TheCluster/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-MLX-bf16

Same weights, converted vs HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive. Same weights stored in another dtype (BF16/F32 → BF16/F32), plus 357 new tensors.

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/81, MLP 0/96, Attention 0/32

On the Hub · Downloads on Model Pulse