Skip to content

Latest commit

 

History

History
7 lines (6 loc) · 665 Bytes

README.md

File metadata and controls

7 lines (6 loc) · 665 Bytes

GordoNN

GordoNN is a model for speech recognition for neuromorphic cochleas based on two implementations of HOTS, the first being the zero dimentional HOTS from Germain, Kevin and Ryad, and the second the classic one dimensional HOTS.
Notice that due to memory limitation the Datasets are not loaded on the repo, therefore, in the case of the On and Off recognition task, the folders containing the data (on_aedats,off_aedats) should be positioned in the DATA/On_Off/ folder

Alt Text
(One of the biggest problems to deal with cochleas is to account for the different gain values of each filter)