Fibo 2.0 from BRIA AI now available as a ComfyUI-ready package
Comfy-Org has repackaged BRIA AI's Fibo 2.0 (dev) and Fibo 2.0 edit-turbo image models for ComfyUI. They come in bf16 and a space-saving int8 version. The weights are licensed for non-commercial use only.
What this is about
Comfy-Org, the team behind the ComfyUI image generation interface, has published the "Fibo-2.0" repository on Hugging Face. It isn't a new in-house development. It contains repackaged files of BRIA AI's image models: Fibo-2.0-dev and Fibo-2.0-edit-turbo. The files are prepared so you can load them directly in ComfyUI. At the time of capture, the repository had about 56,800 downloads and 136 likes.
The Model Card is brief. It gives no technical details on parameter count, benchmarks or training data, and points to BRIA AI's original repositories instead. So what we can say mostly concerns the files, the folder structure, the quantization and the license.
Which files are included
The package has three components, which go into different ComfyUI folders:
| Folder | File | Purpose |
|---|---|---|
| models/diffusion_models | fibo2_dev_bf16.safetensors, fibo2_dev_int8_convrot.safetensors | Main model "dev" |
| models/diffusion_models | fibo2_edit_turbo_bf16.safetensors, fibo2_edit_turbo_int8_convrot.safetensors | Editing model "edit-turbo" |
| models/text_encoders | qwen3vl_4b_bf16.safetensors, qwen3vl_4b_int8_convrot.safetensors | Text encoder (Qwen3-VL 4B) |
| models/vae | flux2-vae.safetensors | VAE (converts between images and the model's internal image space) |
The name "edit-turbo" suggests a model for editing images that is built for speed. That's inferred from the name; the Model Card doesn't explain it further.
What changed during repackaging
The Model Card lists the changes from the original:
- bf16: The transformer files split into shards in Diffusers format were merged into a single file, and the key names were adapted to ComfyUI's layout. The attention weights to_q/to_k/to_v were combined into "qkv". In cross-attention and the perceiver, the to_k/to_v weights were combined into "to_kv". The weight values themselves are unchanged (bf16, a 16-bit number format).
- int8_convrot: The same weights were quantized to 8 bits (quantization means numbers are stored more coarsely to save space). It uses a Hadamard rotation with per-channel scaling factors; the adaLN modulation layers stay in bf16. As a result, outputs differ slightly from the original.
- The w1/w3 weights were also fused into w13, and the feed-forward block's intermediate width was zero-padded to 9728.
License
The weights belong to BRIA AI and are released under CC BY-NC 4.0. You may use them for non-commercial purposes only. Commercial use requires a separate agreement with BRIA AI (see the pricing page on platform.bria.ai). The weights are provided "as is", without warranty.
Background: structured captions
The Model Card cites the paper "Generating an Image From 1,000 Words: Enhancing Text-To-Image With Structured Captions" (arXiv:2511.06876). The title suggests that Fibo is built on very detailed, structured image descriptions. The card gives no details on this.
What you need to try it
The source doesn't state exact requirements. As a rule of thumb: a current ComfyUI installation with an Nvidia GPU makes sense, and the bf16 files need considerably more memory than the int8 variants. Because the parameter count is unknown, the VRAM requirement can't be reliably quantified. Expect to need several gigabytes of disk space for the model, text encoder and VAE. Put the files in the folders above and load them in ComfyUI.
Our take
For ComfyUI users, the package is mainly handy because no conversion is needed and the int8 variants save memory. The download count shows interest. How well Fibo 2.0 stacks up against other image models remains open, as the source offers no benchmarks. The non-commercial license sharply limits use for companies and freelancers. Also, int8 quantization slightly changes the image results.
Sources
- Hugging Face Trending huggingface.co
- Hugging Face Trending huggingface.co
- Google News (DE) news.google.com
Try it yourself: Fibo 2.0
Explained step by step, no prior knowledge needed – on your own computer or free in your browser.
What can I do with it?
Fibo 2.0 generates images from text descriptions, and the edit-turbo variant appears to be meant for editing images. You run it locally in ComfyUI.
Who is it worth it for?
For ComfyUI users who experiment privately or for research. Commercial projects require a license from BRIA AI.
What your computer needs
The int8 files save memory and are a better fit for smaller GPUs, but differ slightly from the original. The bf16 files are faithful to the original but need more VRAM. The source gives no concrete sizes.
- The source gives no file sizes or VRAM requirements; int8 needs less than bf16.
- You also need the text encoder (Qwen3-VL 4B) and flux2-vae.
- You'll likely need a recent ComfyUI version with Fibo support.
How do I find out what my computer has?
- Windows: Press Ctrl + Shift + Esc → Task Manager → Performance tab → click GPU on the left. At the bottom you'll see Dedicated GPU memory (that's your VRAM). You'll find system memory under Memory.
- Mac: Apple menu at the top left → About This Mac. You'll see your chip (e.g. M2) and memory (e.g. 16 GB). On Apple chips, graphics and processor share this memory.
- Linux: Type
nvidia-smiin the terminal (shows your graphics card and its memory) andfree -hfor system memory.
Guide
Google Colab (in your browser, no installation)
Google Colab is a free service from Google. It gives you a computer with a graphics card in your browser for a few hours. All you need is a Google account. For this model: Colab Pro only – Free Colab has only 12.7 GB of RAM – not enough for this model.
- Download the notebook. Click the button and save the file.
