FPGA Implementation of LSTM Based On Automatic Speech Recognition
FPGA Implementation of LSTM Based On Automatic Speech Recognition
FPGA Implementation of LSTM Based On Automatic Speech Recognition
Recognition
Chen-Lu Li, Yu-Jie Huang, Yu-Jie Cai, Jun Han, Xiao-Yang Zeng
1
Department of Microelectronics, Fudan University, Shanghai 200433, China
2
State Key Laboratory of ASIC and System, Fudan University, Shanghai 200433, China
* Email: [email protected]
References
[1] Bengio Y, Simard P, Frasconi P. Learning long-term
dependencies with gradient descent is difficult[J]. IEEE
Transactions on Neural Networks, p.157(2002).
[2] Guan Y, Yuan Z, Sun G, et al. FPGA-based
accelerator for long short-term memory recurrent neural
networks[C].Design Automation Conference. IEEE,
p.629-634(2017).
[3] Chang Andre Xian Ming, Martini B, Culurciello E.
Figure 6. WER of ASR system in different numbers of Recurrent Neural Networks Hardware Implementation
UV Decomposition eigenvalue and different quantization on FPGA[J]. Computer Science, (2015).
ways [4] Han S, Kang J, Mao H, et al. ESE: Efficient Speech
Recognition Engine with Sparse LSTM on FPGA[J].
4. Result and Comparison (2017).
Our design is implemented in Zynq-7000 SOC