please tell me the importance of using bais variable and also it’s gradient
About the bais(b) variable
hey @light007 ,
Lets currently consider that we are not using bias for now.
We know that are weights are randomly initialized , there is case when any input can 0 but the required output needs not to be 0. As the input is 0 , its weight will not perform anything and hence here comes into the rescue. Adding this bias value can be effective in case for learning data.
Bias term makes model more flexible and helps to achieve the correct line more faster and accurately.
for better understanding you can have a look at this link
and for its gradient :
The we need to tune and update our weights similarly we also need to update our bias values to get the best bias terms for our model.
I hope this would have resolved your doubt.
Thank you.
Happy Coding. 