Lewdiculous/Mistral-Nemo-12B-ArliAI-RPMax-v1.2-GGUF-IQ-ARM-Imatrix

Same weights, converted vs ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2. Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Embeddings 0/1, Norms 0/81, MLP 0/120, Attention 0/160, lm_head 0/1

On the Hub · Downloads on Model Pulse