Identificação de Raça de Cachorro (*ImageNet Dogs*) no Kaggle
=============================================================
Nesta seção, abordaremos o desafio da identificação de raças de cães na
Competição Kaggle. O endereço da competição na web é
https://www.kaggle.com/c/dog-breed-identification
Nesta competição, tentamos identificar 120 raças diferentes de cães. O
conjunto de dados usado nesta competição é, na verdade, um subconjunto
do famoso conjunto de dados ImageNet. Diferente das imagens no conjunto
de dados CIFAR-10 usado na seção anterior, as imagens no conjunto de
dados ImageNet são mais altas e mais largas e suas dimensões são
inconsistentes.
:numref:`fig_kaggle_dog` mostra as informações na página da
competição. Para enviar os resultados, primeiro registre uma conta no
site do Kaggle.
.. _fig_kaggle_dog:
.. figure:: ../img/kaggle-dog.jpg
:width: 400px
Site de competição de identificação de raças de cães. O conjunto de
dados da competição pode ser acessado clicando na guia “Data”.
Primeiro, importe os pacotes ou módulos necessários para a competição.
.. raw:: html
.. raw:: html
.. code:: python
import os
from mxnet import autograd, gluon, init, npx
from mxnet.gluon import nn
from d2l import mxnet as d2l
npx.set_np()
.. raw:: html
.. raw:: html
.. code:: python
import os
import torch
import torchvision
from torch import nn
from d2l import torch as d2l
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
#@save
d2l.DATA_HUB['dog_tiny'] = (d2l.DATA_URL + 'kaggle_dog_tiny.zip',
'0cb91d09b814ecdc07b50f31f8dcad3e81d6a86d')
# If you use the full dataset downloaded for the Kaggle competition, change
# the variable below to False
demo = True
if demo:
data_dir = d2l.download_extract('dog_tiny')
else:
data_dir = os.path.join('..', 'data', 'dog-breed-identification')
.. parsed-literal::
:class: output
Downloading ../data/kaggle_dog_tiny.zip from http://d2l-data.s3-accelerate.amazonaws.com/kaggle_dog_tiny.zip...
.. raw:: html
.. raw:: html
.. code:: python
#@save
d2l.DATA_HUB['dog_tiny'] = (d2l.DATA_URL + 'kaggle_dog_tiny.zip',
'0cb91d09b814ecdc07b50f31f8dcad3e81d6a86d')
# If you use the full dataset downloaded for the Kaggle competition, change
# the variable below to False
demo = True
if demo:
data_dir = d2l.download_extract('dog_tiny')
else:
data_dir = os.path.join('..', 'data', 'dog-breed-identification')
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
def reorg_dog_data(data_dir, valid_ratio):
labels = d2l.read_csv_labels(os.path.join(data_dir, 'labels.csv'))
d2l.reorg_train_valid(data_dir, labels, valid_ratio)
d2l.reorg_test(data_dir)
batch_size = 4 if demo else 128
valid_ratio = 0.1
reorg_dog_data(data_dir, valid_ratio)
.. raw:: html
.. raw:: html
.. code:: python
def reorg_dog_data(data_dir, valid_ratio):
labels = d2l.read_csv_labels(os.path.join(data_dir, 'labels.csv'))
d2l.reorg_train_valid(data_dir, labels, valid_ratio)
d2l.reorg_test(data_dir)
batch_size = 4 if demo else 128
valid_ratio = 0.1
reorg_dog_data(data_dir, valid_ratio)
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
transform_train = gluon.data.vision.transforms.Compose([
# Randomly crop the image to obtain an image with an area of 0.08 to 1 of
# the original area and height to width ratio between 3/4 and 4/3. Then,
# scale the image to create a new image with a height and width of 224
# pixels each
gluon.data.vision.transforms.RandomResizedCrop(224, scale=(0.08, 1.0),
ratio=(3.0/4.0, 4.0/3.0)),
gluon.data.vision.transforms.RandomFlipLeftRight(),
# Randomly change the brightness, contrast, and saturation
gluon.data.vision.transforms.RandomColorJitter(brightness=0.4,
contrast=0.4,
saturation=0.4),
