Morph MORPH
SIGN IN SIGN UP
pytorch / examples UNCLAIMED

A set of examples around pytorch in Vision, Text, Reinforcement Learning, etc.

23827 0 0 Python
Normal View History Raw
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
#include <torch/torch.h>
Improve code readability and make number of epochs a command line argument (#1222) * Change the cpp/dcgan * Use an open source argparse implementation
2024-01-24 21:20:29 -08:00
#include <argparse/argparse.hpp>
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
#include <cmath>
#include <cstdio>
#include <iostream>
// The size of the noise vector fed to the generator.
const int64_t kNoiseSize = 100;
// The batch size for training.
const int64_t kBatchSize = 64;
// Where to find the MNIST dataset.
const char* kDataFolder = "./data";
// After how many batches to create a new checkpoint periodically.
const int64_t kCheckpointEvery = 200;
// How many images to sample at every checkpoint.
const int64_t kNumberOfSamplesPerCheckpoint = 10;
// Set to `true` to restore models and optimizers from previously saved
// checkpoints.
const bool kRestoreFromCheckpoint = false;
// After how many batches to log a new update with the loss value.
const int64_t kLogInterval = 10;
using namespace torch;
Update C++ frontend examples for v1.4.0 (#697) * Update C++ DCGAN example for v1.4.0 * fix Reduction::Sum * FeatureDropout -> Dropout2d
2020-01-16 18:21:52 -05:00
struct DCGANGeneratorImpl : nn::Module {
DCGANGeneratorImpl(int kNoiseSize)
: conv1(nn::ConvTranspose2dOptions(kNoiseSize, 256, 4)
.bias(false)),
batch_norm1(256),
conv2(nn::ConvTranspose2dOptions(256, 128, 3)
.stride(2)
.padding(1)
.bias(false)),
batch_norm2(128),
conv3(nn::ConvTranspose2dOptions(128, 64, 4)
.stride(2)
.padding(1)
.bias(false)),
batch_norm3(64),
conv4(nn::ConvTranspose2dOptions(64, 1, 4)
.stride(2)
.padding(1)
.bias(false))
{
// register_module() is needed if we want to use the parameters() method later on
register_module("conv1", conv1);
register_module("conv2", conv2);
register_module("conv3", conv3);
register_module("conv4", conv4);
register_module("batch_norm1", batch_norm1);
register_module("batch_norm2", batch_norm2);
register_module("batch_norm3", batch_norm3);
}
torch::Tensor forward(torch::Tensor x) {
x = torch::relu(batch_norm1(conv1(x)));
x = torch::relu(batch_norm2(conv2(x)));
x = torch::relu(batch_norm3(conv3(x)));
x = torch::tanh(conv4(x));
return x;
}
nn::ConvTranspose2d conv1, conv2, conv3, conv4;
nn::BatchNorm2d batch_norm1, batch_norm2, batch_norm3;
};
TORCH_MODULE(DCGANGenerator);
Improve code readability and make number of epochs a command line argument (#1222) * Change the cpp/dcgan * Use an open source argparse implementation
2024-01-24 21:20:29 -08:00
nn::Sequential create_discriminator() {
return nn::Sequential(
// Layer 1
nn::Conv2d(nn::Conv2dOptions(1, 64, 4).stride(2).padding(1).bias(false)),
nn::LeakyReLU(nn::LeakyReLUOptions().negative_slope(0.2)),
// Layer 2
nn::Conv2d(nn::Conv2dOptions(64, 128, 4).stride(2).padding(1).bias(false)),
nn::BatchNorm2d(128),
nn::LeakyReLU(nn::LeakyReLUOptions().negative_slope(0.2)),
// Layer 3
nn::Conv2d(
nn::Conv2dOptions(128, 256, 4).stride(2).padding(1).bias(false)),
nn::BatchNorm2d(256),
nn::LeakyReLU(nn::LeakyReLUOptions().negative_slope(0.2)),
// Layer 4
nn::Conv2d(nn::Conv2dOptions(256, 1, 3).stride(1).padding(0).bias(false)),
nn::Sigmoid());
}
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
int main(int argc, const char* argv[]) {
Improve code readability and make number of epochs a command line argument (#1222) * Change the cpp/dcgan * Use an open source argparse implementation
2024-01-24 21:20:29 -08:00
argparse::ArgumentParser parser("cpp/dcgan example");
parser.add_argument("--epochs")
.help("Number of epochs to train")
.default_value(std::int64_t{30})
.scan<'i', int64_t>();
try {
parser.parse_args(argc, argv);
} catch (const std::exception& err) {
std::cout << err.what() << std::endl;
std::cout << parser;
std::exit(1);
}
// The number of epochs to train, default value is 30.
