rerender guides

Author: t-kalinowski

vignettes/custom_train_step_in_tensorflow.Rmd

@@ -127,11 +127,11 @@ model |> fit(x, y, epochs = 3)
 
 ```
 ## Epoch 1/3
-## 32/32 - 0s - 6ms/step - loss: 2.9118 - mae: 1.3597
+## 32/32 - 1s - 33ms/step - loss: 2.9118 - mae: 1.3597
 ## Epoch 2/3
-## 32/32 - 0s - 513us/step - loss: 2.6026 - mae: 1.2856
+## 32/32 - 0s - 1ms/step - loss: 2.6026 - mae: 1.2856
 ## Epoch 3/3
-## 32/32 - 0s - 509us/step - loss: 2.3378 - mae: 1.2193
+## 32/32 - 0s - 1ms/step - loss: 2.3378 - mae: 1.2193
 ```
 
 ## Going lower-level
@@ -212,11 +212,11 @@ model |> fit(x, y, epochs = 3)
 
 ```
 ## Epoch 1/3
-## 32/32 - 0s - 5ms/step - loss: 2.6540 - mae: 1.2901
+## 32/32 - 1s - 23ms/step - loss: 2.6540 - mae: 1.2901
 ## Epoch 2/3
-## 32/32 - 0s - 527us/step - loss: 2.4139 - mae: 1.2303
+## 32/32 - 0s - 1ms/step - loss: 2.4139 - mae: 1.2303
 ## Epoch 3/3
-## 32/32 - 0s - 522us/step - loss: 2.2080 - mae: 1.1761
+## 32/32 - 0s - 1ms/step - loss: 2.2080 - mae: 1.1761
 ```
 
 ## Supporting `sample_weight` & `class_weight`
@@ -282,11 +282,11 @@ model |> fit(x, y, sample_weight = sw, epochs = 3)
 
 ```
 ## Epoch 1/3
-## 32/32 - 0s - 7ms/step - loss: 0.1607 - mae: 1.3018
+## 32/32 - 1s - 24ms/step - loss: 0.1607 - mae: 1.3018
 ## Epoch 2/3
-## 32/32 - 0s - 540us/step - loss: 0.1452 - mae: 1.2999
+## 32/32 - 0s - 1ms/step - loss: 0.1452 - mae: 1.2999
 ## Epoch 3/3
-## 32/32 - 0s - 529us/step - loss: 0.1335 - mae: 1.2986
+## 32/32 - 0s - 1ms/step - loss: 0.1335 - mae: 1.2986
 ```
 
 ## Providing your own evaluation step
@@ -332,7 +332,7 @@ model |> evaluate(x, y)
 ```
 
 ```
-## 32/32 - 0s - 2ms/step - loss: 0.0000e+00 - mae: 1.3947
+## 32/32 - 0s - 11ms/step - loss: 0.0000e+00 - mae: 1.3947
 ```
 
 ```
@@ -507,7 +507,7 @@ gan |> fit(
 ```
 
 ```
-## 100/100 - 21s - 206ms/step - d_loss: 0.0000e+00 - g_loss: 0.0000e+00
+## 100/100 - 7s - 70ms/step - d_loss: 0.0000e+00 - g_loss: 0.0000e+00
 ```
 
 The ideas behind deep learning are simple, so why should their implementation be painful?

vignettes/distributed_training_with_tensorflow.Rmd

@@ -115,7 +115,7 @@ Here's a simple end-to-end runnable example:
 
