llmfan46/MiniMax-M2.7-ultra-uncensored-heretic-GGUF

Same weights, converted vs llmfan46/MiniMax-M2.7-BF16-ultra-uncensored-heretic. Same weights stored in another dtype (BF16 → BF16/F32), plus 248 new tensors.

Tensors changed: Embeddings 0/1, other 0/62, Norms 0/249, Attention 0/248, lm_head 0/1

On the Hub · Downloads on Model Pulse