Exemplos de ResNetMCC em Python

Exemplo n.º 1

from loss_functions.multi_angular_loss import multi_angular_loss
from torchvision import transforms
import torch
import torchvision
import os
import time
from thop import profile

dataset = ImageNet(
    train=False,
    transform=transforms.Compose([transforms.ToTensor()]),
    target_transform=transforms.Compose([transforms.ToTensor()]),
)
testloader = DataLoader(dataset, batch_size=1, shuffle=False, num_workers=8)

model = ResNetMCC()
model.to(device=DEVICE)

macs, params = profile(model,
                       inputs=(torch.randn(1, 3, 224, 224).to(DEVICE), ))
print("Model's macs is %f, params is %f" % (macs, params))


def run():
    statistical_angular_errors = StatisticalValue()
    sub_dir = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime())
    print('Test start.')

    with torch.no_grad():
        for idx, (images, labels, names) in enumerate(testloader):
            images, labels = images.to(DEVICE), labels.to(DEVICE)

Exemplo n.º 2

Arquivo: train.py Projeto: YiKeYaTu/Experiments

from os.path import join
import torch
import torchvision
import numpy as np
import os

trainloader = DataLoader(ImageNet(
    train=True,
    transform=transforms.Compose([transforms.ToTensor()]),
    target_transform=transforms.Compose([transforms.ToTensor()]),
),
                         batch_size=10,
                         shuffle=False,
                         num_workers=8)

model = ResNetMCC()
model.to(device=DEVICE)

criterion = torch.nn.MSELoss(reduction='sum')
# criterion = multi_angular_loss
optimizer = torch.optim.Adam(model.parameters(),
                             lr=LEARNING_RATE,
                             weight_decay=WEIGHT_DECAY)
# optimizer = torch.optim.SGD(
#     model.parameters(),
#     momentum=0.9,
#     lr=LEARNING_RATE,
#     weight_decay=WEIGHT_DECAY
# )
scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=1, gamma=0.1)

Exemplo n.º 3

Arquivo: eval_performance.py Projeto: YiKeYaTu/Experiments

from constant import DEVICE, TMP_ROOT
from utils.StatisticalValue import StatisticalValue
from loss_functions.multi_angular_loss import multi_angular_loss
from torchvision import transforms
import torch
import torchvision
import os
import time
from thop import profile

dataset = ImageNet(
    train=False,
    transform=transforms.Compose([
        transforms.ToTensor()
    ]),
    target_transform=transforms.Compose([
        transforms.ToTensor()
    ]),
)
testloader = DataLoader(
    dataset,
    batch_size=1,
    shuffle=False,
    num_workers=8
)

model = ResNetMCC(resnet50=True)
model.to(device=DEVICE)

macs, params = profile(model, inputs=(torch.randn(1, 3, 224, 224).to(DEVICE), ))
print("Model's macs is %f, params is %f" % (macs, params))

Exemplo n.º 4

Arquivo: test.py Projeto: YiKeYaTu/Experiments

from loss_functions.multi_angular_loss import multi_angular_loss
from torchvision import transforms
import torch
import torchvision
import os
import time
from thop import profile

dataset = ImageNet(
    train=False,
    transform=transforms.Compose([transforms.ToTensor()]),
    target_transform=transforms.Compose([transforms.ToTensor()]),
)
testloader = DataLoader(dataset, batch_size=1, shuffle=False, num_workers=8)

model = ResNetMCC(layer_count=152)
model.to(device=DEVICE)

# macs, params = profile(model, inputs=(torch.randn(1, 3, 224, 224).to(DEVICE), ))
# print("Model's macs is %f, params is %f" % (macs, params))


def run():
    statistical_angular_errors = StatisticalValue()
    sub_dir = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime())
    print('Test start.')

    with torch.no_grad():
        for idx, (images, labels, names) in enumerate(testloader):
            images, labels = images.to(DEVICE), labels.to(DEVICE)

Exemplo n.º 5

Arquivo: train.py Projeto: YiKeYaTu/Experiments

trainloader = DataLoader(
    ImageNet(
        train=True,
        transform=transforms.Compose([
            transforms.ToTensor()
        ]),
        target_transform=transforms.Compose([
            transforms.ToTensor()
        ]),
    ),
    batch_size=10,
    shuffle=False,
    num_workers=8
)

model = ResNetMCC(layer_count=152)
model.to(device=DEVICE)

criterion = torch.nn.MSELoss(reduction='sum')
# criterion = multi_angular_loss
optimizer = torch.optim.Adam(
    model.parameters(),
    lr=LEARNING_RATE,
    weight_decay=WEIGHT_DECAY
)
# optimizer = torch.optim.SGD(
#     model.parameters(),
#     momentum=0.9,
#     lr=LEARNING_RATE,
#     weight_decay=WEIGHT_DECAY
# )