Guide for fine-tuning NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4 to add new programming language support

Accelerated Computing DGX Spark / GB10 User Forum DGX Spark / GB10
1

Hi,

Q01: Do you think a dual cluster of DGX Sparks nodes would be sufficient to fine tune the NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4 model, to add support for a new programming language?

Q02: Are there any available guides or recommended resources for fine-tuning NVIDIA Nemotron-3 models?

2

  1. It’s possible with 256GB VRAM,

  2. NVIDIA Nemotron-3-Super: How To Run Guide | Unsloth Documentation
    NVIDIA Nemotron 3 Nano - How To Run Guide | Unsloth Documentation
    Nemotron-3-Super Fine-tuning with NeMo AutoModel — Nemotron