Neural type switch with keen execution and Keras

How would your summer season vacation’s photographs look had Edvard Munch painted them? (Maybe it’s higher to not know).
Let’s take a extra comforting instance: How would a pleasant, summarly river panorama look if painted by Katsushika Hokusai?

Type switch on pictures just isn’t new, however bought a lift when Gatys, Ecker, and Bethge(Gatys, Ecker, and Bethge 2015) confirmed learn how to efficiently do it with deep studying.
The primary thought is easy: Create a hybrid that could be a tradeoff between the content material picture we need to manipulate, and a type picture we need to imitate, by optimizing for maximal resemblance to each on the similar time.

Should you’ve learn the chapter on neural type switch from Deep Studying with R, it’s possible you’ll acknowledge a number of the code snippets that observe.
Nevertheless, there is a crucial distinction: This submit makes use of TensorFlow Keen Execution, permitting for an crucial manner of coding that makes it simple to map ideas to code.
Similar to earlier posts on keen execution on this weblog, this can be a port of a Google Colaboratory pocket book that performs the identical activity in Python.

As standard, please ensure you have the required package deal variations put in. And no want to repeat the snippets – you’ll discover the entire code among the many Keras examples.

Conditions

The code on this submit is determined by the latest variations of a number of of the TensorFlow R packages. You possibly can set up these packages as follows:

set up.packages(c("tensorflow", "keras", "tfdatasets"))

library(tensorflow)
install_tensorflow()

There are extra necessities for utilizing TensorFlow keen execution. First, we have to name tfe_enable_eager_execution() proper initially of this system. Second, we have to use the implementation of Keras included in TensorFlow, moderately than the bottom Keras implementation.

Conditions behind us, let’s get began!

Enter pictures

Right here is our content material picture – substitute by a picture of your personal:

# If in case you have sufficient reminiscence in your GPU, no have to load the photographs
# at such small measurement.
# That is the dimensions I discovered working for a 4G GPU.
img_shape <- c(128, 128, 3)

content_path <- "isar.jpg"

content_image <-  image_load(content_path, target_size = img_shape[1:2])
content_image %>% 
  image_to_array() %>%
  `/`(., 255) %>%
  as.raster() %>%
  plot()

And right here’s the type mannequin, Hokusai’s The Nice Wave off Kanagawa, which you’ll be able to obtain from Wikimedia Commons:

style_path <- "The_Great_Wave_off_Kanagawa.jpg"

style_image <-  image_load(content_path, target_size = img_shape[1:2])
style_image %>% 
  image_to_array() %>%
  `/`(., 255) %>%
  as.raster() %>%
  plot()

We create a wrapper that masses and preprocesses the enter pictures for us.
As we shall be working with VGG19, a community that has been skilled on ImageNet, we have to rework our enter pictures in the identical manner that was used coaching it. Later, we’ll apply the inverse transformation to our mixture picture earlier than displaying it.

load_and_preprocess_image <- operate(path) {
  img <- image_load(path, target_size = img_shape[1:2]) %>%
    image_to_array() %>%
    k_expand_dims(axis = 1) %>%
    imagenet_preprocess_input()
}

deprocess_image <- operate(x) {
  x <- x[1, , ,]
  # Take away zero-center by imply pixel
  x[, , 1] <- x[, , 1] + 103.939
  x[, , 2] <- x[, , 2] + 116.779
  x[, , 3] <- x[, , 3] + 123.68
  # 'BGR'->'RGB'
  x <- x[, , c(3, 2, 1)]
  x[x > 255] <- 255
  x[x < 0] <- 0
  x[] <- as.integer(x) / 255
  x
}

Setting the scene

We’re going to use a neural community, however we received’t be coaching it. Neural type switch is a bit unusual in that we don’t optimize the community’s weights, however again propagate the loss to the enter layer (the picture), with the intention to transfer it within the desired path.

