All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog and this project adheres to Semantic Versioning.
- Fix ignored SIGINT
- Add
ModuleTrainer.evaluate
method - Add CsvWriter to evaluate method
- Add
filename_transform_function
argument toSaveBestCheckpointCallback
- Metric step method now return the intermediate computed values
- Rename
CsvWriter
'sextra
argument toextra_data_function
CsvWriter
can now be called with theextra_data
extra argument to define the extra data that will be logged- Add
ModuleTrainer.load
method
- Rename
dateset_loader
todataloader
ModuleTrainer
objectEarlyStopping
: stop training after a configurable period of stagnation- Checkpointing: save model and estimator at regular intervals
- CSV file writer to output logs
- Several metrics are available: all default PyTorch loss functions, Accuracy, MAE
- Progress bar from console
- SIGINT handling: handle CTRL-C
- Model's data type (float32, float64)
- Full use of Pytorch's Cuda support