王召德

后端可以考虑接入MNN

1

现在移动端部署接入了TFLite，需要实现Torch到TFLite的转换；建议后端接入MNN， MNN后端性能领先于TFLite, 且目前提供了torchscript -> mnn模型的转换工具。

Update README about ChatGLM-MNN.

1

ChatGLM-MNN提供了Android APP Demo，在3G内存下可以流程运行。

Mobile deploy of LLM project

2

Thanks your project~🚀 Perhaps deploying LLM (Language Learning Models) with a size of 1.8 to 2 billion parameters on mobile or edge devices will become the next hotspot. Here is...

How to specify the output dir ?

3

When I used this action build a iOS project using below code: ``` - uses: mxcl/xcodebuild@v1 with: platform: iOS action: build configuration: release working-directory: project/ios scheme: XXX ``` I get...

微信群，最新信息更新

63


Support Alibaba MNN export and predict.

23

Support Alibaba [MNN](https://github.com/alibaba/MNN) export and predict. Test code: ```python from ultralytics import YOLO model = YOLO("yolo11n.pt") model.export(format='mnn') mnn_model = YOLO("yolo11n.mnn") results = mnn_model('../resource/test.jpg') for result in results: result.show() # display...

[Bug] mmlu_pro结果正则提取出错

5

### 先决条件 - [x] 我已经搜索过 [问题](https://github.com/open-compass/opencompass/issues/) 和 [讨论](https://github.com/open-compass/opencompass/discussions) 但未得到预期的帮助。 - [x] 错误在 [最新版本](https://github.com/open-compass/opencompass) 中尚未被修复。 ### 问题类型 我正在使用官方支持的任务/模型/数据集进行评估。 ### 环境 None ### 重现问题 - 代码/配置示例 None ### 重现问题 - 命令或脚本 None...

add MNN to local-apps.

bugfix: bugfix of ui arg no_eagle3.

1

1. Bugfix of `no_eagle3` to `use_eagle3`. 2. Remove `CUDA_VISIBLE_DEVICES = [2, 3]`. 3. add `ds-llama-3` model type.