 ```r
 get_compiled_model <- function() {
-  inputs <- layer_input(shape = 784)
+  inputs <- keras_input(shape = 784)
   outputs <- inputs |>
     layer_dense(units = 256, activation = "relu") |>
     layer_dense(units = 256, activation = "relu") |>
@@ -195,7 +195,7 @@ with(strategy$scope(), {
 ## Epoch 1/2
 ## 782/782 - 7s - 9ms/step - loss: 2.9199 - sparse_categorical_accuracy: 0.8580 - val_loss: 0.9377 - val_sparse_categorical_accuracy: 0.8986
 ## Epoch 2/2
-## 782/782 - 3s - 4ms/step - loss: 0.5622 - sparse_categorical_accuracy: 0.9254 - val_loss: 0.7505 - val_sparse_categorical_accuracy: 0.9044
+## 782/782 - 4s - 5ms/step - loss: 0.5622 - sparse_categorical_accuracy: 0.9254 - val_loss: 0.7505 - val_sparse_categorical_accuracy: 0.9044
 ## 157/157 - 0s - 3ms/step - loss: 0.8129 - sparse_categorical_accuracy: 0.9042
 ```
 
@@ -274,7 +274,7 @@ run_training(epochs = 1)
 ```
 
 ```
-## 782/782 - 4s - 5ms/step - loss: 0.5186 - sparse_categorical_accuracy: 0.9351 - val_loss: 0.7081 - val_sparse_categorical_accuracy: 0.9184
+## 782/782 - 6s - 8ms/step - loss: 3.3305 - sparse_categorical_accuracy: 0.8592 - val_loss: 1.1806 - val_sparse_categorical_accuracy: 0.8930
 ```
 
 ```r
@@ -283,7 +283,7 @@ run_training(epochs = 1)
 ```
 
 ```
-## 782/782 - 4s - 5ms/step - loss: 0.3189 - sparse_categorical_accuracy: 0.9476 - val_loss: 0.5234 - val_sparse_categorical_accuracy: 0.9382
+## 782/782 - 4s - 5ms/step - loss: 0.6629 - sparse_categorical_accuracy: 0.9215 - val_loss: 0.8294 - val_sparse_categorical_accuracy: 0.9130
 ```
 
 ## `tf$data` performance tips

vignettes/functional_api.Rmd

@@ -149,7 +149,7 @@ You can also plot the model as a graph:
 plot(model)
 ```
 
-<img src="functional_api/unnamed-chunk-10-1.png" width="155" />
+<img src="functional_api/unnamed-chunk-10-1.png" width="176" />
 
 And, optionally, display the input and output shapes of each layer
 in the plotted graph:
@@ -159,7 +159,7 @@ in the plotted graph:
 plot(model, show_shapes = TRUE)
 ```
 
-<img src="functional_api/unnamed-chunk-11-1.png" width="261" />
+<img src="functional_api/unnamed-chunk-11-1.png" width="311" />
 
 This figure and the code are almost identical. In the code version,
 the connection arrows are replaced by the call operation.
@@ -202,17 +202,17 @@ history <- model |> fit(
 
 ```
 ## Epoch 1/2
-## 750/750 - 1s - 1ms/step - accuracy: 0.9015 - loss: 0.3513 - val_accuracy: 0.9442 - val_loss: 0.1921
+## 750/750 - 2s - 3ms/step - accuracy: 0.9015 - loss: 0.3513 - val_accuracy: 0.9442 - val_loss: 0.1921
 ## Epoch 2/2
-## 750/750 - 1s - 699us/step - accuracy: 0.9508 - loss: 0.1643 - val_accuracy: 0.9581 - val_loss: 0.1423
+## 750/750 - 1s - 1ms/step - accuracy: 0.9508 - loss: 0.1643 - val_accuracy: 0.9581 - val_loss: 0.1423
 ```
 
 ```r
 test_scores <- model |> evaluate(x_test, y_test, verbose=2)
 ```
 
 ```
-## 313/313 - 0s - 335us/step - accuracy: 0.9578 - loss: 0.1338
+## 313/313 - 0s - 2ms/step - accuracy: 0.9578 - loss: 0.1338
 ```
 
 
@@ -581,7 +581,7 @@ Now plot the model:
 plot(model, show_shapes = TRUE)
 ```
 
-<img src="functional_api/unnamed-chunk-20-1.png" width="851" />
+<img src="functional_api/unnamed-chunk-20-1.png" width="1008" />
 
 When compiling this model, you can assign different losses to each output.
 You can even assign different weights to each loss -- to modulate
@@ -637,9 +637,9 @@ model |> fit(
 
