Efficient Training of Neural Transducer for Speech Recognition

• URL: http://arxiv.org/abs/2204.10586v1
• Date: Fri, 22 Apr 2022 09:16:51 GMT
• Title: Efficient Training of Neural Transducer for Speech Recognition
• Authors: Wei Zhou, Wilfried Michel, Ralf Schl\"uter, Hermann Ney
• Abstract summary: We propose an efficient 3-stage progressive training pipeline to build highly-performing neural transducer models from scratch. The proposed pipeline is able to train transducer models approaching state-of-the-art performance with a single GPU in just 2-3 weeks.
• Score: 44.99337868233026
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: As one of the most popular sequence-to-sequence modeling approaches for speech recognition, the RNN-Transducer has achieved evolving performance with more and more sophisticated neural network models of growing size and increasing training epochs. While strong computation resources seem to be the prerequisite of training superior models, we try to overcome it by carefully designing a more efficient training pipeline. In this work, we propose an efficient 3-stage progressive training pipeline to build highly-performing neural transducer models from scratch with very limited computation resources in a reasonable short time period. The effectiveness of each stage is experimentally verified on both Librispeech and Switchboard corpora. The proposed pipeline is able to train transducer models approaching state-of-the-art performance with a single GPU in just 2-3 weeks. Our best conformer transducer achieves 4.1% WER on Librispeech test-other with only 35 epochs of training.

Related papers

• Self-Supervised Radio Pre-training: Toward Foundational Models for Spectrogram Learning [6.1339395157466425]
  Foundational deep learning (DL) models are general models, trained on diverse, diverse, and unlabelled datasets. We introduce Masked Spectrogram Modeling, a novel self-supervised learning approach for pretraining foundational DL models on radio signals.
  arXiv  Detail & Related papers  (2024-11-14T23:56:57Z)
• Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think [72.48325960659822]
  One main bottleneck in training large-scale diffusion models for generation lies in effectively learning these representations. We study this by introducing a straightforward regularization called REPresentation Alignment (REPA), which aligns the projections of noisy input hidden states in denoising networks with clean image representations obtained from external, pretrained visual encoders. The results are striking: our simple strategy yields significant improvements in both training efficiency and generation quality when applied to popular diffusion and flow-based transformers, such as DiTs and SiTs.
  arXiv  Detail & Related papers  (2024-10-09T14:34:53Z)
• Always-Sparse Training by Growing Connections with Guided Stochastic Exploration [46.4179239171213]
  We propose an efficient always-sparse training algorithm with excellent scaling to larger and sparser models. We evaluate our method on CIFAR-10/100 and ImageNet using VGG, and ViT models, and compare it against a range of sparsification methods.
  arXiv  Detail & Related papers  (2024-01-12T21:32:04Z)
• No Train No Gain: Revisiting Efficient Training Algorithms For Transformer-based Language Models [31.080446886440757]
  In this work, we revisit three categories of such algorithms: dynamic architectures (layer stacking, dropping), batch selection (selective backprop, RHO loss), and efficient layers (Lion, Sophia) We find that their training, validation, and downstream gains vanish compared to a baseline with a fully-decayed learning rate. We define an evaluation protocol that enables machines to be done on arbitrary computation by mapping all computation time to a reference machine which we call reference system time.
  arXiv  Detail & Related papers  (2023-07-12T20:10:14Z)
• Continual learning autoencoder training for a particle-in-cell simulation via streaming [52.77024349608834]
  upcoming exascale era will provide a new generation of physics simulations with high resolution. These simulations will have a high resolution, which will impact the training of machine learning models since storing a high amount of simulation data on disk is nearly impossible. This work presents an approach that trains a neural network concurrently to a running simulation without data on a disk.
  arXiv  Detail & Related papers  (2022-11-09T09:55:14Z)
• Online Convolutional Re-parameterization [51.97831675242173]
  We present online convolutional re- parameterization (OREPA), a two-stage pipeline, aiming to reduce the huge training overhead by squeezing the complex training-time block into a single convolution. Compared with the state-of-the-art re-param models, OREPA is able to save the training-time memory cost by about 70% and accelerate the training speed by around 2x. We also conduct experiments on object detection and semantic segmentation and show consistent improvements on the downstream tasks.
  arXiv  Detail & Related papers  (2022-04-02T09:50:19Z)
• EarlyBERT: Efficient BERT Training via Early-bird Lottery Tickets [106.79387235014379]
  EarlyBERT is a general computationally-efficient training algorithm applicable to both pre-training and fine-tuning of large-scale language models. We are the first to identify structured winning tickets in the early stage of BERT training, and use them for efficient training. EarlyBERT easily achieves comparable performance to standard BERT with 3545% less training time.
  arXiv  Detail & Related papers  (2020-12-31T20:38:20Z)
• FracTrain: Fractionally Squeezing Bit Savings Both Temporally and Spatially for Efficient DNN Training [81.85361544720885]
  We propose FracTrain that integrates progressive fractional quantization which gradually increases the precision of activations, weights, and gradients. FracTrain reduces computational cost and hardware-quantified energy/latency of DNN training while achieving a comparable or better (-0.12%+1.87%) accuracy.
  arXiv  Detail & Related papers  (2020-12-24T05:24:10Z)
• Accelerating Training of Transformer-Based Language Models with Progressive Layer Dropping [24.547833264405355]
  The proposed method achieves a 24% time reduction on average per sample and allows the pre-training to be 2.5 times faster than the baseline. While being faster, our pre-trained models are equipped with strong knowledge transferability, achieving comparable and sometimes higher GLUE score than the baseline.
  arXiv  Detail & Related papers  (2020-10-26T06:50:07Z)
• A Practical Incremental Method to Train Deep CTR Models [37.54660958085938]
  We introduce a practical incremental method to train deep CTR models, which consists of three decoupled modules. Our method can achieve comparable performance to the conventional batch mode training with much better training efficiency.
  arXiv  Detail & Related papers  (2020-09-04T12:35:42Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.