mradermacher/Qwen3.5-4B-Uncensored-GGUF
Same weights, converted vs Felldude/Qwen3.5-4B-Uncensored. Same weights stored in another dtype (F32/BF16 → F16/F32), plus 39 new tensors.
Tensors changed: Norms 0/81, MLP 0/96, Attention 0/32, Embeddings 0/1
Same weights, converted vs Felldude/Qwen3.5-4B-Uncensored. Same weights stored in another dtype (F32/BF16 → F16/F32), plus 39 new tensors.
Tensors changed: Norms 0/81, MLP 0/96, Attention 0/32, Embeddings 0/1