Examples

• Timeseries Classification
  • Dataset info
  • Configure TrainGenerator, DataGenerators, & model
  • Visualize some samples
  • Visualize LSTM weights before training
  • Train
  • Visualize LSTM weights post-training
  • Predictions per iteration
  • Predictions distribution
  • Inference vs. Train histogram
  • Windowed timeseries idea
• Model Health Monitoring
  • Case 1: Large weights
  • Case 2: Exploding/vanishing weights
• Reproducibility
  • Random seed setter
  • Configure & train
  • Clear current session
  • Start new session, load savefile
• Preprocessor & batching logic
  • RandCropPreprocessor
  • TimeseriesPreprocessor
• Flexible batch_size & Faster SSD Loading
  • User batch_size=64, file batch_size=128
  • User batch_size=256, file batch_size=128
  • Faster SSD Loading
• Model Auto-naming