const int64_t kNumberOfEpochs = parser.get<int64_t>("--epochs");
std::cout << "Traning with number of epochs: " << kNumberOfEpochs
<< std::endl;
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
torch::manual_seed(1);
// Create the device we pass around based on whether CUDA is available.
torch::Device device(torch::kCPU);
if (torch::cuda::is_available()) {
std::cout << "CUDA is available! Training on GPU." << std::endl;
device = torch::Device(torch::kCUDA);
}
Update C++ frontend examples for v1.4.0 (#697) * Update C++ DCGAN example for v1.4.0 * fix Reduction::Sum * FeatureDropout -> Dropout2d
2020-01-16 18:21:52 -05:00
DCGANGenerator generator(kNoiseSize);
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
generator->to(device);
Improve code readability and make number of epochs a command line argument (#1222) * Change the cpp/dcgan * Use an open source argparse implementation
2024-01-24 21:20:29 -08:00
nn::Sequential discriminator = create_discriminator();
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
discriminator->to(device);
// Assume the MNIST dataset is available under `kDataFolder`;
auto dataset = torch::data::datasets::MNIST(kDataFolder)
.map(torch::data::transforms::Normalize<>(0.5, 0.5))
.map(torch::data::transforms::Stack<>());
Improve code readability and make number of epochs a command line argument (#1222) * Change the cpp/dcgan * Use an open source argparse implementation
2024-01-24 21:20:29 -08:00
const int64_t batches_per_epoch = static_cast<int64_t>(
std::ceil(dataset.size().value() / static_cast<double>(kBatchSize)));
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
auto data_loader = torch::data::make_data_loader(
std::move(dataset),
torch::data::DataLoaderOptions().batch_size(kBatchSize).workers(2));
torch::optim::Adam generator_optimizer(
Update dcgan.cpp (#765) Fix AdamOptions initialization error with latest libtorch. Error without this change: ``` dcgan.cpp:124:64: error: no member named 'beta1' in 'torch::optim::AdamOptions' generator->parameters(), torch::optim::AdamOptions(2e-4).beta1(0.5)); ```
2020-05-04 02:39:59 +09:00
generator->parameters(), torch::optim::AdamOptions(2e-4).betas(std::make_tuple (0.5, 0.5)));
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
torch::optim::Adam discriminator_optimizer(
Update dcgan.cpp (#765) Fix AdamOptions initialization error with latest libtorch. Error without this change: ``` dcgan.cpp:124:64: error: no member named 'beta1' in 'torch::optim::AdamOptions' generator->parameters(), torch::optim::AdamOptions(2e-4).beta1(0.5)); ```
2020-05-04 02:39:59 +09:00
discriminator->parameters(), torch::optim::AdamOptions(2e-4).betas(std::make_tuple (0.5, 0.5)));
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
if (kRestoreFromCheckpoint) {
torch::load(generator, "generator-checkpoint.pt");
torch::load(generator_optimizer, "generator-optimizer-checkpoint.pt");
torch::load(discriminator, "discriminator-checkpoint.pt");
torch::load(
discriminator_optimizer, "discriminator-optimizer-checkpoint.pt");
}
int64_t checkpoint_counter = 1;
for (int64_t epoch = 1; epoch <= kNumberOfEpochs; ++epoch) {
int64_t batch_index = 0;
Improve code readability and make number of epochs a command line argument (#1222) * Change the cpp/dcgan * Use an open source argparse implementation
2024-01-24 21:20:29 -08:00
for (const torch::data::Example<>& batch : *data_loader) {
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
// Train discriminator with real images.
discriminator->zero_grad();
torch::Tensor real_images = batch.data.to(device);
torch::Tensor real_labels =
torch::empty(batch.data.size(0), device).uniform_(0.8, 1.0);
Fix the DCGAN C++ shape warning (#1207) fix the dcgan shape warning
2024-01-11 20:38:03 -08:00
torch::Tensor real_output = discriminator->forward(real_images).reshape(real_labels.sizes());
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
torch::Tensor d_loss_real =
torch::binary_cross_entropy(real_output, real_labels);
d_loss_real.backward();
// Train discriminator with fake images.
torch::Tensor noise =
torch::randn({batch.data.size(0), kNoiseSize, 1, 1}, device);
torch::Tensor fake_images = generator->forward(noise);
torch::Tensor fake_labels = torch::zeros(batch.data.size(0), device);
Fix the DCGAN C++ shape warning (#1207) fix the dcgan shape warning
2024-01-11 20:38:03 -08:00
torch::Tensor fake_output = discriminator->forward(fake_images.detach()).reshape(fake_labels.sizes());
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
torch::Tensor d_loss_fake =
torch::binary_cross_entropy(fake_output, fake_labels);
d_loss_fake.backward();
torch::Tensor d_loss = d_loss_real + d_loss_fake;
discriminator_optimizer.step();
// Train generator.
