205 citations · 293 across the 6 of their papers we have counts for
6 papers
Accelerating Inference and Language Model Fusion of Recurrent Neural Network Transducers via End-to-End 4-bit Quantization
Andrea Fasoli, Chia-Yu Chen, Mauricio Serrano +5
We report on aggressive quantization strategies that greatly accelerate inference of Recurrent Neural Network Transducers (RNN-T). We use a 4 bit integer representation for both we…
4-bit Quantization of LSTM-based Speech Recognition Models
Andrea Fasoli, Chia-Yu Chen, Mauricio Serrano +9
We investigate the impact of aggressive low-precision representations of weights and activations in two families of large LSTM-based architectures for Automatic Speech Recognition…
ScaleCom: Scalable Sparsified Gradient Compression for Communication-Efficient Distributed Training
Chia-Yu Chen, Jiamin Ni, Songtao Lu +8
Large-scale distributed training of Deep Neural Networks (DNNs) on state-of-the-art platforms is expected to be severely communication constrained. To overcome this limitation, num…
Accumulation Bit-Width Scaling For Ultra-Low Precision Training Of Deep Networks
Charbel Sakr, Naigang Wang, Chia-Yu Chen +4
Efforts to reduce the numerical precision of computations in deep learning training have yielded systems that aggressively quantize weights and activations, yet employ wide high-pr…
Training Deep Neural Networks with 8-bit Floating Point Numbers
Naigang Wang, Jungwook Choi, Daniel Brand +2
The state-of-the-art hardware platforms for training Deep Neural Networks (DNNs) are moving from traditional single precision (32-bit) computations towards 16 bits of precision --…
AdaComp : Adaptive Residual Gradient Compression for Data-Parallel Distributed Training
Chia-Yu Chen, Jungwook Choi, Daniel Brand +3
Highly distributed training of Deep Neural Networks (DNNs) on future compute platforms (offering 100 of TeraOps/s of computational capacity) is expected to be severely communicatio…