Deep Java Library (DJL) v0.6.0 Release Notes
Release Date: 2020-06-25 // almost 4 years ago-
๐ DJL 0.6.0 brings stable Android support, ONNX Runtime experimental inference support, experimental training support for PyTorch.
Key Features
- ๐ Stable Android inference support for PyTorch models
- Provide abstraction for Image processing using ImageFactory
- ๐ Experimental support for inference on ONNX models
- ๐ Initial experimental training and imperative inference support for PyTorch engine
- ๐ Experimental support for using multi-engine
- ๐ Improved usage for NDIndex - support for ellipsis notation, arguments
- ๐ Improvements to AbstractBlock to simplify custom block creation
- โ Added new datasets
๐ Documentation and examples
- โ Added Sentiment Analysis training example
- โก๏ธ Updates the DJL webpage with new demos
๐ฅ Breaking changes
- ๐ง ModelZoo Configuration changes
- ImageFactory changes
- โ Please refer to javadocs for minor API changes
Known issues
- Issue with training with MXNet in multi-gpu instances
- ๐ Stable Android inference support for PyTorch models