evansanih/Qwen2.5-Coder-14B-n8n-Workflow-Generator
Adapter vs Qwen/Qwen2.5-Coder-14B-Instruct. evansanih/Qwen2.5-Coder-14B-n8n-Workflow-Generator is a LoRA adapter, not full weights: rank 32, on down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj in 48 layers (336 adapted matrices). It modifies Qwen/Qwen2.5-Coder-14B-Instruct at load time.
6 other repos hold byte-for-byte identical weight files.