unsloth/Qwen2.5-VL-32B-Instruct
Same weights, converted vs Qwen/Qwen2.5-VL-32B-Instruct. Same weights stored in another dtype (BF16/F32 → BF16).
Tensors changed: Attention 0/576, MLP 0/388, Norms 0/194, Embeddings 0/2, lm_head 0/1
Same weights, converted vs Qwen/Qwen2.5-VL-32B-Instruct. Same weights stored in another dtype (BF16/F32 → BF16).
Tensors changed: Attention 0/576, MLP 0/388, Norms 0/194, Embeddings 0/2, lm_head 0/1