LGJun 6, 2024

Speed of Light Exact Greedy Decoding for RNN-T Speech Recognition Models on GPU

arXiv:2406.03791v114 citationsHas Code
Originality Incremental advance
AI Analysis

This work enables high-throughput inference for RNN-T speech recognition models, making them competitive with CTC models, which is significant for real-time applications like voice assistants.

The paper tackles the problem of inefficient GPU utilization during RNN-T model decoding, which causes ~80% idle time, by introducing an exact GPU-based greedy decoding implementation using CUDA graph conditional nodes, achieving a 2.5x end-to-end speedup for a 1.1 billion parameter model.

The vast majority of inference time for RNN Transducer (RNN-T) models today is spent on decoding. Current state-of-the-art RNN-T decoding implementations leave the GPU idle ~80% of the time. Leveraging a new CUDA 12.4 feature, CUDA graph conditional nodes, we present an exact GPU-based implementation of greedy decoding for RNN-T models that eliminates this idle time. Our optimizations speed up a 1.1 billion parameter RNN-T model end-to-end by a factor of 2.5x. This technique can applied to the "label looping" alternative greedy decoding algorithm as well, achieving 1.7x and 1.4x end-to-end speedups when applied to 1.1 billion parameter RNN-T and Token and Duration Transducer models respectively. This work enables a 1.1 billion parameter RNN-T model to run only 16% slower than a similarly sized CTC model, contradicting the common belief that RNN-T models are not suitable for high throughput inference. The implementation is available in NVIDIA NeMo.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes