qubvel/ttach
Image Test Time Augmentation with PyTorch!
This tool helps machine learning engineers and researchers improve the accuracy of their image-based AI models during evaluation. It takes a trained model and a batch of test images, then applies various transformations like flips, rotations, and scaling to each image. The model then makes predictions on all augmented versions, which are then combined to produce a more robust and accurate final prediction for segmentation, classification, or keypoint detection tasks.
1,029 stars. Used by 1 other package. No commits in the last 6 months. Available on PyPI.
Use this if you want to boost the performance and reliability of your image-based AI models on test data by leveraging augmented predictions.
Not ideal if your primary goal is to speed up model inference time, as this technique inherently increases the computation required for predictions.
Stars
1,029
Forks
69
Language
Python
License
MIT
Category
Last pushed
Jul 28, 2023
Commits (30d)
0
Reverse dependents
1
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/qubvel/ttach"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Compare
Related frameworks
TorchIO-project/torchio
Medical imaging processing for AI applications.
aleju/imgaug
Image augmentation for machine learning experiments.
makcedward/nlpaug
Data augmentation for NLP
mdbloice/Augmentor
Image augmentation library in Python for machine learning.
BloodAxe/pytorch-toolbelt
PyTorch extensions for fast R&D prototyping and Kaggle farming