diff --git a/e2e/docs/testdata/julia_mnist/build.envd b/e2e/docs/testdata/julia_mnist/build.envd
index 67440e761..d8ee4b08a 100644
--- a/e2e/docs/testdata/julia_mnist/build.envd
+++ b/e2e/docs/testdata/julia_mnist/build.envd
@@ -6,3 +6,4 @@ def build():
     install.julia()
     install.julia_packages(name=["Flux", "MLDatasets"])
     runtime.command(commands={"julia-mnist": "julia mlp_mnist.jl"})
+    runtime.environ(env={"DATADEPS_ALWAYS_ACCEPT": "true"})
diff --git a/e2e/docs/testdata/julia_mnist/mlp_mnist.jl b/e2e/docs/testdata/julia_mnist/mlp_mnist.jl
index 050e6dd79..26e0e3cef 100644
--- a/e2e/docs/testdata/julia_mnist/mlp_mnist.jl
+++ b/e2e/docs/testdata/julia_mnist/mlp_mnist.jl
@@ -68,9 +68,9 @@ train_loader = simple_loader(train_data, batchsize = 256)
 
 opt_state = Flux.setup(Adam(3e-4), model);
 
-# Then train for 10 epochs, printing out details as we go:
+# Then train for 3 epochs, printing out details as we go:
 
-for epoch in 1:10
+for epoch in 1:3
     loss = 0.0
     for (x, y) in train_loader
         # Compute the loss and the gradients: