# redistribute it and/or modify it under the terms of the GNU General Public
# License as published by the Free Software Foundation, version 2.
#
# This program is distributed in the hope that it will be useful, but WITHOUT
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
# details.
#
# You should have received a copy of the GNU General Public License along with
# this program; if not, write to the Free Software Foundation, Inc., 51
# Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
#
# Copyright Nils Schaetti <*****@*****.**>
# Imports
import torch.utils.data
import dataset
from echotorch.transforms import text
# Author identification dataset
pan18loader = torch.utils.data.DataLoader(dataset.AuthorIdentificationDataset(
root="./data/", download=True, transform=text.GloveVector(), problem=1),
batch_size=1,
shuffle=True)
# Get training data for this fold
for i, data in enumerate(pan18loader):
# Inputs and labels
inputs, labels = data
# end for
# This program is distributed in the hope that it will be useful, but WITHOUT
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
# details.
#
# You should have received a copy of the GNU General Public License along with
# this program; if not, write to the Free Software Foundation, Inc., 51
# Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
#
# Copyright Nils Schaetti <*****@*****.**>
# Imports
import torch.utils.data
from echotorch import datasets
from echotorch.transforms import text
# Reuters C50 dataset
reutersloader = torch.utils.data.DataLoader(datasets.SFGramDataset(
tokenizer=text.Token(),
root="../../data/sfgram/",
download=True,
transform=text.GloveVector()),
batch_size=1,
shuffle=True)
# Get training data for this fold
for i, data in enumerate(reutersloader):
# Inputs and labels
inputs, labels = data
# end for
type=int,
help="Number of features",
default=10)
parser.add_argument("--fold", type=int, help="Starting fold", default=0)
parser.add_argument("--steps", type=int, help="Steps", default=1)
parser.add_argument("--no-cuda",
action='store_true',
default=False,
help="Enables CUDA training")
args = parser.parse_args()
# Use CUDA?
args.cuda = not args.no_cuda and torch.cuda.is_available()
# Word embedding
transform = text.GloveVector(model='en_vectors_web_lg')
# Reuters C50 dataset
reutersloader = torch.utils.data.DataLoader(datasets.ReutersC50Dataset(
download=True, n_authors=15, transform=transform),
batch_size=1,
shuffle=False)
# Loss function
# loss_function = nn.NLLLoss()
loss_function = nn.CrossEntropyLoss()
# 10-CV
for k in np.arange(args.fold, 10):
# Model
# model = CNNFeatureSelector(embedding_dim=embedding_dim, n_authors=n_authors)
import torch.utils.data
import dataset
from echotorch.transforms import text
# Experience parameter
batch_size = 64
n_epoch = 1
window_size = 700
training_set_size = 10
test_set_size = 2
training_samples = training_set_size + test_set_size
stride = 100
# Style change detection dataset, training set
pan18loader_train = torch.utils.data.DataLoader(dataset.SCDSimpleDataset(
root='./data/', download=True, transform=text.GloveVector(), train=True),
batch_size=1)
# Style change detection dataset, validation set
pan18loader_valid = torch.utils.data.DataLoader(dataset.SCDSimpleDataset(
root='./data/', download=True, transform=text.GloveVector(), train=False),
batch_size=1)
# Get training data
for i, data in enumerate(pan18loader_train):
# Inputs and c
inputs, label = data
# TRAINING
# end for
for space in param_space:
# Params
reservoir_size = int(space['reservoir_size'])
w_sparsity = space['w_sparsity']
leak_rate = space['leak_rate']
input_scaling = space['input_scaling']
input_sparsity = space['input_sparsity']
spectral_radius = space['spectral_radius']
transformer = space['transformer'][0][0]
aggregation = space['aggregation'][0][0]
state_gram = space['state_gram']
dataset_start = int(space['dataset_start'])
# Choose the right transformer
if "wv" in transformer:
reutersloader.dataset.transform = text.GloveVector()
elif "pos" in transformer:
reutersloader.dataset.transform = text.PartOfSpeech()
elif "tag" in transformer:
reutersloader.dataset.transform = text.Tag()
elif "character" in transformer:
reutersloader.dataset.transform = text.Character()
elif "fw" in transformer:
reutersloader.dataset.transform = text.FunctionWord()
else:
print(u"No transformer set!")
exit()
# end if
# Dataset start
reutersloader.dataset.set_start(dataset_start)