In [4]:
# Data
import numpy as np
with open('data/text_data/japan.txt', 'r') as f:
# with open('data/text_data/anna.txt', 'r') as f:
txt = f.read()
X = []
y = []
char_to_idx = {char: i for i, char in enumerate(set(txt))}
idx_to_char = {i: char for i, char in enumerate(set(txt))}
X = np.array([char_to_idx[x] for x in txt])
y = [char_to_idx[x] for x in txt[1:]]
y.append(char_to_idx['.'])
y = np.array(y)
# Looking at the X, y
X.shape, y.shape, X[:10], y[:10]
Out[4]:
In [5]:
# Model or Network
import impl.layer as l
class GRU:
def __init__(self, D, H, L, char2idx, idx2char):
self.D = D
self.H = H
self.L = L
self.char2idx = char2idx
self.idx2char = idx2char
self.vocab_size = len(char2idx)
self.losses = {'train':[], 'smooth train':[]}
# Model params
Z = H + D
m = dict(
Wz=np.random.randn(Z, H) / np.sqrt(Z / 2.),
Wh=np.random.randn(Z, H) / np.sqrt(Z / 2.),
Wy=np.random.randn(H, D) / np.sqrt(H / 2.),
bz=np.zeros((1, H)),
bh=np.zeros((1, H)),
by=np.zeros((1, D))
)
self.model = []
for _ in range(self.L):
self.model.append(m)
# Bi-directional
self.model_l = []
for _ in range(self.L):
self.model_l.append(m)
def initial_state(self):
return np.zeros((1, self.H))
def forward(self, X, h, m):
Wz, Wh, Wy = m['Wz'], m['Wh'], m['Wy']
bz, bh, by = m['bz'], m['bh'], m['by']
X_in = X.copy()
h_in = h.copy()
X = np.column_stack((h_in, X_in))
hz, hz_cache = l.fc_forward(X, Wz, bz)
hz, hz_sigm_cache = l.sigmoid_forward(hz)
hh, hh_cache = l.fc_forward(X, Wh, bh)
hh, hh_tanh_cache = l.tanh_forward(hh)
# h = (1. - hz) * h_old + hz * hh
# or
# h = ((1. - hz) * h_in) + (hz * hh)
# or
h = h_in + (hz * (hh - h_in))
y, y_cache = l.fc_forward(h, Wy, by)
cache = (h_in, hz, hz_cache, hz_sigm_cache, hh, hh_cache, hh_tanh_cache, y_cache)
return y, h, cache
def backward(self, dy, dh, cache):
h_in, hz, hz_cache, hz_sigm_cache, hh, hh_cache, hh_tanh_cache, y_cache = cache
dh_out = dh.copy()
dh, dWy, dby = l.fc_backward(dy, y_cache)
dh += dh_out
dh_in1 = (1. - hz) * dh
dhh = hz * dh
dhh = l.tanh_backward(dhh, hh_tanh_cache)
dXh, dWh, dbh = l.fc_backward(dhh, hh_cache)
# dhz = (hh * dh) - (h_in * dh)
# or
dhz = (hh - h_in) * dh
dhz = l.sigmoid_backward(dhz, hz_sigm_cache)
dXz, dWz, dbz = l.fc_backward(dhz, hz_cache)
dX = dXh + dXz
dh_in2 = dX[:, :self.H]
dX_in = dX[:, self.H:]
dh = dh_in1 + dh_in2
dX = dX_in
grad = dict(Wz=dWz, Wh=dWh, Wy=dWy, bz=dbz, bh=dbh, by=dby)
return dX, dh, grad
def train_forward(self, X_train, h):
ys, caches = [], []
h_init = h.copy()
h = []
for _ in range(self.L):
h.append(h_init.copy())
caches.append([])
# for X in X_train:
for t in range(len(X_train)):
X = X_train[t]
X_one_hot = np.zeros(self.D)
X_one_hot[X] = 1.
X = X_one_hot.reshape(1, -1)
for layer in range(self.L):
y, h[layer], cache = self.forward(X, h[layer], self.model[layer])
caches[layer].append(cache)
X = y.copy()
ys.append(y)
return ys, caches
def train_forward_l(self, X_train, h):
ys, caches = [], []
h_init = h.copy()
h = []
for _ in range(self.L):
h.append(h_init.copy())
caches.append([])
# for X in X_train:
for t in reversed(range(len(X_train))):
X = X_train[t]
X_one_hot = np.zeros(self.D)
X_one_hot[X] = 1.
