he pang

关于spearman评估方法的疑问

2

首先, Nice work! 我看评估的代码tests/model_spearman.py中, 在整个数据集上对pred和labels运行了spearmanr(x, y)方法, 数据集中的标签为1-5之间的整数, 而pred的余弦相似度为-1-1之间的连续浮点数. 由于spearman是按照秩次评估相关性的, 数据集中大量重复的label会不会导致大量等次的秩, 影响相关性的评估准确性? 一个想法, 是不是可以吧计算spearman相关性的步骤拆分成若干个小组, 每个小组里放6个label不相同的结果, 再与预测值计算相关度, 这样更能凸显出文本相似度之间的对比关系. 谢谢作者

Got Issues when `pip install -r requirements.txt`

1

Outputs: ~~~ Collecting fastapi==0.86.0 (from -r requirements.txt (line 1)) Downloading fastapi-0.86.0-py3-none-any.whl (55 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 55.5/55.5 kB 1.1 MB/s eta 0:00:00 Collecting numpy==1.22.3 (from -r requirements.txt (line 2)) Downloading numpy-1.22.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (16.8...

### Do you need to file an issue? - [X] I have searched the existing issues and this feature is not already filed. - [ ] My model is hosted...

作者您好，当运行下面的命令启动基于qwen1.5-1.8b的moe模型时，forward过程的一个assert报错导致进程退出。 ``` python ming/serve/cli.py \ --model_base /root/autodl-tmp/qwen/qwen1.5-1.8b-chat \ --model_path /root/autodl-tmp/qwen/ming-moe \ --conv_template qwen \ --max_new_token 128 \ --beam_size 3 \ --temperature 1.2 ``` ```trace File "/root/autodl-tmp/lora-moe/original-ming-git/MING/ming/model/utils.py", line 231, in forward...

migrate qwen3omni to MLX

1

## Add Support for Qwen/Qwen3-Omni-30B-A3B-Instruct This PR adds support for the Qwen3-Omni-30B-A3B-Instruct model, enabling multi-modal understanding and audio generation capabilities. The model supports text, image, and audio inputs, and can...