We shall be occupied with two sorts of outputs from the community, equivalent to our two objectives.
Firstly, we need to preserve the mix picture much like the content material picture, on a excessive degree. In a convnet, higher layers map to extra holistic ideas, so we’re selecting a layer excessive up within the graph to match outputs from the supply and the mix.

Secondly, the generated picture ought to “appear to be” the type picture. Type corresponds to decrease degree options like texture, shapes, strokes… So to match the mix in opposition to the type instance, we select a set of decrease degree conv blocks for comparability and mixture the outcomes.

content_layers <- c("block5_conv2")
style_layers <- c("block1_conv1",
                 "block2_conv1",
                 "block3_conv1",
                 "block4_conv1",
                 "block5_conv1")

num_content_layers <- size(content_layers)
num_style_layers <- size(style_layers)

get_model <- operate() {
  vgg <- application_vgg19(include_top = FALSE, weights = "imagenet")
  vgg$trainable <- FALSE
  style_outputs <- map(style_layers, operate(layer) vgg$get_layer(layer)$output)
  content_outputs <- map(content_layers, operate(layer) vgg$get_layer(layer)$output)
  model_outputs <- c(style_outputs, content_outputs)
  keras_model(vgg$enter, model_outputs)
}

Losses

When optimizing the enter picture, we’ll take into account three kinds of losses. Firstly, the content material loss: How completely different is the mix picture from the supply? Right here, we’re utilizing the sum of the squared errors for comparability.

content_loss <- operate(content_image, goal) {
  k_sum(k_square(goal - content_image))
}

Our second concern is having the types match as carefully as attainable. Type is often operationalized because the Gram matrix of flattened function maps in a layer. We thus assume that type is said to how maps in a layer correlate with different.

We subsequently compute the Gram matrices of the layers we’re occupied with (outlined above), for the supply picture in addition to the optimization candidate, and evaluate them, once more utilizing the sum of squared errors.

gram_matrix <- operate(x) {
  options <- k_batch_flatten(k_permute_dimensions(x, c(3, 1, 2)))
  gram <- k_dot(options, k_transpose(options))
  gram
}

style_loss <- operate(gram_target, mixture) {
  gram_comb <- gram_matrix(mixture)
  k_sum(k_square(gram_target - gram_comb)) /
    (4 * (img_shape[3] ^ 2) * (img_shape[1] * img_shape[2]) ^ 2)
}

Thirdly, we don’t need the mix picture to look overly pixelated, thus we’re including in a regularization element, the overall variation within the picture:

total_variation_loss <- operate(picture) {
  y_ij  <- picture[1:(img_shape[1] - 1L), 1:(img_shape[2] - 1L),]
  y_i1j <- picture[2:(img_shape[1]), 1:(img_shape[2] - 1L),]
  y_ij1 <- picture[1:(img_shape[1] - 1L), 2:(img_shape[2]),]
  a <- k_square(y_ij - y_i1j)
  b <- k_square(y_ij - y_ij1)
  k_sum(k_pow(a + b, 1.25))
}

The difficult factor is learn how to mix these losses. We’ve reached acceptable outcomes with the next weightings, however be happy to mess around as you see match:

content_weight <- 100
style_weight <- 0.8
total_variation_weight <- 0.01

Get mannequin outputs for the content material and magnificence pictures

We’d like the mannequin’s output for the content material and magnificence pictures, however right here it suffices to do that simply as soon as.
We concatenate each pictures alongside the batch dimension, go that enter to the mannequin, and get again an inventory of outputs, the place each factor of the record is a 4-d tensor. For the type picture, we’re within the type outputs at batch place 1, whereas for the content material picture, we’d like the content material output at batch place 2.

