LLM EXPLORER 56,604 MODELS INDEXED

DeepSeek V4 Flash NVFP4 by amd

By amd · 40 downloads
HF Repo ↗ Copy URL

DeepSeek V4 Flash NVFP4 is an open-source language model by amd. Features: 149.2b LLM, VRAM: 153.1GB, Context: 1024K, License: mit, LLM Explorer Score: 0.29.

  8-bit   Amd-quark Base model:deepseek-ai/deepsee... Base model:quantized:deepseek-...   Deepseek   Deepseek v4   En   Endpoints compatible   Moe   Nvfp4   Ocp-mx   Quantization   Quark   Region:us   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/amd/DeepSeek-V4-Flash-NVFP4 

DeepSeek V4 Flash NVFP4 Parameters and Internals

LLM NameDeepSeek V4 Flash NVFP4
Repository 🤗https://huggingface.co/amd/DeepSeek-V4-Flash-NVFP4 
Base Model(s)  DeepSeek V4 Flash   deepseek-ai/DeepSeek-V4-Flash
Model Size149.2b
Required VRAM153.1 GB
Updated2026-08-08
Maintaineramd
Model Typedeepseek_v4
Model Files  0.0 GB   1.1 GB: 1-of-46   3.8 GB: 2-of-46   3.8 GB: 3-of-46   3.8 GB: 4-of-46   3.8 GB: 5-of-46   3.8 GB: 6-of-46   3.8 GB: 7-of-46   3.8 GB: 8-of-46   3.8 GB: 9-of-46   3.8 GB: 10-of-46   3.8 GB: 11-of-46   3.8 GB: 12-of-46   3.8 GB: 13-of-46   3.8 GB: 14-of-46   3.8 GB: 15-of-46   3.8 GB: 16-of-46   3.8 GB: 17-of-46   3.8 GB: 18-of-46   3.8 GB: 19-of-46   3.8 GB: 20-of-46   3.8 GB: 21-of-46   3.8 GB: 22-of-46   3.8 GB: 23-of-46   3.8 GB: 24-of-46   3.8 GB: 25-of-46   3.8 GB: 26-of-46   3.8 GB: 27-of-46   3.8 GB: 28-of-46   3.8 GB: 29-of-46   3.8 GB: 30-of-46   3.8 GB: 31-of-46   3.8 GB: 32-of-46   3.8 GB: 33-of-46   3.8 GB: 34-of-46   3.8 GB: 35-of-46   3.8 GB: 36-of-46   3.8 GB: 37-of-46   3.8 GB: 38-of-46   3.8 GB: 39-of-46   3.8 GB: 40-of-46   3.8 GB: 41-of-46
Supported Languagesen zh
Model ArchitectureDeepseekV4ForCausalLM
Licensemit
Context Length1048576
Model Max Length1048576
Transformers Version4.57.1
Vocabulary Size129280
Torch Data Typebfloat16