miguelvr/dropblock
Implementation of DropBlock: A regularization method for convolutional networks in PyTorch.
This tool helps deep learning engineers improve the accuracy and robustness of their computer vision models. By selectively disabling contiguous regions of feature maps during training, it takes a raw neural network's feature map data and outputs a regularized feature map. This process helps prevent models from overfitting and allows them to generalize better to new, unseen images.
594 stars. No commits in the last 6 months. Available on PyPI.
Use this if you are a machine learning engineer or researcher building convolutional neural networks and want to improve their performance on tasks like image classification or object detection.
Not ideal if you are working with non-convolutional neural networks or if your primary goal is to speed up model inference rather than improve training stability and accuracy.
Stars
594
Forks
94
Language
Python
License
MIT
Category
Last pushed
Jul 29, 2020
Commits (30d)
0
Dependencies
2
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/miguelvr/dropblock"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
kk7nc/RMDL
RMDL: Random Multimodel Deep Learning for Classification
MaximeVandegar/Papers-in-100-Lines-of-Code
Implementation of papers in 100 lines of code.
OML-Team/open-metric-learning
Metric learning and retrieval pipelines, models and zoo.
DLTK/DLTK
Deep Learning Toolkit for Medical Image Analysis
dasguptar/treelstm.pytorch
Tree LSTM implementation in PyTorch