Skip to content

Update examples #1026

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
skeydan merged 4 commits into from
Apr 27, 2020
Merged

Update examples #1026

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
dd57faa
tf2 compatibility
turgut090 Apr 24, 2020
a411196
minor fixes
turgut090 Apr 24, 2020
56cf6a4
"tfprobability" instead of "reticulate::import"
turgut090 Apr 24, 2020
f42c7b7
"tf==2.0.0" compatibility & "tfprobability"
turgut090 Apr 24, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 4 additions & 11 deletions vignettes/examples/eager_image_captioning.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,9 +5,7 @@
#' https://blogs.rstudio.com/tensorflow/posts/2018-09-17-eager-captioning

library(keras)
use_implementation("tensorflow")
library(tensorflow)
tfe_enable_eager_execution(device_policy = "silent")

np <- import("numpy")

Expand All @@ -32,10 +30,6 @@ debugshapes <- FALSE
restore_checkpoint <- FALSE
saved_features_exist <- FALSE

use_session_with_seed(7777,
disable_gpu = FALSE,
disable_parallel_cpu = FALSE)

annotation_file <- "train2014/annotations/captions_train2014.json"
image_path <- "train2014/train2014"

Expand Down Expand Up @@ -138,7 +132,7 @@ top_k <- 5000
tokenizer <- text_tokenizer(num_words = top_k,
oov_token = "<unk>",
filters = '!"#$%&()*+.,-/:;=?@[\\]^_`{|}~ ')
tokenizer$fit_on_texts(sample_captions)
fit_text_tokenizer(tokenizer, sample_captions)
train_captions_tokenized <-
tokenizer %>% texts_to_sequences(train_captions)
validation_captions_tokenized <-
Expand Down Expand Up @@ -216,7 +210,7 @@ map_func <- function(img_name, cap) {
train_dataset <-
tensor_slices_dataset(list(train_images, train_captions_padded)) %>%
dataset_map(function(item1, item2)
tf$py_func(map_func, list(item1, item2), list(tf$float32, tf$int32))) %>%
tf$py_function(map_func, list(item1, item2), list(tf$float32, tf$int32))) %>%
# dataset_shuffle(buffer_size) %>%
dataset_batch(batch_size)

Expand Down Expand Up @@ -360,7 +354,7 @@ rnn_decoder <-
encoder <- cnn_encoder(embedding_dim)
decoder <- rnn_decoder(embedding_dim, gru_units, vocab_size)

optimizer = tf$train$AdamOptimizer()
optimizer = tf$optimizers$Adam()

cx_loss <- function(y_true, y_pred) {
mask <- 1 - k_cast(y_true == 0L, dtype = "float32")
Expand Down Expand Up @@ -530,8 +524,7 @@ if (!restore_checkpoint) {
variables <- c(encoder$variables, decoder$variables)
gradients <- tape$gradient(loss, variables)

optimizer$apply_gradients(purrr::transpose(list(gradients, variables)),
global_step = tf$train$get_or_create_global_step())
optimizer$apply_gradients(purrr::transpose(list(gradients, variables)))
})
cat(paste0(
"\n\nTotal loss (epoch): ",
Expand Down
17 changes: 14 additions & 3 deletions vignettes/examples/nmt_attention.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,17 @@ library(tibble)
# Assumes you've downloaded and unzipped one of the bilingual datasets offered at
# http://www.manythings.org/anki/ and put it into a directory "data"
# This example translates English to Dutch.
download_data = function(){
if(!dir.exists('data')) {
dir.create('data')
}
if(!file.exists('data/nld-eng.zip')) {
download.file('http://www.manythings.org/anki/nld-eng.zip',
destfile = file.path("data", basename('nld-eng.zip')))
unzip('data/nld-eng.zip', exdir = 'data')
}
}
download_data()

filepath <- file.path("data", "nld.txt")

Expand Down Expand Up @@ -290,7 +301,7 @@ evaluate <-
attention_matrix[t,] <- attention_weights %>% as.double()

pred_idx <-
tf$compat$v1$multinomial(k_exp(preds), num_samples = 1L)[1, 1] %>% as.double()
tf$random$categorical(k_exp(preds), num_samples = 1L)[1, 1] %>% as.double()
pred_word <- index2word(pred_idx, target_index)

if (pred_word == '<stop>') {
Expand Down Expand Up @@ -387,7 +398,7 @@ for (epoch in seq_len(n_epochs)) {
": ",
(loss / k_cast_to_floatx(dim(y)[2])) %>% as.double() %>% round(4),
"\n"
) %>% print()
) %>% cat()

variables <- c(encoder$variables, decoder$variables)
gradients <- tape$gradient(loss, variables)
Expand All @@ -402,7 +413,7 @@ for (epoch in seq_len(n_epochs)) {
": ",
(total_loss / k_cast_to_floatx(buffer_size)) %>% as.double() %>% round(4),
"\n"
) %>% print()
) %>% cat()

walk(train_sentences[1:5], function(pair)
translate(pair[1]))
Expand Down
58 changes: 30 additions & 28 deletions vignettes/examples/tfprob_vae.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,13 +5,8 @@
#' https://blogs.rstudio.com/tensorflow/posts/2019-01-08-getting-started-with-tf-probability/

library(keras)
use_implementation("tensorflow")
library(tensorflow)
tfe_enable_eager_execution(device_policy = "silent")

tfp <- import("tensorflow_probability")
tfd <- tfp$distributions

library(tfprobability)
library(tfdatasets)
library(dplyr)
library(glue)
Expand Down Expand Up @@ -79,6 +74,17 @@ np <- import("numpy")

