NVIDIA DeepSeek-V4.1-Flash-1017-NVFP4 Model Released
Original title: nvidia/DeepSeek-V4.1-Flash-1017-NVFP4 — new model trending #30 on Hugging Face
The quantized NVIDIA DeepSeek-V4.1-Flash-1017 NVFP4 model is now trending on Hugging Face with support for up to 1 million token contexts.
NVIDIA released the DeepSeek-V4.1-Flash-1017 NVFP4 quantized model on Hugging Face. The natively multimodal Mixture-of-Experts model uses a Causal Encoder-Decoder architecture with Compressed Sparse Attention 2 and supports contexts up to one million tokens. It is optimized for NVIDIA GPU systems with vLLM and runs on Blackwell hardware.
- Quantized with Model Optimizer v0.48.0rc0
- 8B activated during prefill, 16B during decode
- Supports up to 1M token context
- Compatible with vLLM on NVIDIA Blackwell
- Vendors
- NVIDIADeepSeek AI
- AI models
- DeepSeek-V4.1-Flash-1017-NVFP4
From the original article
5 min read at huggingface.coimage-text-to-text · 3,648 downloads · 167 likes · license mit
By nvidia · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.