# Add random noise
gluon.data.vision.transforms.RandomLighting(0.1),
gluon.data.vision.transforms.ToTensor(),
# Standardize each channel of the image
gluon.data.vision.transforms.Normalize([0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])])
.. raw:: html
.. raw:: html
.. code:: python
transform_train = torchvision.transforms.Compose([
# Randomly crop the image to obtain an image with an area of 0.08 to 1 of
# the original area and height to width ratio between 3/4 and 4/3. Then,
# scale the image to create a new image with a height and width of 224
# pixels each
torchvision.transforms.RandomResizedCrop(224, scale=(0.08, 1.0),
ratio=(3.0/4.0, 4.0/3.0)),
torchvision.transforms.RandomHorizontalFlip(),
# Randomly change the brightness, contrast, and saturation
torchvision.transforms.ColorJitter(brightness=0.4,
contrast=0.4,
saturation=0.4),
# Add random noise
torchvision.transforms.ToTensor(),
# Standardize each channel of the image
torchvision.transforms.Normalize([0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])])
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
transform_test = gluon.data.vision.transforms.Compose([
gluon.data.vision.transforms.Resize(256),
# Crop a square of 224 by 224 from the center of the image
gluon.data.vision.transforms.CenterCrop(224),
gluon.data.vision.transforms.ToTensor(),
gluon.data.vision.transforms.Normalize([0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])])
.. raw:: html
.. raw:: html
.. code:: python
transform_test = torchvision.transforms.Compose([
torchvision.transforms.Resize(256),
# Crop a square of 224 by 224 from the center of the image
torchvision.transforms.CenterCrop(224),
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize([0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])])
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
train_ds, valid_ds, train_valid_ds, test_ds = [
gluon.data.vision.ImageFolderDataset(
os.path.join(data_dir, 'train_valid_test', folder))
for folder in ('train', 'valid', 'train_valid', 'test')]
.. raw:: html
.. raw:: html
.. code:: python
train_ds, train_valid_ds = [torchvision.datasets.ImageFolder(
os.path.join(data_dir, 'train_valid_test', folder),
transform=transform_train) for folder in ['train', 'train_valid']]
valid_ds, test_ds = [torchvision.datasets.ImageFolder(
os.path.join(data_dir, 'train_valid_test', folder),
transform=transform_test) for folder in ['valid', 'test']]
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
train_iter, train_valid_iter = [gluon.data.DataLoader(
dataset.transform_first(transform_train), batch_size, shuffle=True,
last_batch='discard') for dataset in (train_ds, train_valid_ds)]
valid_iter = gluon.data.DataLoader(
valid_ds.transform_first(transform_test), batch_size, shuffle=False,
last_batch='discard')
test_iter = gluon.data.DataLoader(
test_ds.transform_first(transform_test), batch_size, shuffle=False,
last_batch='keep')
.. raw:: html
.. raw:: html
.. code:: python
train_iter, train_valid_iter = [torch.utils.data.DataLoader(
dataset, batch_size, shuffle=True, drop_last=True)
for dataset in (train_ds, train_valid_ds)]
valid_iter = torch.utils.data.DataLoader(valid_ds, batch_size, shuffle=False,
drop_last=True)
test_iter = torch.utils.data.DataLoader(test_ds, batch_size, shuffle=False,
drop_last=False)
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
def get_net(devices):
finetune_net = gluon.model_zoo.vision.resnet34_v2(pretrained=True)
# Define a new output network
finetune_net.output_new = nn.HybridSequential(prefix='')
finetune_net.output_new.add(nn.Dense(256, activation='relu'))
# There are 120 output categories
finetune_net.output_new.add(nn.Dense(120))
