def main(): (X_train, y_train), (X_test, y_test) = tf.contrib.keras.datasets.mnist.load_data() X_train = X_train / 255. X_test = X_test / 255. model = RNNClassifier(mx.cpu(), n_out=10) model.fit(X_train, y_train) pred = model.predict(X_test) final_acc = (pred == y_test).mean() print("final testing accuracy: %.4f" % final_acc)
def main(): (X_train, y_train), (X_test, y_test) = tf.contrib.keras.datasets.cifar10.load_data() X_train = (X_train / 255.).mean(axis=3) X_test = (X_test / 255.).mean(axis=3) y_train = y_train.ravel() y_test = y_test.ravel() model = RNNClassifier(mx.cpu(), n_out=10) model.fit(X_train, y_train) pred = model.predict(X_test) final_acc = (pred == y_test).mean() print("final testing accuracy: %.4f" % final_acc)
from rnn_clf import RNNClassifier import numpy as np import tensorflow as tf if __name__ == '__main__': (X_train, y_train), (X_test, y_test) = tf.contrib.keras.datasets.mnist.load_data() X_train = X_train / 255.0 X_test = X_test / 255.0 Y_train = tf.contrib.keras.utils.to_categorical(y_train) Y_test = tf.contrib.keras.utils.to_categorical(y_test) clf = RNNClassifier(n_in=28, n_out=10, stateful=True) log = clf.fit(X_train, y_train, keep_prob_tuple=(0.8,1.0), val_data=(X_test, y_test)) pred = clf.predict(X_test) final_acc = (pred == y_test).mean() print("final testing accuracy: %.4f" % final_acc)
from rnn_clf import RNNClassifier import numpy as np import tensorflow as tf if __name__ == '__main__': (X_train, y_train), (X_test, y_test) = tf.contrib.keras.datasets.cifar10.load_data() X_train = (X_train / 255.0).mean(axis=3) X_test = (X_test / 255.0).mean(axis=3) clf = RNNClassifier(n_in=32, n_step=32, n_out=10) log = clf.fit(X_train, y_train, val_data=(X_test, y_test)) pred = clf.predict(X_test) final_acc = (pred == y_test).mean() print("final testing accuracy: %.4f" % final_acc)
from rnn_clf import RNNClassifier import tensorflow as tf n_in = 28 cell_size = 128 n_layer = 2 n_out = 10 batch_size = 128 n_epoch = 1 if __name__ == '__main__': (X_train, y_train), (X_test, y_test) = tf.contrib.keras.datasets.mnist.load_data() X_train = X_train / 255.0 X_test = X_test / 255.0 rnn = RNNClassifier(n_in, n_out, cell_size, n_layer) rnn.fit(X_train, y_train, n_epoch, batch_size) rnn.evaluate(X_test, y_test, batch_size)
# # linear_output = linear(final_output) # print("Shape of linear output: ", linear_output.shape) # # softmax_output = softmax(linear_output) # print("Shape of softmax output: ", softmax_output.shape) # print("Shape of target: ", y.shape) # # loss = criterion(softmax_output, y) # print("Loss value: ", loss.data.numpy()) from rnn_clf import RNNClassifier model = RNNClassifier(vocal_size=vocal_size, embedding_dim=100, hidden_dim=50, output_dim=label_size, batch_size=1) optim = SGD(params=model.parameters(), lr=0.01) criterion = NLLLoss() for i in range(10): total_loss = 0 model.train() for it, ex in enumerate(train_data): f, t = ex X = torch.LongTensor(f) y = torch.LongTensor(t) model.hidden = model.init_hidden() output = model.forward(X)