llmfan46/Qwen3.5-9B-Nikusui-v1-Uncensored-Heretic-Native-MTP-Preserved-GGUF

Same weights, converted vs llmfan46/Qwen3.5-9B-Nikusui-v1-Uncensored-Heretic-Native-MTP-Preserved. Same weights stored in another dtype (BF16 → BF16/F32), plus 39 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