conv_layer.params[0].set_value(savedparams[k])
conv_layer.params[1].set_value(savedparams[k + 1])
k = k + 2
test_set_x = datasets[0][:, :img_h]
test_set_y = np.asarray(datasets[0][:, -1], "int32")
test_pred_layers = []
test_size = 1
test_layer0_input = Words[T.cast(x.flatten(), dtype="int32")].reshape(
(test_size, 1, img_h, Words.shape[1]))
for conv_layer in conv_layers:
test_layer0_output = conv_layer.predict(test_layer0_input, test_size)
test_pred_layers.append(test_layer0_output.flatten(2))
test_layer1_input = T.concatenate(test_pred_layers, 1)
test_y_pred = classifier.predict_p(test_layer1_input)
#test_error = T.mean(T.neq(test_y_pred, y))
pdb.set_trace()
model = theano.function([x], test_y_pred, allow_input_downcast=True)
currentClass = 0
currentBestScore = 0
currentBestSentence = ""
for i in range(0, test_set_x.shape[0]):
rev = revsin[i]
result = model(test_set_x[i, None])
scoresum = sum(result[0][0:currentClass + 1])
if rev['y'] > currentClass:
currentClass = rev['y']
currentBestScore = 0
print "best for class" + str(currentClass)
print currentBestSentence
conv_layer.params[1].set_value( savedparams[k+1])
k = k + 2
test_set_x = datasets[0][:,:img_h]
test_set_y = np.asarray(datasets[0][:,-1],"int32")
test_pred_layers = []
test_size = 1
test_layer0_input = Words[T.cast(x.flatten(),dtype="int32")].reshape((test_size,1,img_h,Words.shape[1]))
for conv_layer in conv_layers:
test_layer0_output = conv_layer.predict(test_layer0_input, test_size)
test_pred_layers.append(test_layer0_output.flatten(2))
test_layer1_input = T.concatenate(test_pred_layers, 1)
test_y_pred = classifier.predict_p(test_layer1_input)
#test_error = T.mean(T.neq(test_y_pred, y))
pdb.set_trace()
model = theano.function([x],test_y_pred,allow_input_downcast=True)
currentClass = 0
currentBestScore = 0
currentBestSentence = ""
correct = 0
docMean = 0
docSent = 0
with open('output.csv', 'w') as fp:
writer = csv.writer(fp, delimiter=',')
writer.writerow(['class','score', 'sdscore','sentence'])
print "document " + str(currentClass)
for i in range(0,test_set_x.shape[0]):
rev = revsin[i]
def __init__(self):
mrppath = os.path.join(this_dir, "mr.p")
x = cPickle.load(open(mrppath,"rb"))
revs, W, W2, word_idx_map, vocab = x[0], x[1], x[2], x[3], x[4]
self.word_idx_map = word_idx_map
U = W
classifierpath = os.path.join(this_dir, "classifier.save")
savedparams = cPickle.load(open(classifierpath,'rb'))
filter_hs=[3,4,5]
conv_non_linear="relu"
hidden_units=[100,2]
dropout_rate=[0.5]
activations=[Iden]
img_h = 56 + 4 + 4
img_w = 300
rng = np.random.RandomState(3435)
batch_size=50
filter_w = img_w
feature_maps = hidden_units[0]
filter_shapes = []
pool_sizes = []
for filter_h in filter_hs:
filter_shapes.append((feature_maps, 1, filter_h, filter_w))
pool_sizes.append((img_h-filter_h+1, img_w-filter_w+1))
#define model architecture
x = T.matrix('x')
Words = theano.shared(value = U, name = "Words")
zero_vec_tensor = T.vector()
layer0_input = Words[T.cast(x.flatten(),dtype="int32")].reshape((x.shape[0],1,x.shape[1],Words.shape[1]))
conv_layers = []
layer1_inputs = []
for i in xrange(len(filter_hs)):
filter_shape = filter_shapes[i]
pool_size = pool_sizes[i]
conv_layer = LeNetConvPoolLayer(rng, input=layer0_input,image_shape=(batch_size, 1, img_h, img_w),
filter_shape=filter_shape, poolsize=pool_size, non_linear=conv_non_linear)
layer1_input = conv_layer.output.flatten(2)
conv_layers.append(conv_layer)
layer1_inputs.append(layer1_input)
layer1_input = T.concatenate(layer1_inputs,1)
hidden_units[0] = feature_maps*len(filter_hs)
classifier = MLPDropout(rng, input=layer1_input, layer_sizes=hidden_units, activations=activations, dropout_rates=dropout_rate)
classifier.params[0].set_value(savedparams[0])
classifier.params[1].set_value(savedparams[1])
k = 2
for conv_layer in conv_layers:
conv_layer.params[0].set_value( savedparams[k])
conv_layer.params[1].set_value( savedparams[k+1])
k = k + 2
test_pred_layers = []
test_size = 1
test_layer0_input = Words[T.cast(x.flatten(),dtype="int32")].reshape((test_size,1,img_h,Words.shape[1]))
for conv_layer in conv_layers:
test_layer0_output = conv_layer.predict(test_layer0_input, test_size)
test_pred_layers.append(test_layer0_output.flatten(2))
test_layer1_input = T.concatenate(test_pred_layers, 1)
test_y_pred = classifier.predict_p(test_layer1_input)
#test_error = T.mean(T.neq(test_y_pred, y))
self.model = theano.function([x],test_y_pred,allow_input_downcast=True)
