base_model: Qwen/Qwen2.5-Coder-7B gate_mode: random dtype: bfloat16 architecture: qwen experts: - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen shared_experts: - source_model: Qwen/Qwen2.5-Coder-7B architecture: qwen