EmoPy
github.com
- Free plan
- Windows
- macOS
- Linux
- Self-hosted
7.2 Club score Visit github.com
Plans
Every plan and priceEmoPy
Free
Free plan
What the maker publishes
- Accuracy caveat
- The documented models overfit, and are less accurate on new images than on training images. source
- Architecture
- It includes five neural-network architectures implemented with Keras and TensorFlow. source
- Dependencies
- The documented setup requires Python 3.6 or newer; GraphViz is required for visualization functions. source
- Emotion labels
- The project describes seven standard labels: happiness, anger, fear, surprise, disgust, sadness, and calm/neutral. source
- Input conditions
- The project says it works best with evenly lit images and images resembling the training data's style, framing, and cropping. source
- Installation
- The README documents installation through pip or from the source repository. source
- Integrations
- The repository lists RIOT AI and a ROS wrapper as projects built on EmoPy. source
- License
- The repository lists an AGPL-3.0 license. source
- Pretrained models
- The FERModel module uses pretrained models to simplify getting a prediction model running. source
- Purpose
- EmoPy is a Python toolkit that predicts human emotional-expression classifications from face images. source
- Support and contribution
- The maintainers invite suggestions and contributions and direct users to contact an EmoPy maintainer. source
- Training data limit
- EmoPy is built around public datasets, which the project identifies as a constraint on training accuracy. source
Checked against the maker's pages on 14 Nov 2026.