def readEntertainDB(filename):
try:
with open(filename, mode='rt', encoding='UTF8') as file:
datareader = csv.reader(file)
# row_count = sum(1 for row in datareader)
# print('row count of file: ' + filename + ' is ' + str(row_count))
row = next(datareader) # yield the header row
for row in datareader:
imdb_id = ''
type = ''
title = ''
genre = ''
if (len(row) > 0):
imdb_id = row[0]
if (len(row) > 1):
type = row[1]
if (len(row) > 2):
title = row[2]
if (len(row) > 3):
genre = row[3]
if (not imdb_id in imdb_data): # imdb id already exist
imdb_item = imdb.Imdb()
imdb_item.id = imdb_id
imdb_item.type = type
imdb_item.title = title
imdb_item.genre = genre
imdb_data[imdb_id] = imdb_item
except Exception as e:
print('exception while reading entertainment db : ' + str(e))
def train():
paddle.init(use_gpu=False, trainer_count=1)
data = paddle.layer.data(
name='data', type=paddle.data_type.integer_value_sequence(102100))
label = paddle.layer.data(name='label',
type=paddle.data_type.integer_value(2))
output = model(data)
loss = paddle.layer.classification_cost(input=output, label=label)
parameters = paddle.parameters.create(loss)
print(parameters.keys())
optimizer = paddle.optimizer.Adam(learning_rate=0.001)
trainer = paddle.trainer.SGD(parameters=parameters,
update_equation=optimizer,
cost=loss)
path = 'data/imdb.pkl'
dataset = imdb.Imdb(path)
train_data_reader = dataset.create_reader('train')
test_data_reader = dataset.create_reader('test')
feeding = {'data': 0, 'label': 1}
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 5 == 0:
class_error_rate = event.metrics[
'classification_error_evaluator']
print("\npass %d, Batch: %d cost: %f error: %s" %
(event.pass_id, event.batch_id, event.cost,
class_error_rate))
else:
sys.stdout.write('.')
sys.stdout.flush()
if isinstance(event, paddle.event.EndPass):
# save parameters
with gzip.open('output/params_pass_%d.tar.gz' % event.pass_id,
'w') as f:
parameters.to_tar(f)
result = trainer.test(reader=paddle.batch(test_data_reader,
batch_size=32),
feeding=feeding)
class_error_rate = result.metrics['classification_error_evaluator']
print("\nTest with Pass %d, cost: %s error: %f" %
(event.pass_id, result.cost, class_error_rate))
trainer.train(reader=paddle.batch(train_data_reader, batch_size=32),
event_handler=event_handler,
num_passes=10,
feeding=feeding)
# http://www.gnu.org/copyleft/gpl.html
#
import os
import random
import datetime
import re
import imdb
import game
import library
import xbmcvfs
from strings import *
IMDB = imdb.Imdb()
class Answer(object):
def __init__(self, id, text, image=None, sortWeight=None, correct=False):
self.correct = correct
self.id = id
self.text = text
self.coverFile = image
self.sortWeight = sortWeight
def setCoverFile(self, coverFile):
self.coverFile = coverFile
def __repr__(self):
return "<Answer(id=%s, text=%s, correct=%s)>" % (self.id, self.text, self.correct)