import torch.optim as optim
import torch.nn as nn
import torch.nn.functional as F
from datasets import mnist, loader
from logger import struct, Logger
from train import train, acc_metric, nll_metric
import numpy as np
H_proto = struct(
batch_size=64,
val_batch_size=1000,
epochs=10,
lr=0.01,
caps=[10, 20, 50, 70, 100, 200, 500],
momentum=0.5,
)
S_proto = struct(epoch=0, bn=0)
log = Logger('mnist_capacity_nodrop',
H_proto,
S_proto,
load=True,
metric_show_freq=1)
tn_loader, val_loader = loader(mnist(), H_proto.batch_size,
H_proto.val_batch_size)
for epoch in range(S_proto.epoch, len(H_proto.caps)):
S_proto.epoch = epoch
log.flush()
cap = H_proto.caps[epoch]
from logger import Logger, struct
import matplotlib.pyplot as plt
# from pprint import pprint
H = struct()
S = struct()
log = Logger('mnist_momentum', H, S, load=True)
losses = [m.best_loss for m in log.metrics()]
# pprint(log.metrics())
# pprint(H.momenta)
plt.plot(H.momenta, losses)
plt.xscale('log')
plt.yscale('log')
plt.show()
import torch.optim as optim
import torch.nn as nn
import torch.nn.functional as F
from datasets import cifar10, loader
from logger import struct, Logger
from train import train, acc_metric, nll_metric
from models import vgg
H = struct(
batch_size=128,
val_batch_size=100,
epochs=10,
lr=0.1,
momentum=0.9,
)
S = struct(epoch=1, bn=1)
log = Logger('cifar10_standard', H, S, overwrite=True, metric_show_freq=100)
tn_loader, val_loader = loader(cifar10(), H.batch_size, H.val_batch_size)
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.features = vgg('vgg11')
self.classifier = nn.Linear(512, 10)
def forward(self, x):
out = self.features(x)
out = out.view(out.size(0), -1)
from logger import Logger, struct
import matplotlib.pyplot as plt
from pprint import pprint
H = struct()
S = struct()
log = Logger('mnist_capacity', H, S, load=True)
losses = [m.best_loss for m in log.metrics()]
H_nodrop = struct()
S_nodrop = struct()
log_nodrop = Logger('mnist_capacity_nodrop', H_nodrop, S_nodrop, load=True)
losses_nodrop = [m.best_loss for m in log_nodrop.metrics()]
H_mlp = struct()
S_mlp = struct()
log_mlp = Logger('mnist_capacity_mlp', H_mlp, S_mlp, load=True)
losses_mlp = [m.best_loss for m in log_mlp.metrics()]
pprint(log.metrics())
pprint(log_nodrop.metrics())
# pprint(H.caps)
plt.plot(H.caps, losses, label='dropout')
plt.plot(H_nodrop.caps, losses_nodrop, label='nodrop')
plt.plot(H_mlp.caps, losses_mlp, label='mlp')
plt.xscale('log')
plt.yscale('log')
plt.legend(loc='best')
plt.show()
import torch.optim as optim
import torch.nn as nn
import torch.nn.functional as F
import numpy as np
from datasets import mnist, loader
from logger import struct, Logger
from train import train, acc_metric, nll_metric
H_proto = struct(
batch_size=64,
val_batch_size=1000,
epochs=10,
lrs=[
0.001, 0.003, 0.005, 0.01, 0.02, 0.03, 0.04, 0.05, 0.07, 0.1, 0.3, 0.5,
1
],
momentum=0.5,
)
S_proto = struct(epoch=0, bn=0)
log = Logger('mnist_lr2', H_proto, S_proto, load=True, metric_show_freq=1)
S_proto.epoch = 0
tn_loader, val_loader = loader(mnist(), H_proto.batch_size,
H_proto.val_batch_size)
for epoch in range(S_proto.epoch, len(H_proto.lrs)):
S_proto.epoch = epoch
log.flush()
lr = H_proto.lrs[epoch]
H = H_proto.copy()
import torch.optim as optim
import torch.nn as nn
import torch.nn.functional as F
from datasets import mnist, loader
from logger import struct, Logger
from train import train
H = struct(
batch_size=64,
val_batch_size=1000,
epochs=2,
lr=0.01,
momentum=0.5,
)
S = struct(epoch=1, bn=1)
log = Logger('mnist_scratch', H, S, overwrite=True, metric_show_freq=100)
tn_loader, val_loader = loader(mnist(), H.batch_size, H.val_batch_size)
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, 50)
self.fc2 = nn.Linear(50, 10)
def forward(self, x):
import torch.optim as optim
import torch.nn as nn
import torch.nn.functional as F
from datasets import mnist, loader
from logger import struct
from train import (cycle_hyperparams, acc_metric, nll_metric, max_metametric,
argmax_metametric, min_metametric, argmin_metametric)
H_proto = struct(batch_size=64,
val_batch_size=1000,
epochs=10,
lr=0.01,
momenta=[0.0, 0.1, 0.3, 0.5, 0.7, 0.9, 0.95, 0.99])
def model(H):
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, 50)
self.fc2 = nn.Linear(50, 10)
def forward(self, x):
x = F.relu(F.max_pool2d(self.conv1(x), 2))
x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
x = x.view(-1, 320)
x = F.relu(self.fc1(x))
x = F.dropout(x, training=self.training)
import torch.optim as optim
import torch.nn as nn
import torch.nn.functional as F
from datasets import mnist, loader
from logger import struct
from train import (cycle_hyperparams, acc_metric, nll_metric, max_metametric,
argmax_metametric, min_metametric, argmin_metametric)
H_proto = struct(batch_size=64,
val_batch_size=1000,
epochs=10,
lr=0.01,
caps=[10, 20, 50, 70, 100, 200, 500],
momentum=0.5)
def model(H):
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, H.cap)
self.fc2 = nn.Linear(H.cap, 10)
def forward(self, x):
x = F.relu(F.max_pool2d(self.conv1(x), 2))
x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
x = x.view(-1, 320)
x = F.relu(self.fc1(x))
import torch.optim as optim
import torch.nn as nn
import torch.nn.functional as F
from datasets import mnist, loader
from logger import struct
from train import (cycle_hyperparams, acc_metric, nll_metric, max_metametric,
argmax_metametric, min_metametric, argmin_metametric)
H_proto = struct(batch_size=64,
val_batch_size=1000,
epochs=10,
lrs=[
0.001, 0.003, 0.005, 0.01, 0.02, 0.03, 0.04, 0.05, 0.07,
0.1, 0.3, 0.5, 1
],
momentum=0.5)
def model(H):
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, 50)
self.fc2 = nn.Linear(50, 10)
def forward(self, x):
x = F.relu(F.max_pool2d(self.conv1(x), 2))
x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))