• Jasper speech recognition model and documentation
• Audio classification model for Speech Commands dataset and documentation
• Improved documentation (LARC, distributed training, WaveNet)
• Minor enhancements (speech recognition data layer, custom decoder, TRT)
• Various bug fixes

• Significantly improved Speech recognition (WER 4.32%)
• BPE support for Speech recognition
• WaveNet
• Enhanced cuDNN RNN support for Language Modeling
• Improvements in distributed text2text (thanks to @vsuthichai )
• Ability to freeze some layers (thanks to @ka-bu )
• Various bug-fixes

What's new

• Improved and updated models with checkpoints (see the documentation)
• New models:
  • Transformer Big (for NMT)
  • Sentiment Analysis (based on universal LM)
  • Joint CTC Attention based ASR
• Improved scalability
• Speech Synthesis audio samples (in the documentation)
• Support for CUDA10, Horovod 0.14

What's new

• Improved and updated models with checkpoints (see the documentation)
• Dropped Python2 support
• Switched to TensorFlow 1.10
• Added TensorRT support for fast inference
• Refactored and updated documentation: https://nvidia.github.io/OpenSeq2Seq
• Switched versioning to month-based labels

New Modality text2speech - spectrogram synthesis from text

New Models

• Tacotron 2 - like model for text2speech (English)

Various improvements

• in ConvS2S model for translation
• in Wav2Letter - like model for speech2text
• in DeepSpeech2 - like model for speech2text
• Bugfixes

Other

• Tensorflow's version increased to 1.9

New models:

• ConvS2S model for translation.
• Wav2Letter model for speech recognition.
• CIFAR-10 dataset support.
• CNNEncoder that can be used to construct arbitrary (almost) CNN models. Based on that, integrated AlexNet and cifar10-nv.

New features:

• Support for "iter_size" (accumulating gradients for "iter_size" steps without update).
• "objects" benchmarking to evaluation and inference modes.
• cuDNN compatible cells support for GNMT.
• 8-padding for transformer.
• Improved config overwriting by train/eval/infer params (will not replace whole dicts, but update incrementally).
• Audio normalization before preprocessing for speech2text models.
• More summaries/parameters for different models.

Bug fixes:

• Regularization in mixed precision mode (loss scaling was not applied, disabling regularizer).
• Overwriting bool values from command line.
• Multi-GPU evaluation in towers mode.
• Multi-GPU inference for speech2text.
• "reflect" padding changed to use zeros for audio preprocessing.
• Unicode support for Python 2.

Important config/API changes:

• Unified static/dynamic loss scaling into a single parameter.
• Made RNN cells accept arbitrary parameters.
• Exposed training step into maybe_print_logs and evaluate functions.

Other changes:

• Improved unit tests and documentation.

New models:

• Added ResNet model and ImageNet data layer.
• Improved DeepSpeech-2 models and reached 4.59% WER.
• Added Transformers model.

New features:

• Implemented evaluation in Horovod mode.
• Added mixed precision support for Horovod mode.
• Fixed evaluation in multi-GPU mode.
• All string/numerical config parameters can now be rewritten from command line (nested dicts are separated with "/").
• Moved start_experiment.sh functionality to run.py (--enable_logs parameter). Additionally it now logs exact command line arguments used to invoke the script.
• Added new benchmarking functionality: now models can also report number of objects (e..g tokens or images) per second.
• Added more summaries/parameters for different models.

API changes:

• Replaced Seq2Seq class with EncoderDecoderModel to support arbitrary models that can be expressed in encoder-decoder-loss paradigm.
• Changed data layer API to only work with tf.data (dropped placeholders support).
• Hid Horovod/non Horovod differences from users (no need to take care about that when creating new models / data layers).

Other changes:

• Improved unit tests and documentation.

• Massive API changes
• Add mixed precision training support
• Add speech-to-text models support
• Improved documentation