Merge pull request #215 from FluxML/dev

For a 0.2.9 release

Author: ablaom

Project.toml

@@ -1,7 +1,7 @@
 name = "MLJFlux"
 uuid = "094fc8d1-fd35-5302-93ea-dabda2abf845"
 authors = ["Anthony D. Blaom <anthony.blaom@gmail.com>", "Ayush Shridhar <ayush.shridhar1999@gmail.com>"]
-version = "0.2.8"
+version = "0.2.9"
 
 [deps]
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"

src/core.jl

@@ -36,7 +36,7 @@ function train!(loss, penalty, chain, optimiser, X, y)
         parameters = Flux.params(chain)
         gs = Flux.gradient(parameters) do
             yhat = chain(X[i])
-            batch_loss = loss(yhat, y[i]) + penalty(parameters)
+            batch_loss = loss(yhat, y[i]) + penalty(parameters)/n_batches
             training_loss += batch_loss
             return batch_loss
         end
@@ -96,7 +96,7 @@ function  fit!(loss, penalty, chain, optimiser, epochs, verbosity, X, y)
 
     parameters = Flux.params(chain)
     losses = (loss(chain(X[i]), y[i]) +
-              penalty(parameters) for i in 1:n_batches)
+              penalty(parameters)/n_batches for i in 1:n_batches)
     history = [mean(losses),]
 
     for i in 1:epochs