Auptimizer v2.0 Release

Major Features

• Add Dashboard to provide insightful visualizations to help users track experiment progress, analyze and contrast jobs, experiments, and optimization approaches
• Add Compressor - a model compression tool that helps reduce memory complexity and inference time of neural networks
• Add 4 early stopping strategies to improve the efficiency of HPO experiments
• Add feature to track intermediate results of both HPO and compression experiments
• Add feature to save the best model with the best hyperparameter combination
• Add feature to track multiple results returned from the HPO experiment

Enhancements

• Refactor documentation
• Minor bug fixes

Auptimizer v1.4 Release

Major Features

• Add Converter to convert models in various formats (SavedModel/Protobuf/Checkpoint/Keras/Pytorch) into TensorFlow Lite or ONNX
• Re-establish SSH connection upon connection failure
• Resubmit failed jobs and resume experiment from job failure

Enhancements

• Refactor documentation

Auptimizer v1.3 Release

Major Features

• Adding Profiler - a simulator for profiling performance of Machine Learning model scripts.