Abiray/Fibo-2.0-edit-turbo-GGUF

Fibo-2.0-edit-turbo GGUF

GGUF quantizations of briaai/Fibo-2.0-edit-turbo optimized for direct use in ComfyUI via ComfyUI-GGUF.

📦 Available Quantizations

Filename	Quantization	Size	Recommended Use
fibo2_edit_turbo-Q3_K_M.gguf	Q3_K_M	4 GB	Ultra low-VRAM setups
fibo2_edit_turbo-Q4_K_S.gguf	Q4_K_S	4.84 GB	Fast generation, low VRAM
fibo2_edit_turbo-Q4_K_M.gguf	Q4_K_M	5.28 GB	Best balance of speed and quality
fibo2_edit_turbo-Q5_K_M.gguf	Q5_K_M	6.12 GB	High fidelity generation
fibo2_edit_turbo-Q6_K.gguf	Q6_K	6.76 GB	Near-lossless quality
fibo2_edit_turbo-Q8_0.gguf	Q8_0	8.75 GB	Maximum precision quant

🛠️ Required Custom Nodes & Dependencies

To run Fibo 2.0 models in ComfyUI, you need:

1. ComfyUI-GGUF (placed in ComfyUI/custom_nodes/) to load the GGUF UNet weights.
2. ComfyUI Core Native Support: Ensure your ComfyUI installation supports Fibo 2.0 (such as via ComfyUI PR #17240 or equivalent native integration).

🚀 Setup & Usage in ComfyUI

1. Place your chosen GGUF file from this repository into your ComfyUI/models/unet/ folder.
2. Download and place the required text encoder (qwen3vl_4b_int8_convrot.safetensors) from the Comfy-Org/Fibo-2.0 Text Encoders repository into ComfyUI/models/text_encoders/.
3. Download and place the VAE (flux2-vae.safetensors) from the Comfy-Org/Fibo-2.0 VAE into ComfyUI/models/vae/.
4. Load the model using the UnetLoaderGGUF node coupled with CLIPLoader (type: fibo2) and TextEncodeFibo2.

📋 Example Workflow

You can download and import the ready-to-use ComfyUI workflow directly from this repository: 👉 Fibo2_turbo_edit_test_GGUF.json

📄 License & Attribution

These GGUF weights are derived from briaai/Fibo-2.0-edit-turbo. Please refer to the original model card for licensing terms and commercial use guidelines.