 ```
 ## Epoch 1/2
-## 40/40 - 2s - 39ms/step - loss: 0.3948
+## 40/40 - 3s - 68ms/step - loss: 0.3948
 ## Epoch 2/2
-## 40/40 - 1s - 16ms/step - loss: 0.1971
+## 40/40 - 0s - 5ms/step - loss: 0.1971
 ```
 
 When calling fit with a `Dataset` object, it should yield either a
@@ -741,7 +741,7 @@ Plot the model:
 plot(model, show_shapes = TRUE)
 ```
 
-<img src="functional_api/unnamed-chunk-25-1.png" width="389" />
+<img src="functional_api/unnamed-chunk-25-1.png" width="468" />
 
 Now train the model:
 
@@ -770,7 +770,7 @@ model |> fit(
 ```
 
 ```
-## 13/13 - 1s - 93ms/step - acc: 0.1238 - loss: 2.3109 - val_acc: 0.1150 - val_loss: 2.3210
+## 13/13 - 9s - 674ms/step - acc: 0.1325 - loss: 2.3035 - val_acc: 0.1150 - val_loss: 2.3076
 ```
 
 ## Shared layers

vignettes/getting_started.Rmd

@@ -196,15 +196,15 @@ model |> evaluate(x_test, y_test)
 ```
 
 ```
-## 313/313 - 0s - 431us/step - accuracy: 0.9825 - loss: 0.0907
+## 313/313 - 1s - 3ms/step - accuracy: 0.9816 - loss: 0.0906
 ```
 
 ```
 ## $accuracy
-## [1] 0.9825
+## [1] 0.9816
 ##
 ## $loss
-## [1] 0.09072524
+## [1] 0.09057788
 ```
 
 Generate predictions on new data:
@@ -215,7 +215,7 @@ probs <- model |> predict(x_test)
 ```
 