Within the under feedback, please be aware that the sizes of dimensions 2 and three will differ should you’re loading pictures at a unique measurement.

get_feature_representations <-
  operate(mannequin, content_path, style_path) {
    
    # dim == (1, 128, 128, 3)
    style_image <-
      load_and_process_image(style_path) %>% k_cast("float32")
    # dim == (1, 128, 128, 3)
    content_image <-
      load_and_process_image(content_path) %>% k_cast("float32")
    # dim == (2, 128, 128, 3)
    stack_images <- k_concatenate(record(style_image, content_image), axis = 1)
    
    # size(model_outputs) == 6
    # dim(model_outputs[[1]]) = (2, 128, 128, 64)
    # dim(model_outputs[[6]]) = (2, 8, 8, 512)
    model_outputs <- mannequin(stack_images)
    
    style_features <- 
      model_outputs[1:num_style_layers] %>%
      map(operate(batch) batch[1, , , ])
    content_features <- 
      model_outputs[(num_style_layers + 1):(num_style_layers + num_content_layers)] %>%
      map(operate(batch) batch[2, , , ])
    
    record(style_features, content_features)
  }

Computing the losses

On each iteration, we have to go the mix picture by the mannequin, receive the type and content material outputs, and compute the losses. Once more, the code is extensively commented with tensor sizes for straightforward verification, however please needless to say the precise numbers presuppose you’re working with 128×128 pictures.

compute_loss <-
  operate(mannequin, loss_weights, init_image, gram_style_features, content_features) {
    
    c(style_weight, content_weight) %<-% loss_weights
    model_outputs <- mannequin(init_image)
    style_output_features <- model_outputs[1:num_style_layers]
    content_output_features <-
      model_outputs[(num_style_layers + 1):(num_style_layers + num_content_layers)]
    
    # type loss
    weight_per_style_layer <- 1 / num_style_layers
    style_score <- 0
    # dim(style_zip[[5]][[1]]) == (512, 512)
    style_zip <- transpose(record(gram_style_features, style_output_features))
    for (l in 1:size(style_zip)) {
      # for l == 1:
      # dim(target_style) == (64, 64)
      # dim(comb_style) == (1, 128, 128, 64)
      c(target_style, comb_style) %<-% style_zip[[l]]
      style_score <- style_score + weight_per_style_layer * 
        style_loss(target_style, comb_style[1, , , ])
    }
    
    # content material loss
    weight_per_content_layer <- 1 / num_content_layers
    content_score <- 0
    content_zip <- transpose(record(content_features, content_output_features))
    for (l in 1:size(content_zip)) {
      # dim(comb_content) ==  (1, 8, 8, 512)
      # dim(target_content) == (8, 8, 512)
      c(target_content, comb_content) %<-% content_zip[[l]]
      content_score <- content_score + weight_per_content_layer *
        content_loss(comb_content[1, , , ], target_content)
    }
    
    # whole variation loss
    variation_loss <- total_variation_loss(init_image[1, , ,])
    
    style_score <- style_score * style_weight
    content_score <- content_score * content_weight
    variation_score <- variation_loss * total_variation_weight
    
    loss <- style_score + content_score + variation_score
    record(loss, style_score, content_score, variation_score)
  }

Computing the gradients

As quickly as we’ve the losses, acquiring the gradients of the general loss with respect to the enter picture is only a matter of calling tape$gradient on the GradientTape. Be aware that the nested name to compute_loss, and thus the decision of the mannequin on our mixture picture, occurs contained in the GradientTape context.

compute_grads <- 
  operate(mannequin, loss_weights, init_image, gram_style_features, content_features) {
    with(tf$GradientTape() %as% tape, {
      scores <-
        compute_loss(mannequin,
                     loss_weights,
                     init_image,
                     gram_style_features,
                     content_features)
    })
    total_loss <- scores[[1]]
    record(tape$gradient(total_loss, init_image), scores)
  }

Coaching section

Now it’s time to coach! Whereas the pure continuation of this sentence would have been “… the mannequin,” the mannequin we’re coaching right here just isn’t VGG19 (that one we’re simply utilizing as a software), however a minimal setup of simply:

a Variable that holds our to-be-optimized picture
the loss capabilities we outlined above
an optimizer that may apply the calculated gradients to the picture variable (tf$prepare$AdamOptimizer)

Under, we get the type options (of the type picture) and the content material function (of the content material picture) simply as soon as, then iterate over the optimization course of, saving the output each 100 iterations.

