def create_network(window, input): network_first = create_regressor(batchsize=batchsize, window=window, input=input, dropout=0.0) network_second = create_core(batchsize=batchsize, window=window, dropout=0.0, depooler=lambda x, **kw: x / 2) network_second.load(np.load('network_core.npz')) network = Network(network_first, network_second[1], params=network_first.params) network.load(np.load('network_regression.npz')) return network_first, network_second, network
feet = np.array([9, 10, 11, 12, 13, 14, 21, 22, 23, 24, 25, 26]) Y = X[:, feet] I = np.arange(len(X)) rng.shuffle(I) X, Y = X[I], Y[I] batchsize = 1 window = X.shape[2] network_second = create_core(batchsize=batchsize, window=window, dropout=0.0, depooler=lambda x, **kw: x / 2) network_second.load(np.load('network_core.npz')) network_first = create_regressor(batchsize=batchsize, window=window, input=Y.shape[1]) network = Network(network_first, network_second[1], params=network_first.params) E = theano.shared(X, borrow=True) F = theano.shared(Y, borrow=True) trainer = AdamTrainer(rng=rng, batchsize=batchsize, epochs=100, alpha=0.00001) trainer.train(network, F, E, filename='network_regression_kick.npz')
feet = np.array([27, 28, 29]) Y = X[:, feet] print('X:', X) print('Y: ', Y) #Y shape: 121,12,240 #m.getch() frames = np.hstack(np.arange(116, 125)) test = Y[:, :, frames] test_gt = X[:, :, frames] print('test:', test.shape) batchsize = 1 window = test.shape[2] network_first = create_regressor(batchsize=batchsize, window=window, input=test.shape[1], dropout=0.0) network_second = create_core(batchsize=batchsize, window=window, dropout=0.0, depooler=lambda x, **kw: x / 2) #print('network_second[1]: ',network_second[0]) #print('network_second params: ',network_second.params) #network_second.params=[W,b,W,b] #print('after network_second') network_second.load(np.load('network_core.npz')) #print('after network_second load') #print('network_second params: ',network_second.params) #network_second.params=[W,b,W,b] network = Network(network_first,
rng.shuffle(I) data_train = data[I[:len(data) // 2]] data_valid = data[I[len(data) // 2:]] X = data_valid X = np.swapaxes(X, 1, 2).astype(theano.config.floatX) preprocess = np.load('preprocess_core.npz') X = (X - preprocess['Xmean']) / preprocess['Xstd'] Y = X[:, -7:] batchsize = 1 network_first = create_regressor(batchsize=batchsize, window=X.shape[2], input=Y.shape[1], dropout=0.0) network_second = create_core(batchsize=batchsize, window=X.shape[2], dropout=0.0, depooler=lambda x, **kw: x / 2) network = Network(network_first, network_second[1], params=network_first.params) network.load(np.load('network_regression.npz')) from AnimationPlot import animation_plot for i in range(5): index = rng.randint(len(X) - 1)