---
title: DeepSeek-V4-Pro-NVFP4-FP8
canonical_url: "https://www.modelscope.ai/models/RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8"
md_url: "https://www.modelscope.ai/models/RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8.md"
repository: RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8
last_updated: 2026-08-29
license: mit
model_type:
  - deepseek_v4
architectures:
  - DeepseekV4ForCausalLM
base_model:
  - deepseek-ai/DeepSeek-V4-Pro
base_model_relation: quantized
parameters: 894.3B
tensor_type:
  - I64
  - U8
  - BF16
  - F32
  - F8_E4M3
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 76
stars: 0
tags:
  - compressed-tensors
  - vLLM
---

# DeepSeek-V4-Pro-NVFP4-FP8

> DeepSeek-V4-Pro-NVFP4-FP8 - An open-source model by RedHatAI on ModelScope. RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8

RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8 is a 894.3B-parameter machine learning model on ModelScope. licensed under mit. derived from deepseek-ai/DeepSeek-V4-Pro.

- **Repository**: RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8
- **License**: mit
- **Parameters**: 894.3B
- **Base model**: deepseek-ai/DeepSeek-V4-Pro
- **Tags**: compressed-tensors, vLLM
- **Downloads**: 76
- **Stars**: 0
- **Last updated**: 2026-08-29

Source: https://www.modelscope.ai/models/RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8

---

# RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8

This is a quantized version of `deepseek-ai/DeepSeek-V4-Pro` with MoE layers quantized to NVFP4 and attention layers quantized to FP8 block

## Usage

This model is intended for deployment with vLLM. You can serve the model using

```bash
vllm serve RedHatAI/DeepSeek-V4-Pro-NVFP4-FP8-BLOCK --tensor_parallel_size 8 --kv_cache_dtype=fp8
```

## Creation Process

This model was created using [LLM Compressor](https://github.com/vllm-project/llm-compressor). The example script can be found in `examples/quantizing_moe/deepseek_v4_pro_example.py` [[DSV4] DeepSeekV4 Pro](https://github.com/vllm-project/llm-compressor/pull/2858). Quantizing the model with data parallelism and 6xA100 takes about 3 hours.