MiniCPM

MiniCPM copied to clipboard

[Bad Case]: 安卓部署问题

1

### Description / 描述 在使用powerinfer部署时，执行pip install -r requirements.txt时，显示安卓不能安装torch ### Case Explaination / 案例解释 _No response_

No3cat

### Feature request / 功能建议 你好，不太清楚在哪里修改。。 full log： ``` Traceback (most recent call last): File "/root/mambaforge/envs/habitat-sim/lib/python3.9/site-packages/urllib3/connection.py", line 198, in _new_conn sock = connection.create_connection( File "/root/mambaforge/envs/habitat-sim/lib/python3.9/site-packages/urllib3/util/connection.py", line 60, in create_connection for...

Yonggie

[Bad Case]: how can i use model.forward to inference model?

1

### Description / 描述 i want to use model(data) to inference model instead of model.chat(input) to inference. how can i prepared the data?? ### Case Explaination / 案例解释 i can...

ziyanxzy

[Bug]: Android 版MiniCPM-V 2.0初始化之后一直闪退

1

### Is there an existing issue ? / 是否已有相关的 issue ? - [X] I have searched, and there is no existing issue. / 我已经搜索过了，没有相关的 issue。 ### Describe the bug /...

leevin101

您好，我在使用该框架测试gsm8k时遇到了复现不一致的问题。 用minicpm-2b-sft-bf16模型在gsm8k任务上测的值只有38.13， ``` "overall_result": { "accuracy": 0.3813495072024261 } ``` 使用的配置参数如下： ``` { "task_name": "gsm8k_gsm8k_gen", "path": "datasets/gsm8k/data/gsm8k.jsonl", "description": "", "transform": "datasets/gsm8k/transform_gen_v0.py", "fewshot": 8, "batch_size": 1, "generate": { "method": "generate", "params": "models/model_params/vllm_sample_v1.json",...

Abigail61

[Feature Request]: 哭着求MiniCPM 4和Ollama的集成？

18

### Feature request / 功能建议 [MiniCPM](https://github.com/OpenBMB/MiniCPM)真的是好用的感动的我要哭！！强烈请求尽快出Ollama模型，最好是能在ollama官网上上线，兼容最新版ollama。先谢过了哪位大神了~

fishfree

生成速度比Qwen2 7B还慢

5

测试了一下在V100上速度较慢，这是为何

lucasjinreal

[Feature Request]: emebdding&reranker服务问题

1

### Feature request / 功能建议 请问什么框架可以提供mini-embedding和mini-reranker的推理服务，transformers推理好像不能直接支持异步

lyj157175

[Bug]: openai_api_server_demo.generate_minicpm:158 - Input length larger than 20

2

### Is there an existing issue ? / 是否已有相关的 issue ? - [X] I have searched, and there is no existing issue. / 我已经搜索过了，没有相关的 issue。 ### Describe the bug /...

cxgreat2014

[Feature Request]: 请教关于显存和LLM x MapReduce的问题

5

### Feature request / 功能建议 我用vllm进行部署,命令如下 ```bash vllm serve /hestia/model/MiniCPM3-4B --trust-remote-code --max-model-len 12288 --num-gpu-blocks-override 768 --port 8001 --max-num-seqs 32 --served-model-name minicpm --swap-space 0 ``` 12288的上下文长度就消耗了22G的显存，我看readme里提到了LLM x MapReduce可以低显存处理无限上下文，请问要如何开启

wciq1208