bartowski/nvidia_Nemotron-H-47B-Reasoning-128K-GGUF

Same weights, converted vs nvidia/Nemotron-H-47B-Reasoning-128K. Same weights stored in another dtype (BF16 → BF16/F32), plus 45 new tensors.

Tensors changed: Embeddings 0/1, Norms 0/99, lm_head 0/1

On the Hub · Downloads on Model Pulse