OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Falco

Falco is an open-source Python and PyTorch research implementation of latent-code optimization for face-dataset anonymization. It uses a pretrained GAN latent space to obfuscate identity while matching selected facial features, aiming to keep images useful for downstream machine-learning tasks.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Falco — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Consider Falco when a research workflow needs face-dataset anonymization that seeks to retain selected image attributes for downstream tasks, and you can manage Python dependencies, datasets and pretrained models.

Who it is for

  • Researchers studying face-image anonymization
  • Dataset developers preparing face data for downstream machine-learning work

Documented uses

  • Extract image features in CLIP, FaRL, DINO or ArcFace feature spaces for the documented CelebA-HQ dataset.
  • Prepare a pool of synthetic face images, latent codes and feature representations using a pretrained GAN.
  • Explore an anonymization approach designed to obscure identity while preserving facial attributes.
Before you choose

Requirements & limits

The repository is public and lists an Apache-2.0 license.

  • The method works through latent-code optimization in a pretrained GAN rather than training an additional purpose-built neural network, as described by the project.
  • Setup requires multiple pretrained model families; the listed prerequisite downloads total about 3.5 GB.
  • The documented real-image feature-extraction command offers CelebA-HQ as its dataset choice.
  • Feature extraction supports CLIP, FaRL, DINO and ArcFace selections, batch size options and CUDA selection.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Falco and prepare its pretrained models

Set up the documented Python environment and download the model prerequisites.

Before you start

  • Python with the native venv module
  • The Falco repository and its requirements.txt file
  • About 3.5 GB for the listed pretrained models

Steps

  1. Create and activate a Python virtual environment.

  2. Upgrade pip, install setuptools, and install the packages listed in requirements.txt.

  3. Run python download_models.py to retrieve the prerequisite pretrained models.

Expected outcome

A virtual environment with the repository requirements installed and pretrained model files placed under models/pretrained.

Guide limitations

  • CLIP and DINO models are downloaded automatically only when requested.
  • This setup does not include running the anonymization workflow.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
chi0tzp/FALCO · Official homepage title · retained in its source language
How we review sourcesSuggest a correction