 ```
-## 313/313 - 0s - 494us/step
+## 313/313 - 0s - 2ms/step
 ```
 
 ```r

vignettes/getting_started/unnamed-chunk-12-1.png

@@ -174,25 +174,25 @@ model |> fit(
 
 ```
 ## Epoch 1/10
-## 399/399 - 293s - 734ms/step - acc: 0.7466 - loss: 0.7492 - val_acc: 0.9618 - val_loss: 0.1349
+## 399/399 - 10s - 24ms/step - acc: 0.7362 - loss: 0.7750 - val_acc: 0.9622 - val_loss: 0.1377
 ## Epoch 2/10
-## 399/399 - 40s - 100ms/step - acc: 0.9326 - loss: 0.2262 - val_acc: 0.9756 - val_loss: 0.0807
+## 399/399 - 2s - 5ms/step - acc: 0.9304 - loss: 0.2354 - val_acc: 0.9702 - val_loss: 0.0983
 ## Epoch 3/10
-## 399/399 - 39s - 98ms/step - acc: 0.9529 - loss: 0.1580 - val_acc: 0.9807 - val_loss: 0.0671
+## 399/399 - 2s - 5ms/step - acc: 0.9480 - loss: 0.1740 - val_acc: 0.9791 - val_loss: 0.0685
 ## Epoch 4/10
-## 399/399 - 38s - 95ms/step - acc: 0.9612 - loss: 0.1268 - val_acc: 0.9826 - val_loss: 0.0576
+## 399/399 - 2s - 5ms/step - acc: 0.9593 - loss: 0.1395 - val_acc: 0.9848 - val_loss: 0.0545
 ## Epoch 5/10
-## 399/399 - 39s - 98ms/step - acc: 0.9695 - loss: 0.1079 - val_acc: 0.9857 - val_loss: 0.0490
+## 399/399 - 2s - 5ms/step - acc: 0.9656 - loss: 0.1179 - val_acc: 0.9872 - val_loss: 0.0451
 ## Epoch 6/10
-## 399/399 - 39s - 98ms/step - acc: 0.9726 - loss: 0.0937 - val_acc: 0.9878 - val_loss: 0.0433
+## 399/399 - 2s - 5ms/step - acc: 0.9692 - loss: 0.1026 - val_acc: 0.9868 - val_loss: 0.0466
 ## Epoch 7/10
-## 399/399 - 38s - 95ms/step - acc: 0.9749 - loss: 0.0869 - val_acc: 0.9879 - val_loss: 0.0430
+## 399/399 - 2s - 5ms/step - acc: 0.9737 - loss: 0.0916 - val_acc: 0.9890 - val_loss: 0.0394
 ## Epoch 8/10
-## 399/399 - 38s - 96ms/step - acc: 0.9776 - loss: 0.0757 - val_acc: 0.9894 - val_loss: 0.0367
+## 399/399 - 2s - 5ms/step - acc: 0.9753 - loss: 0.0834 - val_acc: 0.9898 - val_loss: 0.0357
 ## Epoch 9/10
-## 399/399 - 38s - 95ms/step - acc: 0.9792 - loss: 0.0701 - val_acc: 0.9896 - val_loss: 0.0398
+## 399/399 - 2s - 5ms/step - acc: 0.9778 - loss: 0.0750 - val_acc: 0.9910 - val_loss: 0.0313
 ## Epoch 10/10
-## 399/399 - 39s - 97ms/step - acc: 0.9815 - loss: 0.0632 - val_acc: 0.9902 - val_loss: 0.0378
+## 399/399 - 2s - 5ms/step - acc: 0.9799 - loss: 0.0695 - val_acc: 0.9882 - val_loss: 0.0381
 ```
 
 ```r
@@ -223,7 +223,7 @@ predictions <- model |> predict(x_test)
 ```
 
 ```
-## 313/313 - 3s - 9ms/step
+## 313/313 - 1s - 2ms/step
 ```
 
 ```r
@@ -358,7 +358,7 @@ model |> fit(
 ```
 
 ```
-## 399/399 - 39s - 99ms/step - acc: 0.7494 - loss: 0.7321 - val_acc: 0.9291 - val_loss: 0.2322
+## 399/399 - 9s - 22ms/step - acc: 0.7409 - loss: 0.7537 - val_acc: 0.9232 - val_loss: 0.2499
 ```
 
 ## Training models on arbitrary data sources
@@ -435,7 +435,7 @@ model |> fit(train_dataset, epochs = 1, validation_data = test_dataset)
 ```
 
 ```
-## 469/469 - 46s - 98ms/step - acc: 0.7528 - loss: 0.7348 - val_acc: 0.9235 - val_loss: 0.2524
+## 469/469 - 10s - 21ms/step - acc: 0.7581 - loss: 0.7151 - val_acc: 0.8985 - val_loss: 0.3161
 ```
 
 ## Further reading

vignettes/intro_to_keras_for_engineers.Rmd

@@ -480,7 +480,7 @@ model |> fit(random_normal(c(2, 3)), random_normal(c(2, 3)), epochs = 1)
 ```
 
 ```
-## 1/1 - 0s - 37ms/step - loss: 1.8971
+## 1/1 - 0s - 132ms/step - loss: 1.8971
 ```
 
 ```r
@@ -492,7 +492,7 @@ model |> fit(random_normal(c(2, 3)), random_normal(c(2, 3)), epochs = 1)
 ```
 
 ```
-## 1/1 - 0s - 26ms/step - loss: -3.3344e-03
+## 1/1 - 0s - 104ms/step - loss: -3.3344e-03
 ```
 
 ## You can optionally enable serialization on your layers
@@ -834,7 +834,7 @@ vae |> fit(x_train, x_train, epochs = 2, batch_size = 64)
 
 ```
 ## Epoch 1/2
-## 938/938 - 2s - 2ms/step - loss: 0.0748
+## 938/938 - 6s - 7ms/step - loss: 0.0748
 ## Epoch 2/2
 ## 938/938 - 1s - 1ms/step - loss: 0.0676
 ```