Skip to content

Comparison of Machine Learning Frameworks for In-browser Model Execution

Focus: Implementation, Machine Learning, Web Apps, Literature Review, Comparison. Requirements: good web development skills, basic ML/AI knowledge.

Research Questions

  1. Which frameworks are commonly used for creating ML models? (PyTorch, OpenVino, Caffe)
  2. Which datasets have become established? (COCO, ImageNet, MNIST)
  3. Which pretrained models are frequently used? (MobileNet, SAM, BERT, YOLO)

Comparison Criteria

  • Compatibility — Support for other models and frameworks
  • Functionality — GPU acceleration, parallel execution
  • Performance — Speed, memory usage, scalability across browsers
  • Adoption — Community size, maintenance
  • Integration — Ease of use, API design (optional)

Implementation

A browser application using TensorFlow.js and ONNX Runtime Web to validate results. Optionally integrated into a React application.

Sources