philipperemy/keract
Layers Outputs and Gradients in Keras. Made easy.
This tool helps machine learning engineers and researchers understand the inner workings of their Keras deep learning models. It takes a Keras model and input data, then outputs the activation values (what each layer 'sees') and gradients (how much each part of the model contributes to the error) for every layer. This allows practitioners to debug, interpret, and visualize the decision-making process within complex neural networks.
1,055 stars. No commits in the last 6 months.
Use this if you need to inspect the activations or gradients of individual layers within your Keras or TensorFlow models to understand their behavior or troubleshoot issues.
Not ideal if you are using TensorFlow version 2.16 or newer, as this tool is not compatible with those versions.
Stars
1,055
Forks
191
Language
Python
License
MIT
Category
Last pushed
Apr 07, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/philipperemy/keract"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
lutzroeder/netron
Visualizer for neural network, deep learning and machine learning models
mert-kurttutan/torchview
torchview: visualize pytorch models
raghakot/keras-vis
Neural network visualization toolkit for keras
janosh/diagrams
Diagrams of concepts in physics/chemistry/ML
dvgodoy/deepreplay
Deep Replay - Generate visualizations as in my "Hyper-parameters in Action!" series!