---
title: flux1-dev-bnb-nf4
canonical_url: "https://www.modelscope.cn/models/ai-modelscope/flux1-dev-bnb-nf4"
md_url: "https://www.modelscope.cn/models/ai-modelscope/flux1-dev-bnb-nf4.md"
repository: ai-modelscope/flux1-dev-bnb-nf4
last_updated: 2024-11-21
license: other
parameters: 22.2B
tensor_type:
  - U8
  - F16
  - F8_E4M3
  - BF16
  - F32
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 2475
stars: 3
---

# flux1-dev-bnb-nf4

> flux1-dev-bnb-nf4 - ai-modelscope 在 ModelScope 开源的模型。Main page: https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/981

ai-modelscope/flux1-dev-bnb-nf4 是 ModelScope 魔搭社区上的 22.2B 参数机器学习模型，采用 other 许可。

- **Repository**: ai-modelscope/flux1-dev-bnb-nf4
- **License**: other
- **Parameters**: 22.2B
- **Downloads**: 2475
- **Stars**: 3
- **Last updated**: 2024-11-21

Source: https://www.modelscope.cn/models/ai-modelscope/flux1-dev-bnb-nf4

---

Main page: https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/981

---

Update:

Always use V2 by default. 

V2 is quantized in a better way to turn off the second stage of double quant. 

V2 is 0.5 GB larger than the previous version, since the chunk 64 norm is now stored in full precision float32, making it much more precise than the previous version. Also, since V2 does not have second compression stage, it now has less computation overhead for on-the-fly decompression, making the inference a bit faster.

The only drawback of V2 is being 0.5 GB larger.

---

Main model in bnb-nf4 (v1 with chunk 64 norm in nf4, v2 with chunk 64 norm in float32)

T5xxl in fp8e4m3fn

CLIP-L in fp16

VAE in bf16