# assume data have been downloaded from https://github.com/rois-codh/kmnist
# and stored in /tmp
download_data = function(){
if(!dir.exists('tmp')) {
dir.create('tmp')
}
if(!file.exists('tmp/kmnist-train-imgs.npz')) {
download.file('http://codh.rois.ac.jp/kmnist/dataset/kmnist/kmnist-train-imgs.npz',
destfile = file.path("tmp", basename('kmnist-train-imgs.npz')))
}
}
download_data()

kuzushiji <- np$load("/tmp/kmnist-train-imgs.npz")
kuzushiji <- kuzushiji$get("arr_0")

Expand All @@ -98,8 +104,8 @@ train_dataset <- tensor_slices_dataset(train_images) %>%

# Params ------------------------------------------------------------------

latent_dim <- 2
mixture_components <- 16
latent_dim <- 2L
mixture_components <- 16L


# Model -------------------------------------------------------------------
Expand Down Expand Up @@ -132,8 +138,8 @@ encoder_model <- function(name = NULL) {
self$conv2() %>%
self$flatten() %>%
self$dense()
tfd$MultivariateNormalDiag(loc = x[, 1:latent_dim],
scale_diag = tf$nn$softplus(x[, (latent_dim + 1):(2 * latent_dim)] + 1e-5))
tfd_multivariate_normal_diag(loc = x[, 1:latent_dim],
scale_diag = tf$nn$softplus(x[, (latent_dim + 1):(2 * latent_dim)] + 1e-5))
}
})
}
Expand Down Expand Up @@ -178,7 +184,7 @@ decoder_model <- function(name = NULL) {
self$deconv2() %>%
self$deconv3()

tfd$Independent(tfd$Bernoulli(logits = x),
tfd_independent(tfd_bernoulli(logits = x),
reinterpreted_batch_ndims = 3L)

}
Expand All @@ -192,30 +198,30 @@ learnable_prior_model <-

keras_model_custom(name = name, function(self) {
self$loc <-
tf$get_variable(
tf$compat$v1$get_variable(
name = "loc",
shape = list(mixture_components, latent_dim),
dtype = tf$float32
)
self$raw_scale_diag <- tf$get_variable(
self$raw_scale_diag <- tf$compat$v1$get_variable(
name = "raw_scale_diag",
shape = c(mixture_components, latent_dim),
dtype = tf$float32
)
self$mixture_logits <-
tf$get_variable(
tf$compat$v1$get_variable(
name = "mixture_logits",
shape = c(mixture_components),
dtype = tf$float32
)

function (x, mask = NULL) {
tfd$MixtureSameFamily(
components_distribution = tfd$MultivariateNormalDiag(
tfd_mixture_same_family(
components_distribution = tfd_multivariate_normal_diag(
loc = self$loc,
scale_diag = tf$nn$softplus(self$raw_scale_diag)
),
mixture_distribution = tfd$Categorical(logits = self$mixture_logits)
mixture_distribution = tfd_categorical(logits = self$mixture_logits)
)
}
})
Expand All @@ -234,8 +240,7 @@ compute_kl_loss <-
}


global_step <- tf$train$get_or_create_global_step()
optimizer <- tf$train$AdamOptimizer(1e-4)
optimizer <- tf$optimizers$Adam(1e-4)


# Training loop -----------------------------------------------------------
Expand All @@ -253,7 +258,6 @@ checkpoint_prefix <- file.path(checkpoint_dir, "ckpt")
checkpoint <-
tf$train$Checkpoint(
optimizer = optimizer,
global_step = global_step,
encoder = encoder,
decoder = decoder,
latent_prior_model = latent_prior_model
Expand Down Expand Up @@ -284,7 +288,7 @@ for (epoch in seq_len(num_epochs)) {
compute_kl_loss(latent_prior,
approx_posterior,
approx_posterior_sample)

loss <- kl_loss + avg_nll
})

Expand All @@ -299,18 +303,15 @@ for (epoch in seq_len(num_epochs)) {

optimizer$apply_gradients(purrr::transpose(list(
encoder_gradients, encoder$variables
)),
global_step = tf$train$get_or_create_global_step())
)))
optimizer$apply_gradients(purrr::transpose(list(
decoder_gradients, decoder$variables
)),
global_step = tf$train$get_or_create_global_step())
)))
optimizer$apply_gradients(purrr::transpose(list(
prior_gradients, latent_prior_model$variables
)),
global_step = tf$train$get_or_create_global_step())
)))

