TastyPancakes
/
Flux.jl
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

clearer docs

This commit is contained in:
Mike J Innes 2018-06-26 16:07:58 +01:00
parent 88c16e62dd
commit bed6d2311e
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
docs/src/models/basics.md
View File

@ -28,13 +28,15 @@ l = loss(x, y)
back!(l)
```
`loss(x, y)` returns the same number, but it's now a *tracked* value that records gradients as it goes along. Calling `back!` then calculates the gradient of `W` and `b`. We can see what this gradient is, and modify `W` to train the model.
`loss(x, y)` returns the same number, but it's now a *tracked* value that records gradients as it goes along. Calling `back!` then accumulates the gradient of `W` and `b`. We can see what this gradient is, and modify `W` to train the model.
```julia
W.grad
# Update the parameter
W.data .-= 0.1(W.grad)
# Reset the gradient
W.grad .= 0
loss(x, y) # ~ 2.5
```