LiveBenchDetailedResults
收藏魔搭社区2025-11-12 更新2024-10-12 收录
下载链接：
https://modelscope.cn/datasets/lmms-lab/LiveBenchDetailedResults
下载链接
链接失效反馈官方服务：
资源简介：
## Upload Results to HuggingFace
1. Evaluate the model using [lmms-eval](https://github.com/EvolvingLMMs-Lab/lmms-eval).
2. Upload logs using [upload_results.py](https://huggingface.co/datasets/lmms-lab/LiveBenchDetailedResults/blob/main/upload_results.py).
### Usage
```sh
python upload_results.py -f <log_folder> -m <model_name> [-F]
```
`[-F]` means the script will automatically upload the results without human checking. Otherwise, the script will print the results and ask for confirmation before uploading.
Example:
```sh
python upload_results.py -f logs/0706_0959_model_outputs_gpt4v_model_args_c974bc -m gpt-4o -F
```
## 将结果上传至 HuggingFace
1. 使用[lmms-eval](https://github.com/EvolvingLMMs-Lab/lmms-eval)对模型进行评估。
2. 使用[upload_results.py](https://huggingface.co/datasets/lmms-lab/LiveBenchDetailedResults/blob/main/upload_results.py)上传日志文件。
### 使用方法
sh
python upload_results.py -f <日志文件夹路径> -m <模型名称> [-F]
其中`[-F]`参数表示脚本将自动上传结果，无需人工审核；若不添加该参数，脚本将先打印结果并在上传前征求人工确认。
### 示例
sh
python upload_results.py -f logs/0706_0959_model_outputs_gpt4v_model_args_c974bc -m gpt-4o -F
提供机构：
maas
创建时间：
2024-10-07