})
})

checkpoint$save(file_prefix = checkpoint_prefix)

Expand All @@ -329,3 +330,4 @@ for (epoch in seq_len(num_epochs)) {
show_grid(epoch)
}
}

50 changes: 27 additions & 23 deletions vignettes/examples/vq_vae.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,20 +5,19 @@
#' https://blogs.rstudio.com/tensorflow/posts/2019-01-24-vq-vae/

library(keras)
use_implementation("tensorflow")
library(tensorflow)
tfe_enable_eager_execution(device_policy = "silent")

use_session_with_seed(7778,
disable_gpu = FALSE,
disable_parallel_cpu = FALSE)

tfp <- import("tensorflow_probability")
tfd <- tfp$distributions

library(tfprobability)
library(tfdatasets)

library(dplyr)
library(glue)

# curry has to be installed from github because CRAN version has no "set_defaults" function
if(!('devtools' %in% rownames(installed.packages()) )) {
install.packages('devtools')
}
devtools::install_github('thomasp85/curry')

library(curry)

moving_averages <- tf$python$training$moving_averages
Expand Down Expand Up @@ -63,7 +62,14 @@ write_png <- function(dataset, epoch, desc, images) {

np <- import("numpy")

# download from: https://github.com/rois-codh/kmnist
# download from: https://github.com/rois-codh/kmnist via "download_data()" function
download_data = function(){
if(!file.exists('kmnist-train-imgs.npz')) {
download.file('http://codh.rois.ac.jp/kmnist/dataset/kmnist/kmnist-train-imgs.npz',
destfile = 'kmnist-train-imgs.npz')
}
}
download_data()
kuzushiji <- np$load("kmnist-train-imgs.npz")
kuzushiji <- kuzushiji$get("arr_0")

Expand All @@ -90,7 +96,7 @@ batch %>% dim()
# Params ------------------------------------------------------------------

learning_rate <- 0.001
latent_size <- 1
latent_size <- 1L
num_codes <- 64L
code_size <- 16L
base_depth <- 32
Expand Down Expand Up @@ -214,7 +220,7 @@ decoder_model <- function(name = NULL,
self$deconv6() %>%
# output shape: 7 28 28 1
self$conv1()
tfd$Independent(tfd$Bernoulli(logits = x),
tfd_independent(tfd_bernoulli(logits = x),
reinterpreted_batch_ndims = length(output_shape))
}
})
Expand All @@ -228,16 +234,16 @@ vector_quantizer_model <-
keras_model_custom(name = name, function(self) {
self$num_codes <- num_codes
self$code_size <- code_size
self$codebook <- tf$get_variable("codebook",
self$codebook <- tf$compat$v1$get_variable("codebook",
shape = c(num_codes, code_size),
dtype = tf$float32)
self$ema_count <- tf$get_variable(
self$ema_count <- tf$compat$v1$get_variable(
name = "ema_count",
shape = c(num_codes),
initializer = tf$constant_initializer(0),
trainable = FALSE
)
self$ema_means = tf$get_variable(
self$ema_means = tf$compat$v1$get_variable(
name = "ema_means",
initializer = self$codebook$initialized_value(),
trainable = FALSE
Expand Down Expand Up @@ -308,7 +314,7 @@ update_ema <- function(vector_quantizer,
updated_ema_means <-
updated_ema_means / tf$expand_dims(updated_ema_count, axis = -1L)

tf$assign(vector_quantizer$codebook, updated_ema_means)
tf$compat$v1$assign(vector_quantizer$codebook, updated_ema_means)
}


Expand All @@ -321,7 +327,7 @@ decoder <- decoder_model(input_size = latent_size * code_size,
vector_quantizer <-
vector_quantizer_model(num_codes = num_codes, code_size = code_size)

optimizer <- tf$train$AdamOptimizer(learning_rate = learning_rate)
optimizer <- tf$optimizers$Adam(learning_rate = learning_rate)

checkpoint_dir <- "./vq_vae_checkpoints"

Expand Down Expand Up @@ -365,7 +371,7 @@ for (epoch in seq_len(num_epochs)) {

commitment_loss <- tf$reduce_mean(tf$square(codes - tf$stop_gradient(nearest_codebook_entries)))

prior_dist <- tfd$Multinomial(total_count = 1,
prior_dist <- tfd_multinomial(total_count = 1,
logits = tf$zeros(c(latent_size, num_codes)))
prior_loss <- -tf$reduce_mean(tf$reduce_sum(prior_dist$log_prob(one_hot_assignments), 1L))

Expand All @@ -379,12 +385,10 @@ for (epoch in seq_len(num_epochs)) {

optimizer$apply_gradients(purrr::transpose(list(
encoder_gradients, encoder$variables
)),
global_step = tf$train$get_or_create_global_step())
)))
optimizer$apply_gradients(purrr::transpose(list(
decoder_gradients, decoder$variables
)),
global_step = tf$train$get_or_create_global_step())
)))

update_ema(vector_quantizer,
one_hot_assignments,
Expand Down