v0.7.0

Notable (possibly breaking) changes:

• d45db80: Remove checkpoint utility functions from utils.py into checkpoint_utils.py
• f2563c2: Move LM definitions into separate files
• dffb167: Updates to model API:
  • FairseqModel -> FairseqEncoderDecoderModel
  • add FairseqDecoder.extract_features and FairseqDecoder.output_layer
  • encoder_out_dict -> encoder_out
  • rm unused remove_head functions
• 34726d5: Move distributed_init into DistributedFairseqModel
• cf17068: Simplify distributed launch by automatically launching multiprocessing on each node for all visible GPUs (allows launching just one job per node instead of one per GPU)
• d45db80: Change default LR scheduler from reduce_lr_on_plateau to fixed
• 96ac28d: Rename --sampling-temperature -> --temperature
• fc1a19a: Deprecate dummy batches
• a1c997b: Add memory mapped datasets
• 0add50c: Allow cycling over multiple datasets, where each one becomes an "epoch"

Plus many additional features and bugfixes

v0.6.2

Changelog:

• 998ba4f: Add language models from Baevski & Auli (2018)
• 4294c4f: Add mixture of experts code from Shen et al. (2019)
• 0049349: Add example for multilingual training
• 48d9afb: Speed improvements, including fused operators from apex
• 44d27e6: Add Tensorboard support
• d17fa85: Add Adadelta optimizer
• 9e1c880: Add FairseqEncoderModel
• b65c579: Add FairseqTask.inference_step to modularize generate.py
• 2ad1178: Add back --curriculum
• Misc bug fixes and other features

v0.6.1

Bumping version number for PyPI release.

v0.6.0

Changelog:

• 4908863: Switch to DistributedDataParallelC10d and bump version 0.5.0 -> 0.6.0
  • no more FP16Trainer, we just have an FP16Optimizer wrapper
  • most of the distributed code is moved to a new wrapper class called DistributedFairseqModel, which behaves like DistributedDataParallel and a FairseqModel at the same time
  • Trainer now requires an extra dummy_batch argument at initialization, which we do fwd/bwd on when there's an uneven number of batches per worker. We hide the gradients from these dummy batches by multiplying the loss by 0
  • Trainer.train_step now takes a list of samples, which will allow cleaner --update-freq
• 1c56b58: parallelize preprocessing
• Misc bug fixes and features

v0.5.0: 0.4.0 -> 0.5.0

Changelog:
- 97b58b4: add Transformer model from Vaswani et al. (2017)
- b2374e5: faster Transformer inference with improved caching
- 2d27ae0: simulate large mini-batch training with delayed updates (`--update-freq`)
- 7ee1d28: add FP16 training support (`--fp16`)
- 2a84f46: faster inference by removing completed sentences from the batch
- 663fd80: batched interactive generation
- 4c2ef2d: add language modeling / gated convolutional model from Dauphin et al. (2017)
- b59815b: add Hierarchical Neural Story Generation model from Fan et al. (2018)
- ff68a9e: add FairseqTask to modularize task definitions (e.g., translation, language modeling)

v0.4.0

Merge internal changes (#163)