def main(dataset="babl"):
if dataset == "nlvr":
data = prepare.load_data(train_json)
data = prepare.tokenize_data(data, mxlen)
imgs, ws, labels = prepare.load_images(train_img_folder,
data,
debug=True)
data.clear()
imgs_mean = np.mean(imgs)
imgs_std = np.std(imgs - imgs_mean)
imgs = (imgs - imgs_mean) / imgs_std
else:
inputs_train, queries_train, answers_train, inputs_test, queries_test, answers_test = prepare.get_babl_data(
)
epochs = 100
model = build_model.model()
model.fit([inputs_train, queries_train],
answers_train,
validation_split=0.1,
epochs=epochs)
model.save('model')
from keras.optimizers import Adam
import gc
import prepare
import subprocess
mxlen = 32
embedding_dim = 50
lstm_unit = 128
MLP_unit = 128
epochs = 100
train_json = 'nlvr\\train\\train.json'
train_img_folder = 'nlvr\\train\\images'
data = prepare.load_data(train_json)
data = prepare.tokenize_data(data, mxlen)
imgs, ws, labels = prepare.load_images(train_img_folder, data, debug=True)
data.clear()
imgs_mean = np.mean(imgs)
imgs_std = np.std(imgs - imgs_mean)
imgs = (imgs - imgs_mean) / imgs_std
epochs = 100
batch_size = 64
def bn_layer(x, conv_unit):
def f(inputs):
md = Conv2D(x, (conv_unit, conv_unit), padding='same')(inputs)
md = BatchNormalization()(md)
return Activation('relu')(md)
return f
mxlen, seed)
EPOCHS = 60
BATCH_SIZE = 256
train_json = 'data/nlvr/train/train.json'
train_img_folder = 'data/nlvr/train/images'
dev_json = 'data/nlvr/dev/dev.json'
dev_img_folder = 'data/nlvr/dev/images'
print('Processing input data...')
data = prepare.load_data(train_json)
prepare.init_tokenizer(data)
data = prepare.tokenize_data(data, mxlen)
imgs, ws, labels = prepare.load_images(train_img_folder, data, section=True)
data.clear()
dev_data = prepare.load_data(dev_json)
dev_data = prepare.tokenize_data(dev_data, mxlen)
dev_imgs, dev_ws, dev_labels = prepare.load_images(dev_img_folder, dev_data, section=True)
dev_data.clear()
imgs_mean = np.mean(imgs)
imgs_std = np.std(imgs - imgs_mean)
imgs = (imgs - imgs_mean) / (imgs_std + 1e-7)
dev_imgs = (dev_imgs - imgs_mean) / (imgs_std + 1e-7)
imgs_1, imgs_2, imgs_3 = imgs[:,0,:,:], imgs[:,1,:,:], imgs[:,2,:,:]
dev_imgs_1, dev_imgs_2, dev_imgs_3 = dev_imgs[:,0,:,:], dev_imgs[:,1,:,:], dev_imgs[:,2,:,:]
mxlen = 32 embedding_dim = 50 lstm_unit = 64 MLP_unit = 128 epochs = 200 batch_size = 256 l2_norm = 1e-4 train_json = 'nlvr\\train\\train.json' train_img_folder = 'nlvr\\train\\images' test_json = 'nlvr\\dev\\dev.json' test_img_folder = 'nlvr\\dev\\images' data = prepare.load_data(train_json) prepare.init_tokenizer(data) data = prepare.tokenize_data(data, mxlen) imgs, ws, labels = prepare.load_images(train_img_folder, data) data.clear() test_data = prepare.load_data(test_json) test_data = prepare.tokenize_data(test_data, mxlen) test_imgs, test_ws, test_labels = prepare.load_images(test_img_folder, test_data) test_data.clear() imgs_mean = np.mean(imgs) imgs_std = np.std(imgs - imgs_mean) imgs = (imgs - imgs_mean) / imgs_std test_imgs = (test_imgs - imgs_mean) / imgs_std def bn_layer(x, conv_unit):
import numpy as np
import matplotlib.pyplot as plt
import prepare, align, classify, present
CLSNAMES = ['phnat', 'bualex']
CLSCOUNT = len(CLSNAMES)
# compose a video path for each classname
VIDPATHS = ['./data/vid' + n + '.mp4' for n in range(CLSCOUNT)]
images = [prepare.load_images(VIDPATHS[n]) for n in range(CLSCOUNT)]
X = []
y = []
for i in range(CLSCOUNT):
for img in images[i]:
X.append(img)
y.append(i)
X = np.asarray(X)
print "composed X matrix. Shape:", X.shape, "y shape:", y.shape
X_f = align.frontalize(X)
print "perormed frontalzation."
present.two_img(X[1], X_f[1])
net = classify.DeepRtlNet()