NiklasvonM/Self-Training
Iterative training on pseudo-labeled data experiment on the MNIST-dataset
This project helps machine learning researchers understand how to effectively train image classification models when only a small amount of labeled data is available. It takes a small set of labeled images and a large set of unlabeled images, then iteratively uses the model's own predictions to expand the training data. The output demonstrates how accuracy improves over iterations and how confidence thresholds affect the training process, providing insights for researchers studying semi-supervised learning.
No commits in the last 6 months.
Use this if you are a machine learning researcher exploring semi-supervised learning techniques and want to experiment with iterative pseudo-labeling for image classification.
Not ideal if you are looking for a ready-to-use solution for production image classification or a tool for datasets other than images.
Stars
12
Forks
2
Language
Python
License
MIT
Category
Last pushed
Sep 03, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/NiklasvonM/Self-Training"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
greydanus/mnist1d
A 1D analogue of the MNIST dataset for measuring spatial biases and answering Science of Deep...
potterhsu/SVHNClassifier
A TensorFlow implementation of Multi-digit Number Recognition from Street View Imagery using...
potterhsu/SVHNClassifier-PyTorch
A PyTorch implementation of Multi-digit Number Recognition from Street View Imagery using Deep...
kensanata/numbers
Handwritten digits, a bit like the MNIST dataset.
kdexd/digit-classifier
A single handwritten digit classifier, using the MNIST dataset. Pure Numpy.