The post-training data has a cutoff date of November 28, 2025. The pre-training data has a cutoff date of June 25, 2025. NVIDIA-Nemotron-Labs-3-Elastic-30B-A3B-NVFP4 is a 3-in-1 elastic large language model (LLM) developed by NVIDIA. It contains three nested model variants (30B, 23B, and 12B parameters) within a single NVFP4 checkpoint, all sharing the same parameter space. The 23B and 12B variants can be extracted zero-shot from this checkpoint using the provided slicing script. This is the NVFP4 quantized version of the BF16 elastic model. Weights are quantized using NVIDIA’s FP4 format via Quantization-Aware Distillation (QAD) with the BF16 elastic model as teacher. The NVFP4 quantization preserves the nested weight-sharing structure, enabling zero-shot extraction of the 23B and 12B variants from this checkpoint. This model was derived from NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 using the Elastic post-training framework, which produces elastic (many-in-one) reasoning LLMs from hybrid Mamba-Transformer-MoE architectures. The method takes the parent reasoning model and embeds N nested submodels within it using the compute of a single training run. The technology was originally…
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: DemomanCA
Теги: gguf, nvidia, elastic, en, es, fr, de, ja
Лайков: 3 | Загрузок: 75
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
