 4a758564e4 fix top_p float bug jingyaogong 2025-04-09 16:52:20 +0800
  • 27c7bcc491 fix: 解决#329中的数据类型定义问题 cn-farmer 2025-04-09 09:56:49 +0800
  • 2b47673cdc
         chore: auto detect mps for pre train Wei Zhang 2025-04-05 20:59:50 +0800
  • c688c4fd79 update rlaif jingyaogong 2025-04-05 16:06:08 +0800
  • 4a7c1c49e8 update rlaif jingyaogong 2025-04-05 16:06:08 +0800
  • 5701ff1aaa update generate jingyaogong 2025-04-05 15:53:55 +0800
  • 9e67798397 update generate jingyaogong 2025-04-05 15:53:55 +0800
  • aab57d5dc5 add hidden state jingyaogong 2025-04-05 14:39:56 +0800
  • 399d526fbd add hidden state jingyaogong 2025-04-05 14:39:56 +0800
  • c37b851caa update inference jingyaogong 2025-04-05 12:04:38 +0800
  • 885661f47d update inference jingyaogong 2025-04-05 12:04:38 +0800
  • 85cac44a38 update inference jingyaogong 2025-04-05 12:03:04 +0800
  • ed01c5d84a update inference jingyaogong 2025-04-05 12:03:04 +0800
  • 844dfb7e3b update seed set jingyaogong 2025-04-04 11:39:41 +0800
  • 7fcc46b39a update seed set jingyaogong 2025-04-04 11:39:41 +0800
  • b1552809c8 update web_demo jingyaogong 2025-04-04 11:25:40 +0800
  • 08e9a22a25 update web_demo jingyaogong 2025-04-04 11:25:40 +0800
  • 6e72678589 Add Load ckpt LH-and-FPGA 2025-04-03 18:37:48 +0100
  • ad741354b0 fix ( model/dataset.py ) 在预训练阶段，pretrain_hq.jsonl文件中已经用和将句子分割，在dataset.py文件构建输入文本时，再次在头部和尾部添加了和，有重复。 cn-farmer 2025-04-02 10:37:58 +0800
  • 2d7bb67ce2 fix（model/dataset.py) 原始文件中已经包含了<s>和</s>句子分割符，无需在代码中重复添加 cn-farmer 2025-04-02 10:18:04 +0800
  • a4707f50c6 update dpo_loss jingyaogong 2025-04-01 17:32:50 +0800
  • 278ec760a1 update dpo_loss jingyaogong 2025-04-01 17:32:50 +0800
  • ec0cfdc7ca update structure jingyaogong 2025-04-01 16:21:15 +0800
  • 6c24a7a5e7 update structure image jingyaogong 2025-04-01 16:15:26 +0800
  • 4f95e23a98 update structure image jingyaogong 2025-04-01 16:15:26 +0800
  • 077e7c3d66 update structure image jingyaogong 2025-04-01 16:11:54 +0800