Picture for Erik McDermott

Erik McDermott

Neural Transducer Training: Reduced Memory Consumption with Sample-wise Computation

Nov 29, 2022
Stefan Braun, Erik McDermott, Roger Hsiao

Figure 1 for Neural Transducer Training: Reduced Memory Consumption with Sample-wise Computation

Figure 2 for Neural Transducer Training: Reduced Memory Consumption with Sample-wise Computation

Figure 3 for Neural Transducer Training: Reduced Memory Consumption with Sample-wise Computation

Figure 4 for Neural Transducer Training: Reduced Memory Consumption with Sample-wise Computation

* 5 pages, 4 figures, 1 table, 1 algorithm

Access Paper or Ask Questions

A Density Ratio Approach to Language Model Fusion in End-To-End Automatic Speech Recognition

Feb 28, 2020
Erik McDermott, Hasim Sak, Ehsan Variani

Figure 1 for A Density Ratio Approach to Language Model Fusion in End-To-End Automatic Speech Recognition

Figure 2 for A Density Ratio Approach to Language Model Fusion in End-To-End Automatic Speech Recognition

Figure 3 for A Density Ratio Approach to Language Model Fusion in End-To-End Automatic Speech Recognition

Figure 4 for A Density Ratio Approach to Language Model Fusion in End-To-End Automatic Speech Recognition

* 8 pages, 4 figures, presented at 2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU 2019)

Access Paper or Ask Questions

Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Loss

Feb 14, 2020
Qian Zhang, Han Lu, Hasim Sak, Anshuman Tripathi, Erik McDermott, Stephen Koo, Shankar Kumar

Figure 1 for Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Loss

Figure 2 for Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Loss

Figure 3 for Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Loss

Figure 4 for Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Loss

* This is the final version of the paper submitted to the ICASSP 2020 on Oct 21, 2019

Access Paper or Ask Questions