The GNMT v2 model is similar to the one discussed in the Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation paper.

The most important difference between the two models is in the attention mechanism. In our model, the output from the first LSTM layer of the decoder goes into the attention module, then the re-weighted context is concatenated with inputs to all subsequent LSTM layers in the decoder at the current timestep.

The same attention mechanism is also implemented in the default GNMT-like models from TensorFlow Neural Machine Translation Tutorial and NVIDIA OpenSeq2Seq Toolkit.

This model is trained with mixed precision using Tensor Cores on Volta, Turing, and the NVIDIA Ampere GPU architectures. Therefore, researchers can get results 2x faster than training without Tensor Cores, while experiencing the benefits of mixed precision training. This model is tested against each NGC monthly container release to ensure consistent accuracy and performance over time.

Model architecture

The following image shows the GNMT model architecture:

Default configuration

The following features were implemented in this model:

• general:

• encoder and decoder are using shared embeddings

• data-parallel multi-GPU training

• dynamic loss scaling with backoff for Tensor Cores (mixed precision) training

• trained with label smoothing loss (smoothing factor 0.1)

• encoder:

• 4-layer LSTM, hidden size 1024, first layer is bidirectional, the rest are unidirectional

• with residual connections starting from 3rd layer

• dropout is applied on input to all LSTM layers, probability of dropout is set to 0.2

• hidden state of LSTM layers is initialized with zeros

• weights and bias of LSTM layers is initialized with uniform (-0.1, 0.1) Distribution

• decoder:

• 4-layer unidirectional LSTM with hidden size 1024 and fully-connected classifier

• with residual connections starting from 3rd layer

• dropout is applied on input to all LSTM layers, probability of dropout is set to 0.2

• hidden state of LSTM layers is initialized with the last hidden state from encoder

• weights and bias of LSTM layers is initialized with uniform (-0.1, 0.1) distribution

• weights and bias of fully-connected classifier is initialized with uniform (-0.1, 0.1) distribution

• attention:

• normalized Bahdanau attention

• output from first LSTM layer of decoder goes into attention, then re-weighted context is concatenated with the input to all subsequent LSTM layers of the decoder at the current timestep

• linear transform of keys and queries is initialized with uniform (-0.1, 0.1), normalization scalar is initialized with 1.0 / sqrt(1024), normalization bias is initialized with zero

• inference:

• beam search with default beam size of 5

• with coverage penalty and length normalization, coverage penalty factor is set to 0.1, length normalization factor is set to 0.6 and length normalization constant is set to 5.0

• de-tokenized BLEU computed by SacreBLEU

• motivation for choosing SacreBLEU

When comparing the BLEU score, there are various tokenization approaches and BLEU calculation methodologies; therefore, ensure you align similar metrics.

Code from this repository can be used to train a larger, 8-layer GNMT v2 model. Our experiments show that a 4-layer model is significantly faster to train and yields comparable accuracy on the public WMT16 English-German dataset. The number of LSTM layers is controlled by the --num_layers parameter in the nmt.py script.

Feature support matrix

The following features are supported by this model.

Features

The following features are supported by this model.

• Automatic Mixed Precision (AMP) - Computation graphs can be modified by TensorFlow on runtime to support mixed precision training. Detailed explanation of mixed precision can be found in the next section.

Mixed precision training

Mixed precision is the combined use of different numerical precisions in a computational method. Mixed precision training offers significant computational speedup by performing operations in half-precision format, while storing minimal information in single-precision to retain as much information as possible in critical parts of the network. Since the introduction of Tensor Cores in Volta, and following with both the Turing and Ampere architectures, significant training speedups are experienced by switching to mixed precision -- up to 3x overall speedup on the most arithmetically intense model architectures. Using mixed precision training previously required two steps:

1. Porting the model to use the FP16 data type where appropriate.
2. Adding loss scaling to preserve small gradient values.

This can now be achieved using Automatic Mixed Precision (AMP) for TensorFlow to enable the full mixed precision methodology in your existing TensorFlow model code. AMP enables mixed precision training on Volta, Turing, and NVIDIA Ampere GPU architectures automatically. The TensorFlow framework code makes all necessary model changes internally.

In TF-AMP, the computational graph is optimized to use as few casts as necessary and maximize the use of FP16, and the loss scaling is automatically applied inside of supported optimizers. AMP can be configured to work with the existing tf.contrib loss scaling manager by disabling the AMP scaling with a single environment variable to perform only the automatic mixed-precision optimization. It accomplishes this by automatically rewriting all computation graphs with the necessary operations to enable mixed precision training and automatic loss scaling.

For information about:

• How to train using mixed precision, see the Mixed Precision Training paper and Training With Mixed Precision documentation.
• Techniques used for mixed precision training, see the Mixed-Precision Training of Deep Neural Networks blog.
• How to access and enable AMP for TensorFlow, see Using TF-AMP from the TensorFlow User Guide.
• APEX tools for mixed precision training, see the NVIDIA Apex: Tools for Easy Mixed-Precision Training in PyTorch.

Enabling mixed precision

Mixed precision is enabled in TensorFlow by using the Automatic Mixed Precision (TF-AMP) extension which casts variables to half-precision upon retrieval, while storing variables in single-precision format. Furthermore, to preserve small gradient magnitudes in backpropagation, a loss scaling step must be included when applying gradients. In TensorFlow, loss scaling can be applied statically by using simple multiplication of loss by a constant value or automatically, by TF-AMP. Automatic mixed precision makes all the adjustments internally in TensorFlow, providing two benefits over manual operations. First, programmers need not modify network model code, reducing development and maintenance effort. Second, using AMP maintains forward and backward compatibility with all the APIs for defining and running TensorFlow models.

To enable mixed precision, you can simply add the values to the environmental variables inside your training script:

• Enable TF-AMP graph rewrite:

os.environ["TF_ENABLE_AUTO_MIXED_PRECISION_GRAPH_REWRITE"] = "1"

• Enable Automated Mixed Precision:

os.environ['TF_ENABLE_AUTO_MIXED_PRECISION'] = '1'

Enabling TF32

TensorFloat-32 (TF32) is the new math mode in NVIDIA A100 GPUs for handling the matrix math also called tensor operations. TF32 running on Tensor Cores in A100 GPUs can provide up to 10x speedups compared to single-precision floating-point math (FP32) on Volta GPUs.

TF32 Tensor Cores can speed up networks using FP32, typically with no loss of accuracy. It is more robust than FP16 for models which require high dynamic range for weights or activations.

For more information, refer to the TensorFloat-32 in the A100 GPU Accelerates AI Training, HPC up to 20x blog post.

TF32 is supported in the NVIDIA Ampere GPU architecture and is enabled by default.