DevQuasar/nvidia.NVIDIA-Nemotron-Nano-12B-v2-Base-GGUF

Same weights, converted vs nvidia/NVIDIA-Nemotron-Nano-12B-v2-Base. Same weights stored in another dtype (BF16 → F16/F32), plus 28 new tensors.

Tensors changed: Embeddings 0/1, Norms 0/63, lm_head 0/1

On the Hub · Downloads on Model Pulse