Neural Machine Translation with 4-Bit Precision and Beyond
arXiv:1909.06091
Abstract
Neural Machine Translation (NMT) is resource intensive. We design a quantization procedure to compress NMT models better for devices with limited hardware capability. Because most neural network parameters are near zero, we employ logarithmic quantization in lieu of fixed-point quantization. However, we find bias terms are less amenable to log quantization but note they comprise a tiny fraction of the model, so we leave them uncompressed. We also propose to use an error-feedback mechanism during retraining, to preserve the compressed model as a stale gradient. We empirically show that NMT models based on Transformer or RNN architecture can be compressed up to 4-bit precision without any noticeable quality degradation. Models can be compressed up to binary precision, albeit with lower quality. The RNN architecture seems to be more robust to quantization, compared to the Transformer.
References in corpus (10)
- Adam: A Method for Stochastic Optimization
- Neural Machine Translation by Jointly Learning to Align and Translate
- Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding
- Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation
- Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
- Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Training
- Binarized Neural Networks
- Fixed Point Quantization of Deep Convolutional Networks
- Convolutional Neural Networks using Logarithmic Data Representation
- Making Asynchronous Stochastic Gradient Descent Work for Transformers