# Initialize the output network
finetune_net.output_new.initialize(init.Xavier(), ctx=devices)
# Distribute the model parameters to the CPUs or GPUs used for computation
finetune_net.collect_params().reset_ctx(devices)
return finetune_net
.. raw:: html
.. raw:: html
.. code:: python
def get_net(devices):
finetune_net = nn.Sequential()
finetune_net.features = torchvision.models.resnet34(pretrained=True)
# Define a new output network
# There are 120 output categories
finetune_net.output_new = nn.Sequential(nn.Linear(1000, 256),
nn.ReLU(),
nn.Linear(256, 120))
# Move model to device
finetune_net = finetune_net.to(devices[0])
# Freeze feature layer params
for param in finetune_net.features.parameters():
param.requires_grad = False
return finetune_net
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
loss = gluon.loss.SoftmaxCrossEntropyLoss()
def evaluate_loss(data_iter, net, devices):
l_sum, n = 0.0, 0
for features, labels in data_iter:
X_shards, y_shards = d2l.split_batch(features, labels, devices)
output_features = [net.features(X_shard) for X_shard in X_shards]
outputs = [net.output_new(feature) for feature in output_features]
ls = [loss(output, y_shard).sum() for output, y_shard
in zip(outputs, y_shards)]
l_sum += sum([float(l.sum()) for l in ls])
n += labels.size
return l_sum / n
.. raw:: html
.. raw:: html
.. code:: python
loss = nn.CrossEntropyLoss(reduction='none')
def evaluate_loss(data_iter, net, devices):
l_sum, n = 0.0, 0
for features, labels in data_iter:
features, labels = features.to(devices[0]), labels.to(devices[0])
outputs = net(features)
l = loss(outputs, labels)
l_sum = l.sum()
n += labels.numel()
return l_sum / n
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
def train(net, train_iter, valid_iter, num_epochs, lr, wd, devices, lr_period,
lr_decay):
# Only train the small custom output network
trainer = gluon.Trainer(net.output_new.collect_params(), 'sgd',
{'learning_rate': lr, 'momentum': 0.9, 'wd': wd})
num_batches, timer = len(train_iter), d2l.Timer()
animator = d2l.Animator(xlabel='epoch', xlim=[1, num_epochs],
legend=['train loss', 'valid loss'])
for epoch in range(num_epochs):
metric = d2l.Accumulator(2)
if epoch > 0 and epoch % lr_period == 0:
trainer.set_learning_rate(trainer.learning_rate * lr_decay)
for i, (features, labels) in enumerate(train_iter):
timer.start()
X_shards, y_shards = d2l.split_batch(features, labels, devices)
output_features = [net.features(X_shard) for X_shard in X_shards]
with autograd.record():
outputs = [net.output_new(feature)
for feature in output_features]
ls = [loss(output, y_shard).sum() for output, y_shard
in zip(outputs, y_shards)]
for l in ls:
l.backward()
trainer.step(batch_size)
metric.add(sum([float(l.sum()) for l in ls]), labels.shape[0])
timer.stop()
if (i + 1) % (num_batches // 5) == 0 or i == num_batches - 1:
animator.add(epoch + (i + 1) / num_batches,
(metric[0] / metric[1], None))
if valid_iter is not None:
valid_loss = evaluate_loss(valid_iter, net, devices)
animator.add(epoch + 1, (None, valid_loss))
if valid_iter is not None:
print(f'train loss {metric[0] / metric[1]:.3f}, '
f'valid loss {valid_loss:.3f}')
else:
print(f'train loss {metric[0] / metric[1]:.3f}')
print(f'{metric[1] * num_epochs / timer.sum():.1f} examples/sec '
f'on {str(devices)}')
.. raw:: html
.. raw:: html
.. code:: python
def train(net, train_iter, valid_iter, num_epochs, lr, wd, devices, lr_period,
lr_decay):
# Only train the small custom output network
net = nn.DataParallel(net, device_ids=devices).to(devices[0])
trainer = torch.optim.SGD((param for param in net.parameters()
if param.requires_grad), lr=lr,
momentum=0.9, weight_decay=wd)