- Open Colab. Go to colab.research.google.com and sign in with your Google account.
- Upload the notebook. In the window that appears, click Upload on the left (or choose File → Upload notebook at the top) and select the file
colab.ipynb. - Turn on the GPU. In the top menu, go to Runtime → Change runtime type → under Hardware accelerator, select T4 GPU (with Colab Pro: L4 or A100) → Save.
- Start everything. Runtime → Run all (keyboard shortcut Ctrl + F9). If you see the warning “This notebook was not authored by Google,” click Run anyway.
- Wait. The cells run one after another. A green checkmark appears to the left of each cell when it's done. Installation and download take about 10 minutes here.
- Try it out. The result appears below the last cell. To test something else, change the text in the cell and run it again with Shift + Enter.
- Clean up. When you're done: Runtime → Disconnect and delete runtime. This saves your free quota.
Good to know: The free GPU is limited (usually a few hours per day). After about 90 minutes of inactivity, Colab disconnects, and you'll have to start everything over.
Windows 10 / 11
This model runs on Python. Our script handles the setup automatically: it creates an isolated Python environment, installs the required packages (PyTorch, approx. 3 GB) and downloads the model (16 GB (about 22 min at 100 Mbit/s)).
- Update your graphics driver (NVIDIA): install the latest driver. Without an NVIDIA card, it runs on the CPU – much more slowly.
- Install Python. Download Python 3.12 from python.org and run the installer. Important: At the bottom of the first window, tick Add python.exe to PATH, then click Install Now.
- Create a folder and download the files. Create a folder, e.g.
Documents\AI-Test, and save both files into it: - Open a terminal in the folder. Open the folder in File Explorer, right-click an empty spot → Open in Terminal.
- Run it. Paste this command and press Enter:
Installation and download are running now. The first time, this can take 10–30 minutes.powershell -ExecutionPolicy Bypass -File try_windows.ps1 - Result. ComfyUI is installed and the model is downloaded. Start ComfyUI (the command is shown at the end in the window) and open
http://127.0.0.1:8188in your browser. - Try your own input: Open
demo.pyin an editor, change the sample text, and save. Then run this in the terminal:.\.venv-fibo-2-0\Scripts\Activate.ps1 python demo.py
macOS
Macs with an Apple chip (M1 to M5) are ideal. Older Intel Macs are very slow.
- Open Terminal. Press Cmd + Space, type Terminal, and press Enter.
- Install the developer tools (first time only; includes Python):
Click Install in the window. If you see "already installed," everything is already in place.xcode-select --install - Download the files – both end up in your Downloads folder:
- Run it:
The first time, installation and download take 10–30 minutes.cd ~/Downloads bash try_mac.sh - Result. ComfyUI is installed and the model is downloaded. Start ComfyUI (the command is shown at the end in the window) and open
http://127.0.0.1:8188in your browser.
Linux
- Check your GPU.
nvidia-smishould show your card. If it doesn't, install the driver on Ubuntu and restart:sudo ubuntu-drivers install - Install the Python basics (Ubuntu/Debian):
sudo apt update && sudo apt install -y python3-venv python3-pip git - Put the files in a folder and open a terminal there:
- Run it:
bash try_linux.sh - Result. ComfyUI is installed and the model is downloaded. Start ComfyUI (the command is shown at the end in the window) and open
http://127.0.0.1:8188in your browser.
Try it
A cozy café in the rain, window view of a wet street, warm light, photorealisticA fox in a space suit on a Martian plain, watercolor-style illustrationChange the jacket of the person in the image to red (edit-turbo)If something goes wrong
Files must go in the right folders: diffusion_models, text_encoders and vae.
The workflow won't run without the matching text encoder and VAE.
CC BY-NC 4.0 rules out commercial use.
If ComfyUI doesn't recognize the model, update to the latest version.
“out of memory,” a crash, or your computer freezes
The model is too big for your VRAM (video memory). Pick a smaller variant from the table above and close other programs, especially games and browsers with lots of tabs.
It runs, but extremely slowly
Your CPU is probably doing the work instead of your graphics card. In Ollama, the PROCESSOR column of ollama ps shows how much runs on the GPU. If it says "100% CPU", update your graphics driver or pick a smaller variant that fits entirely into VRAM (video memory).
“ollama” or “python” is not found
Close the terminal window and open it again so it recognizes the newly installed program. For Python on Windows: reinstall it and check Add python.exe to PATH.
Windows: "Running scripts is disabled on this system"
Run the script with exactly the command above: powershell -ExecutionPolicy Bypass -File try_windows.ps1. This allows execution for this one time only.
Error message when loading the model (version, "trust_remote_code", unknown architecture)
Brand-new models often need the latest software version. With Ollama: update Ollama. With Python: pip install -U transformers in the activated environment. The Model Card usually lists the exact version.
Tokenstrom generates the scripts and notebook automatically. Please take a quick look before running them – they're plain, readable text. Mind the model's license: cc-by-nc-4.0.
Ask the AI assistant
Questions about this article, installation or background? Our assistant knows this page and runs on our own machine, not in someone else's cloud.
The Assistant is an AI (Gemma 4) and can make mistakes – please check important details against the source. Questions are used only to generate an answer and are deleted after one hour. Please don't enter any personal data.
Comments
What do you think? Already tried it? Write it here, no sign-up needed.
Comments are public. Links aren't allowed, and we delete abusive posts. You can delete your own comment at any time in this browser. Learn more in the Privacy Policy.