dangerkhan0/nsfwvision-qwen3-vl-8b-v3-safetensors
LoRA-style vs Qwen/Qwen3-VL-8B-Instruct. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, linear_fc1, linear_fc2, o_proj, proj, qkv changed, in 36 of 36 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/4, Norms 0/261, MLP 95/232, Attention 90/252, lm_head 0/1
2 derivatives checked
- mradermacher/nsfwvision-qwen3-vl-8b-v3-safetensors-i1-GGUF: Quantization (fewer layers)
- mradermacher/nsfwvision-qwen3-vl-8b-v3-safetensors-GGUF: Same weights, converted (fewer layers)
2 other repos hold byte-for-byte identical weight files.