Tensorized Embedding Layers for Efficient Model Compression
This addresses the deployment challenge of NLP models in resource-limited settings by compressing embedding layers, though it is incremental as it builds on existing tensor decomposition techniques.
The paper tackles the problem of large embedding layers in NLP models by introducing a Tensor Train decomposition method, achieving significant compression with minimal performance loss or slight gains across various architectures.
The embedding layers transforming input words into real vectors are the key components of deep neural networks used in natural language processing. However, when the vocabulary is large, the corresponding weight matrices can be enormous, which precludes their deployment in a limited resource setting. We introduce a novel way of parametrizing embedding layers based on the Tensor Train (TT) decomposition, which allows compressing the model significantly at the cost of a negligible drop or even a slight gain in performance. We evaluate our method on a wide range of benchmarks in natural language processing and analyze the trade-off between performance and compression ratios for a wide range of architectures, from MLPs to LSTMs and Transformers.