In distinction to the unique article and the Deep Studying with R guide, however following the Google pocket book as an alternative, we’re not utilizing L-BFGS for optimization, however Adam, as our objective right here is to offer a concise introduction to keen execution.
Nevertheless, you can plug in one other optimization technique should you needed, changing
optimizer$apply_gradients(record(tuple(grads, init_image)))
by an algorithm of your selection (and naturally, assigning the results of the optimization to the Variable holding the picture).

run_style_transfer <- operate(content_path, style_path) {
  mannequin <- get_model()
  stroll(mannequin$layers, operate(layer) layer$trainable = FALSE)
  
  c(style_features, content_features) %<-% 
    get_feature_representations(mannequin, content_path, style_path)
  # dim(gram_style_features[[1]]) == (64, 64)
  gram_style_features <- map(style_features, operate(function) gram_matrix(function))
  
  init_image <- load_and_process_image(content_path)
  init_image <- tf$contrib$keen$Variable(init_image, dtype = "float32")
  
  optimizer <- tf$prepare$AdamOptimizer(learning_rate = 1,
                                      beta1 = 0.99,
                                      epsilon = 1e-1)
  
  c(best_loss, best_image) %<-% record(Inf, NULL)
  loss_weights <- record(style_weight, content_weight)
  
  start_time <- Sys.time()
  global_start <- Sys.time()
  
  norm_means <- c(103.939, 116.779, 123.68)
  min_vals <- -norm_means
  max_vals <- 255 - norm_means
  
  for (i in seq_len(num_iterations)) {
    # dim(grads) == (1, 128, 128, 3)
    c(grads, all_losses) %<-% compute_grads(mannequin,
                                            loss_weights,
                                            init_image,
                                            gram_style_features,
                                            content_features)
    c(loss, style_score, content_score, variation_score) %<-% all_losses
    optimizer$apply_gradients(record(tuple(grads, init_image)))
    clipped <- tf$clip_by_value(init_image, min_vals, max_vals)
    init_image$assign(clipped)
    
    end_time <- Sys.time()
    
    if (k_cast_to_floatx(loss) < best_loss) {
      best_loss <- k_cast_to_floatx(loss)
      best_image <- init_image
    }
    
    if (i %% 50 == 0) {
      glue("Iteration: {i}") %>% print()
      glue(
        "Whole loss: {k_cast_to_floatx(loss)},
        type loss: {k_cast_to_floatx(style_score)},
        content material loss: {k_cast_to_floatx(content_score)},
        whole variation loss: {k_cast_to_floatx(variation_score)},
        time for 1 iteration: {(Sys.time() - start_time) %>% spherical(2)}"
      ) %>% print()
      
      if (i %% 100 == 0) {
        png(paste0("style_epoch_", i, ".png"))
        plot_image <- best_image$numpy()
        plot_image <- deprocess_image(plot_image)
        plot(as.raster(plot_image), fundamental = glue("Iteration {i}"))
        dev.off()
      }
    }
  }
  
  glue("Whole time: {Sys.time() - global_start} seconds") %>% print()
  record(best_image, best_loss)
}

Able to run

Now, we’re prepared to start out the method:

c(best_image, best_loss) %<-% run_style_transfer(content_path, style_path)

In our case, outcomes didn’t change a lot after ~ iteration 1000, and that is how our river panorama was trying:

… positively extra inviting than had it been painted by Edvard Munch!

Conclusion

With neural type switch, some fiddling round could also be wanted till you get the end result you need. However as our instance reveals, this doesn’t imply the code must be difficult. Moreover to being simple to understand, keen execution additionally helps you to add debugging output, and step by the code line-by-line to verify on tensor shapes.
Till subsequent time in our keen execution sequence!

Gatys, Leon A., Alexander S. Ecker, and Matthias Bethge. 2015. “A Neural Algorithm of Creative Type.” CoRR abs/1508.06576. http://arxiv.org/abs/1508.06576.