mradermacher/Mistral-NeMo-Minitron-8B-Instruct-GGUF

Same weights, converted vs nvidia/Mistral-NeMo-Minitron-8B-Instruct. 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