Timeline for How to update first layer weights?
Current License: CC BY-SA 4.0
2 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 29 at 22:52 | comment | added | Allo Bonjour | Thanks for the answer! After reading some code examples, i've learned that: To get the gradient of the last layer, I have to mutliply the error by the derivative of the activation function of the output with respect to the output. Then, I can get the gradient of layer n by multiplying the gradient of the previous layer by the weights of layer n and then multiplying that with the derivative of the activation function with respect to the activation of layer n. Is this correct? | |
| May 28 at 6:20 | history | answered | Tomasz Witkowski | CC BY-SA 4.0 |