def main():
trainX, trainY, testX, testY = load_mnist()
# print "Shapes: ", trainX.shape, trainY.shape, testX.shape, testY.shape
# print "\nDigit sample"
#print_digit(trainX[1], trainY[1])
for i in xrange(9995,10000):
print testY[i]
train_cnn.train(trainX, trainY)
def main():
trainX, trainY, testX, testY = load_mnist()
print "Shapes: ", trainX.shape, trainY.shape, testX.shape, testY.shape
print "\nDigit sample"
print_digit(trainX[1], trainY[1])
train_cnn.train(trainX, trainY)
labels = train_cnn.test(testX)
accuracy = np.mean((labels == testY)) * 100.0
print "\nCNN Test accuracy: %lf%%" % accuracy
def main():
trainX, trainY, testX, testY = load_mnist()
print("Shapes: ", trainX.shape, trainY.shape, testX.shape, testY.shape)
print("\nDigit sample")
print_digit(trainX[1], trainY[1])
#train_dense.train(trainX, trainY,testX,testY)
#labels = train_dense.test(testX)
#accuracy = np.mean((labels == testY)) * 100.0
#print ("\nDNN Test accuracy: %lf%%" % accuracy)
train_cnn.train(trainX, trainY, testX, testY)
model = multichannel_CNN.CNN(args)
# model.word_embeddings.weight.data.copy_(torch.from_numpy(pretrained_weight))
if args.multi_cnn2:
model = multi_CNN.CNN(args)
# model.word_embeddings.weight.data.copy_(torch.from_numpy(pretrained_weight))
if args.cnn_simple:
model = CNN.CNN(args)
# model.word_embeddings.weight.data.copy_(torch.from_numpy(pretrained_weight))
if args.cnn_char:
model = CNN_char.CNN(args)
# model.word_embeddings.weight.data.copy_(torch.from_numpy(pretrained_weight))
if args.gru:
model = gru.GRU(args)
if args.bi_gru:
model = bi_gru.BiGRU(args)
if args.bnlstm:
# model = LSTM_bn(cell_class=BNLSTMCell, input_size=1, hidden_size=args.hidden_size, max_length=1e8)
model = model_bnlstm.BNLSTM(args)
# train
print("Training start")
if args.train_cnn:
# if args.cnn_char:
# train_cnn.train_char(train_iter, dev_iter, test_iter, train_iter_char, dev_iter_char, test_iter_char, model,
# text_field, label_field, args)
# else:
train_cnn.train(train_iter, dev_iter, test_iter, model, text_field, label_field, args)
else:
train_lstm.train(train_iter, dev_iter, test_iter, model, text_field, label_field, args)
# print(padID)
# wv_cat = loader.vector_loader(count_words_reset)
# pretrained_weight = wv_cat
# args.pretrained_weight = pretrained_weight
# update args and print
args.embed_num = len(text_field.vocab)
args.class_num = len(label_field.vocab) - 1
args.save_dir = os.path.join(
args.save_dir,
datetime.datetime.now().strftime('%Y-%m-%d_%H-%M-%S'))
# model
if args.bilstm:
model = Bi_LSTM_random_emb.BiLSTM(args)
if args.use_cuda:
model = model.cuda()
if os.path.exists("./Test_Result.txt"):
os.remove("./Test_Result.txt")
# train
print("Training start")
if args.train_cnn:
train_cnn.train(train_iter, dev_iter, test_iter, model, text_field,
label_field, args)
else:
train_lstm.train(train_iter, dev_iter, test_iter, model, text_field,
label_field, args)
import train_cnn
from dataset_sorter import DatasetSorter
from feature_database import FeatureDatabase
from caffe_config import CaffeConfig
import sys
sys.path.insert(0, 'src/grasp_selection/')
import experiment_config as ec
if __name__ == '__main__':
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('config')
args = parser.parse_args()
config = ec.ExperimentConfig(args.config)
feature_db = feature_database.FeatureDatabase(config)
caffe_config = CaffeConfig()
dataset_sorter = feature_db.feature_dataset_sorter()
if dataset_sorter == None:
dataset_sorter = DatasetSorter(feature_db)
feature_db.save_dataset_sorter(dataset_sorter)
train_cnn.train(feature_db, caffe_config, dataset_sorter)
