Skip to content

Created a Tanh gate for LSTM cell, a type of neural network using RTL with the goal of least area and execution time product

Notifications You must be signed in to change notification settings

SidharthMehta/Tanh-unit

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Tanh Unit for LSTM cell

The project implements a g(t) gate within an LSTM cell. LSTM stands for Long Short Term Memory and is a type of neural network. Screenshot Screenshot

g(t) gate is a tanh function performed on the weighted sum of inputs. Weighted sum is calculated using matrix multiplication of weight Wg with input Xt. • Wg is a matrix of size 16*16. Each weight is 16-bit fixed point. • Xt is a vector of size 16 with each element being 16-bit fixed point.

g(t) gate finds its output using interpolation on weighted sum with help of a lookup table.

The aim of this project was to design g(t) gate while achieving minimal area-delay product.

Veiw project report

About

Created a Tanh gate for LSTM cell, a type of neural network using RTL with the goal of least area and execution time product

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published