Authors: Dhruv Nathawani, Shuoyang Ding US, Vitaly Lavrukhin US, Jane Polak Scowcroft US, Oleksii Kuchaiev US.

NVIDIA continues releasing permissive datasets in support of the open ecosystem with 6 Million Multilingual Reasoning Dataset.

Continuing the success of the recent Nemotron Post-Training Dataset v1 release used in Llama Nemotron Super model, and our Llama Nemotron Post-Training Dataset release earlier this year, we’re excited to release the reasoning dataset translated into five target languages: French, Spanish, German, Italian, and Japanese.

The newly released NVIDIA Nemotron Nano 2 9B brings these capabilities to the edge with leading accuracy and efficiency with a hybrid Transformer–Mamba architecture and a configurable thinking budget—so you can dial accuracy, throughput, and cost to match your real‑world needs.

The release represents a significant step forward in our continued commitment to openness and transparency in model development and improvement. By releasing training data, in addition to the training tools and final model weights, NVIDIA supports continued improvement of open‑weight models.

At a high level, the Nemotron Post-Training Dataset V2 takes our previously released English reasoning data and translates them into five target languages (French, German, Italian, Japanese, Spanish). To best take advantage of English knowledge instilled during pre‑training, we translate the user prompt and model response while preserving the original English reasoning chain.

In a recent development, according to results from the WMT 2024 general translation shared task, LLMs are achieving state‑of‑the‑art results for machine translation tasks. However, for synthetic generation of post‑training data, our preliminary studies have shown that:.

It has been revealed that hence, we incorporate several mechanisms to maintain high translation quality and easy hallucination detection. To summarize:.

Table 1: Ratio of discarded data (measured by bytes) by enforcing output format.

After benchmarking, we selected Qwen2.5-32B-Instruct-AWQ (for German) and Qwen2.5-14B-Instruct (for others) to conduct the translation. The considerations for selecting these models include:.

👉 Explore the dataset here: Hugging Face dataset page.

→ Read original article