MNIST Deep Learning Project
Overview 🚀
This repository explores the application of three distinct techniques using the MNIST dataset. Our primary objective is to assess the efficacy of supervised, semi-supervised, and self-supervised learning methods.
Techniques 🎯
-
Supervised Learning
- Objective: Train a model to recognize numbers using 100 MNIST examples.
- Approach: Employ various methodologies to optimize model performance and utilize data augmentation to enhance the model's accuracy.
-
Semi-Supervised Learning
- Objective: Enable the model to autonomously comprehend data without labeled guidance.
- Task: Classify unlabeled MNIST data into distinct number groups and reconstruct the 10 classes of the target (label).
-
Self-Supervised Learning
- Objective: Evaluate the independent performance of trained models.
- Significance: Assess the model's ability to make accurate predictions without external guidance.
📋 Project Deliverables
- Supervised Learning: Present the model's proficiency in recognizing numbers.
- Semi-Supervised Learning: Demonstrate the model's autonomous categorization of unlabeled data.
- Self-Supervised Learning: Evaluate the model's capability to operate independently.