Lewdiculous/mini-magnum-12b-v1.1-GGUF-IQ-Imatrix
Same weights, converted vs intervitens/mini-magnum-12b-v1.1. 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
Same weights, converted vs intervitens/mini-magnum-12b-v1.1. 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