26 Commits

Author SHA1 Message Date
zh-zheng
9ac6cf3e38 update readme and add more demos 2024-09-06 20:29:35 +08:00
zh-zheng
9a968f85fb Update repo for MiniCPM3 2024-09-05 17:41:40 +08:00
root
c95a1f1cb7 增加了package的版本 2024-07-16 20:34:43 +08:00
root
ba499a19ad 增加了所有argument的注释 2024-07-16 20:30:57 +08:00
root
319f2301a4 增加了langchain对本地MiniCPM的支持,增加了多文件小显存rag的demo 2024-07-16 17:27:39 +08:00
winter
a77a34759f use default gpu 0 2024-07-04 19:54:18 +08:00
cxcz
8c6d6f8615
Merge branch 'main' into add_cpmV_hfdemo 2024-06-28 18:13:19 +08:00
zR
3e7c74ec7a 128k模型无法推理的问题修复 2024-04-14 19:00:46 +08:00
zR
8272667430 更新VLLM的书写方式 2024-04-14 14:23:11 +08:00
zR
e58d99f8ca openai api需要的依赖,requirement推理依赖 2024-04-13 23:16:59 +08:00
zR
b9b53e2e19 OpenAI 推理简单应用
用的是Linux transformers载入推理的,只测试了常规对话,在跑的时候没有看到支持Function 功能就没写
2024-04-11 19:31:08 +08:00
zR
fdaab94f1e 更新训练集类型和推理fp16部分 2024-04-03 14:47:45 +08:00
winter
ce391fb099 add MiniCPMV in hf_demo 2024-03-27 16:30:09 +08:00
zR
9e1438682e mlx inference 2024-03-26 21:07:54 +08:00
Y.W. Fang
600e00dba3 add repetition_penalty and set opk=0 in hf-based demo 2024-02-05 21:55:22 +08:00
ywfang
74f5bd04b2
Merge pull request #16 from soulteary/fix/pad-token-id-warnning
fix: set pad token id #15; closes #15
2024-02-04 11:59:36 +08:00
ywfang
6eec0fa36b
Merge pull request #14 from soulteary/fix/user-warning-typedstorage
fix: ignore typedstorage deprecated message #13
2024-02-04 11:54:50 +08:00
Y.W. Fang
75b144b04d fix typo in demo 2024-02-02 21:28:14 +08:00
ywfang
d39a6a10bb
Merge branch 'main' into feat/allow-set-torch-dtype 2024-02-02 17:00:22 +08:00
Su Yang
d64fe362fc
feat: allow user set torch dtype 2024-02-02 14:46:41 +08:00
Su Yang
6c4bbad9ed
feat: allow user change the demo host and port 2024-02-02 14:32:09 +08:00
Su Yang
0dab357653
fix: set pad token id 2024-02-02 14:15:24 +08:00
Su Yang
5c43592042
fix: ignore typedstorage deprecated message 2024-02-02 14:02:20 +08:00
Y.W. Fang
6cfeec2e54 update demo with vllm 2024-02-01 13:13:27 +08:00
Y.W. Fang
9c0fb61c32 Optimize gradio-based demo 2024-01-31 15:56:26 +08:00
Y.W. Fang
0590944a92 Update Gradio-based demo 2024-01-30 15:46:26 +08:00