qingy2024/Qwen2.5-Coder-Draft-1.5B-Instruct

Qwen2.5-Coder-Draft-1.5B-Instruct

A draft model suitable for Qwen2.5 Coder 32B Instruct

It uses a vocabulary size of 152064, which is the same as Qwen2.5 Coder 32B Instruct (can be used in vLLM directly without any hack)