BytedanceDouyinContent/SAIL-VL-1d6-8B
Different architecture vs Qwen/Qwen2.5-1.5B-Instruct. Different architecture: only 0 of 338/518 tensors match by name and shape.
Different architecture vs Qwen/Qwen2.5-1.5B-Instruct. Different architecture: only 0 of 338/518 tensors match by name and shape.