Fixed: - removed extra loss computation during validation/testing
0.2.4
New: - N/A
Changed: - fixed image link in README
Fixed: - N/A
0.2.3
New: - N/A
Changed: - changed `loss` to `criterion` in `Learner`
Fixed: - N/A
0.2.2
New: - added support for gradient clipping
Changed: - re-factored steps in training to support subclassing `Learner` for custom overrides - added warning about learning rate setting when user-configured scheduler is supplied to `fit` - `fit` accepts custom learning rate schedulers
Fixed: - N/A
0.2.1
New: - Support lists of inputs (e.g., [text, offsets] for `EmbeddingBag`) - Added `predict_example` method - Added text classification example
Changed: - N/A
Fixed: - N/A
0.2.0
New: - Added `gpus` parameter to `Learner` to enable multi-gpu training. - Support lists of inputs (e.g., [text, offsets] for `EmbeddingBag`)