--- base_model: - Qwen/Qwen2.5-Coder-1.5B-Instruct --- Use the latest transformers for this model.