Dataset · nvidia
NVIDIA Nemotron SFT dataset
nvidia/Nemotron-Post-Training-Dataset-v1
Large compilation of NVIDIA SFT data for math, code, STEM, reasoning, and tool calling.
What's inside
This is a compilation of SFT data for improving math, code, STEM, general reasoning, and tool calling in Llama-3.3-Nemotron-Super-49B-v1.5. The model is derived from Meta Llama-3.3-70B-Instruct. The data is text-based, ranging in volume from 10 to 100 million examples; the language is not specified in the card.
What you can do with it
- SFT for Llama-3.3-Nemotron-Super-49B
- Fine-tuning tool calling models
- Improving math/code/reasoning via SFT
- Size
- 10M–100M rows
- Downloads this month
- 38.9K
- Likes
- 199
Models trained on it
- nicosuter/Qwen3.8-27B-AWQImage text → text 436.5K
- nvidia/NVIDIA-Nemotron-Nano-9B-v2Text generation 204.4K
- nvidia/NVIDIA-Nemotron-Nano-9B-v2-FP8Text generation 149.7K
- nvidia/NVIDIA-Nemotron-Nano-9B-v2-NVFP4Text generation 49.8K
- nvidia/NVIDIA-Nemotron-Nano-12B-v2Text generation 3.1K
- mradermacher/GaMS-27B-Instruct-Nemotron-GGUF 2.1K
- cyankiwi/NVIDIA-Nemotron-Nano-12B-v2-AWQ-8bitText generation 1.7K
- cyankiwi/NVIDIA-Nemotron-Nano-12B-v2-AWQ-4bitText generation 1.5K
How to load it
from datasets import load_dataset
ds = load_dataset("nvidia/Nemotron-Post-Training-Dataset-v1")