scheduler = torch.optim.lr_scheduler.StepLR(trainer, lr_period, lr_decay)
num_batches, timer = len(train_iter), d2l.Timer()
animator = d2l.Animator(xlabel='epoch', xlim=[1, num_epochs],
legend=['train loss', 'valid loss'])
for epoch in range(num_epochs):
metric = d2l.Accumulator(2)
for i, (features, labels) in enumerate(train_iter):
timer.start()
features, labels = features.to(devices[0]), labels.to(devices[0])
trainer.zero_grad()
output = net(features)
l = loss(output, labels).sum()
l.backward()
trainer.step()
metric.add(l, labels.shape[0])
timer.stop()
if (i + 1) % (num_batches // 5) == 0 or i == num_batches - 1:
animator.add(epoch + (i + 1) / num_batches,
(metric[0] / metric[1], None))
if valid_iter is not None:
valid_loss = evaluate_loss(valid_iter, net, devices)
animator.add(epoch + 1, (None, valid_loss))
scheduler.step()
if valid_iter is not None:
print(f'train loss {metric[0] / metric[1]:.3f}, '
f'valid loss {valid_loss:.3f}')
else:
print(f'train loss {metric[0] / metric[1]:.3f}')
print(f'{metric[1] * num_epochs / timer.sum():.1f} examples/sec '
f'on {str(devices)}')
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
devices, num_epochs, lr, wd = d2l.try_all_gpus(), 5, 0.01, 1e-4
lr_period, lr_decay, net = 10, 0.1, get_net(devices)
net.hybridize()
train(net, train_iter, valid_iter, num_epochs, lr, wd, devices, lr_period,
lr_decay)
.. parsed-literal::
:class: output
train loss 2.366, valid loss 2.387
61.3 examples/sec on [gpu(0), gpu(1)]
.. figure:: output_kaggle-dog_571091_93_1.svg
.. raw:: html
.. raw:: html
.. code:: python
devices, num_epochs, lr, wd = d2l.try_all_gpus(), 5, 0.001, 1e-4
lr_period, lr_decay, net = 10, 0.1, get_net(devices)
train(net, train_iter, valid_iter, num_epochs, lr, wd, devices, lr_period,
lr_decay)
.. parsed-literal::
:class: output
train loss 3.740, valid loss 0.125
107.6 examples/sec on [device(type='cuda', index=0), device(type='cuda', index=1)]
.. figure:: output_kaggle-dog_571091_96_1.svg
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
net = get_net(devices)
net.hybridize()
train(net, train_valid_iter, None, num_epochs, lr, wd, devices, lr_period,
lr_decay)
preds = []
for data, label in test_iter:
output_features = net.features(data.as_in_ctx(devices[0]))
output = npx.softmax(net.output_new(output_features))
preds.extend(output.asnumpy())
ids = sorted(os.listdir(
os.path.join(data_dir, 'train_valid_test', 'test', 'unknown')))
with open('submission.csv', 'w') as f:
f.write('id,' + ','.join(train_valid_ds.synsets) + '\n')
for i, output in zip(ids, preds):
f.write(i.split('.')[0] + ',' + ','.join(
[str(num) for num in output]) + '\n')
.. parsed-literal::
:class: output
train loss 2.510
119.1 examples/sec on [gpu(0), gpu(1)]
.. figure:: output_kaggle-dog_571091_102_1.svg
.. raw:: html
.. raw:: html
.. code:: python
net = get_net(devices)
train(net, train_valid_iter, None, num_epochs, lr, wd, devices, lr_period,
lr_decay)
preds = []
for data, label in test_iter:
output = torch.nn.functional.softmax(net(data.to(devices[0])), dim=0)
preds.extend(output.cpu().detach().numpy())
ids = sorted(os.listdir(
os.path.join(data_dir, 'train_valid_test', 'test', 'unknown')))
with open('submission.csv', 'w') as f:
f.write('id,' + ','.join(train_valid_ds.classes) + '\n')
for i, output in zip(ids, preds):
f.write(i.split('.')[0] + ',' + ','.join(
[str(num) for num in output]) + '\n')
.. parsed-literal::
:class: output
train loss 3.779
126.4 examples/sec on [device(type='cuda', index=0), device(type='cuda', index=1)]
.. figure:: output_kaggle-dog_571091_105_1.svg
.. raw:: html
.. raw:: html
.. raw:: html
`Discussões `__
.. raw:: html
.. raw:: html
`Discussões `__
.. raw:: html
.. raw:: html