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

Extend docs about train!

Browse Source 
Related to #921: explain why it is not needed to pass the model as argument.
This commit is contained in:
Helios De Rosario 2019-11-08 21:39:00 +01:00 committed by GitHub
parent 9d05afaccc
commit 7e1ffd6507
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
docs/src/training/training.md
View File

@ -1,6 +1,6 @@
# Training
To actually train a model we need three things:
To actually train a model we need three things, in addition to the tracked parameters that will be fitted:
* A *objective function*, that evaluates how well a model is doing given some input data.
* A collection of data points that will be provided to the objective function.
@ -11,6 +11,7 @@ With these we can call `Flux.train!`:
```julia
Flux.train!(objective, params, data, opt)
```
At first glance it may seem strange that the model that we want to train is not part of the input arguments of `Flux.train!`. However the target of the optimizer is not the model itself, but the objective function that represents the departure between modelled and observed data. In other words, the model is implicitly defined in the objective function, and there is no need to give it explicitly. Passing the objective function instead of the model and a cost function separately (see below) provides more flexibility, and the possibility of optimizing the calculations.
There are plenty of examples in the [model zoo](https://github.com/FluxML/model-zoo).