generator->zero_grad();
fake_labels.fill_(1);
Fix the DCGAN C++ shape warning (#1207) fix the dcgan shape warning
2024-01-11 20:38:03 -08:00
fake_output = discriminator->forward(fake_images).reshape(fake_labels.sizes());
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
torch::Tensor g_loss =
torch::binary_cross_entropy(fake_output, fake_labels);
g_loss.backward();
generator_optimizer.step();
[C++][DCGAN] Increment batch_index before the if statements? (#523) * [C++][DCGAN] Increment batch_index before the if statements? I ran the example codes for [C++][DCGAN] but the 2 if statements in the nested for loop are not behaving as expected. The first one just prints the losses every epoch essentially since batch_index gets value 1 through the life of the program. As for the second one it seems to not save anything for the same reason as the first. Not sure if this code runs perfectly fine on other versions as I ran it on C++14. If I'm in the wrong then sorry I wasted your time with this PR. Thanks * Update dcgan.cpp Fixed the missing semi-colon after batch_index++. Ran it and can confirm that it runs (on C++14), and outputs are similar to what is shown on the README.MD of the original post here https://github.com/pytorch/examples/blob/master/cpp/dcgan/README.md . Example output : CUDA is available! Training on GPU. [ 1/30][200/938] D_loss: 0.4970 | G_loss: 3.6134 -> checkpoint 2 [ 1/30][400/938] D_loss: 0.3550 | G_loss: 4.8183 -> checkpoint 3 [ 1/30][600/938] D_loss: 0.5576 | G_loss: 4.4176 -> checkpoint 4 [ 1/30][800/938] D_loss: 0.3619 | G_loss: 4.5095 -> checkpoint 5
2019-03-10 15:11:37 -07:00
batch_index++;
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
if (batch_index % kLogInterval == 0) {
std::printf(
fix log cannot be printed right (#541)
2019-04-10 23:02:53 +08:00
"\r[%2ld/%2ld][%3ld/%3ld] D_loss: %.4f | G_loss: %.4f\n",
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
epoch,
kNumberOfEpochs,
[C++][DCGAN] Increment batch_index before the if statements? (#523) * [C++][DCGAN] Increment batch_index before the if statements? I ran the example codes for [C++][DCGAN] but the 2 if statements in the nested for loop are not behaving as expected. The first one just prints the losses every epoch essentially since batch_index gets value 1 through the life of the program. As for the second one it seems to not save anything for the same reason as the first. Not sure if this code runs perfectly fine on other versions as I ran it on C++14. If I'm in the wrong then sorry I wasted your time with this PR. Thanks * Update dcgan.cpp Fixed the missing semi-colon after batch_index++. Ran it and can confirm that it runs (on C++14), and outputs are similar to what is shown on the README.MD of the original post here https://github.com/pytorch/examples/blob/master/cpp/dcgan/README.md . Example output : CUDA is available! Training on GPU. [ 1/30][200/938] D_loss: 0.4970 | G_loss: 3.6134 -> checkpoint 2 [ 1/30][400/938] D_loss: 0.3550 | G_loss: 4.8183 -> checkpoint 3 [ 1/30][600/938] D_loss: 0.5576 | G_loss: 4.4176 -> checkpoint 4 [ 1/30][800/938] D_loss: 0.3619 | G_loss: 4.5095 -> checkpoint 5
2019-03-10 15:11:37 -07:00
batch_index,
Add cpp folder for C++ frontend examples (#492) * Create C++ version of MNIST example * Create C++ version of DCGAN example * Update for Normalize transform
2019-01-15 07:22:30 -08:00
batches_per_epoch,
d_loss.item<float>(),
g_loss.item<float>());
}
if (batch_index % kCheckpointEvery == 0) {
// Checkpoint the model and optimizer state.
torch::save(generator, "generator-checkpoint.pt");
torch::save(generator_optimizer, "generator-optimizer-checkpoint.pt");
torch::save(discriminator, "discriminator-checkpoint.pt");
torch::save(
discriminator_optimizer, "discriminator-optimizer-checkpoint.pt");
// Sample the generator and save the images.
torch::Tensor samples = generator->forward(torch::randn(
{kNumberOfSamplesPerCheckpoint, kNoiseSize, 1, 1}, device));
torch::save(
(samples + 1.0) / 2.0,
torch::str("dcgan-sample-", checkpoint_counter, ".pt"));
std::cout << "\n-> checkpoint " << ++checkpoint_counter << '\n';
}
}
}
std::cout << "Training complete!" << std::endl;
}
Reference in New Issue Copy Permalink