---

FLUX.1-dev-GGUF

Original Model

black-forest-labs/FLUX.1-dev

Run with sd-api-server

• sd-api-server version: 0.1.4

• Run as LlamaEdge service

  wasmedge --dir .:. sd-api-server.wasm \
    --model-name flux1-dev \
    --diffusion-model flux1-dev-Q4_0.gguf \
    --vae ae.safetensors \
    --clip-l clip_l.safetensors \
    --t5xxl t5xxl-Q8_0.gguf
  

  • Run with LoRA

    Assume that the LoRA model is located in the lora-models directory

    wasmedge --dir .:. \
      --dir lora-models:lora-models \
      sd-api-server.wasm \
      --model-name flux1-dev \
      --diffusion-model flux1-dev-Q4_0.gguf \
      --vae ae.safetensors \
      --clip-l clip_l.safetensors \
      --t5xxl t5xxl-Q8_0.gguf \
      --lora-model-dir lora-models
    

    For details, see https://github.com/LlamaEdge/sd-api-server/blob/main/examples/flux_with_lora.md

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
ae.safetensors	f32	32	335 MB
clip_l-Q8_0.gguf	Q8_0	8	131 MB
clip_l.safetensors	f16	16	246 MB
flux1-dev-Q2_K.gguf	Q2_K	2	4.15 GB
flux1-dev-Q3_K.gguf	Q3_K	3	5.35 GB
flux1-dev-Q4_0.gguf	Q4_0	4	6.93 GB
flux1-dev-Q4_1.gguf	Q4_1	4	7.67 GB
flux1-dev-Q4_K.gguf	Q4_K	4	6.93 GB
flux1-dev-Q5_0.gguf	Q5_0	5	8.40 GB
flux1-dev-Q5_1.gguf	Q5_1	5	9.14 GB
flux1-dev-Q8_0.gguf	Q8_0	8	12.6 GB
flux1-dev.safetensors	f16	16	23.8 GB
t5xxl-Q2_K.gguf	Q2_K	2	1.61 GB
t5xxl-Q3_K.gguf	Q3_K	3	2.10 GB
t5xxl-Q4_0.gguf	Q4_0	4	2.75 GB
t5xxl-Q4_1.gguf	Q4_0	4	3.06 GB
t5xxl-Q4_K.gguf	Q4_K	4	2.75 GB
t5xxl-Q5_0.gguf	Q5_0	5	3.36 GB
t5xxl-Q5_1.gguf	Q5_1	5	3.67 GB
t5xxl-Q8_0.gguf	Q8_0	8	5.20 GB
t5xxl_fp16.safetensors	f16	16	9.79 GB

Quantized with stable-diffusion.cpp master-e71ddce.