reneix

Results 1 comments of reneix

> You can try to use the qwen2.5-14b model after INT4 quantization to reduce the GPU memory. got, will have a try