X = X_one_hot.reshape(1, -1)
for layer in range(self.L):
y, h[layer], cache = self.forward(X, h[layer], self.model_l[layer])
caches[layer].append(cache)
X = y.copy()
ys.append(y)
return ys, caches
def cross_entropy(self, y_pred, y_train):
m = y_pred.shape[0]
prob = l.softmax(y_pred)
log_like = -np.log(prob[range(m), y_train])
data_loss = np.sum(log_like) / m
return data_loss
def dcross_entropy(self, y_pred, y_train):
m = y_pred.shape[0]
grad_y = l.softmax(y_pred)
grad_y[range(m), y_train] -= 1.0
grad_y /= m
return grad_y
def loss_function(self, y_train, ys):
loss, dys = 0.0, []
for y_pred, y in zip(ys, y_train):
loss += self.cross_entropy(y_pred, y)
dy = self.dcross_entropy(y_pred, y)
dys.append(dy)
return loss, dys
def train_backward(self, dys, caches):
dh, grad, grads = [], [], []
for layer in range(self.L):
dh.append(np.zeros((1, self.H)))
grad.append({key: np.zeros_like(val) for key, val in self.model[layer].items()})
grads.append({key: np.zeros_like(val) for key, val in self.model[layer].items()})
dXs = []
for t in reversed(range(len(dys))):
dy = dys[t]
for layer in reversed(range(self.L)):
dX, dh[layer], grad[layer] = self.backward(dy, dh[layer], caches[layer][t])
for k in grad[layer].keys():
grads[layer][k] += grad[layer][k]
dy = dX.copy()
dXs.append(dX)
return dXs, grads
def train_backward_l(self, dys, caches):
dh, grad, grads = [], [], []
for layer in range(self.L):
dh.append(np.zeros((1, self.H)))
grad.append({key: np.zeros_like(val) for key, val in self.model_l[layer].items()})
grads.append({key: np.zeros_like(val) for key, val in self.model_l[layer].items()})
dXs = []
for t in range(len(dys)):
dy = dys[t]
for layer in reversed(range(self.L)):
dX, dh[layer], grad[layer] = self.backward(dy, dh[layer], caches[layer][t])
for k in grad[layer].keys():
grads[layer][k] += grad[layer][k]
dy = dX.copy()
dXs.append(dX)
return dXs, grads
# def test(self, X_seed, h, size):
# chars = [self.idx2char[X_seed]]
# idx_list = list(range(self.vocab_size))
# X = X_seed
# h_init = h.copy()
# h = []
# for _ in range(self.L):
# h.append(h_init.copy())
# for _ in range(size):
# # Right-directional
# X_one_hot = np.zeros(self.D)
# X_one_hot[X] = 1.
# X = X_one_hot.reshape(1, -1)
# for layer in range(self.L):
# y, h[layer], _ = self.forward(X, h[layer], self.model[layer])
# X = y.copy()
# # Left-directional
# X_one_hot = np.zeros(self.D)
# X_one_hot[X] = 1.
# X = X_one_hot.reshape(1, -1)
# for layer in range(self.L):
# y, h[layer], _ = self.forward(X, h[layer], self.model[layer])
# X = y.copy()
# prob = l.softmax(y)
# idx = np.random.choice(idx_list, p=prob.ravel())
# chars.append(self.idx2char[idx])
# X = idx
# return ''.join(chars)
In [6]:
def get_minibatch(X, y, minibatch_size, shuffle):
minibatches = []
#for i in range(0, X.shape[0], minibatch_size):
for i in range(0, X.shape[0] - minibatch_size + 1, 1):
X_mini = X[i:i + minibatch_size]
y_mini = y[i:i + minibatch_size]
minibatches.append((X_mini, y_mini))
return minibatches
def adam_rnn(nn, X_train, y_train, alpha, mb_size, n_iter, print_after):
M, R = [], []
for layer in range(nn.L):
M.append({k: np.zeros_like(v) for k, v in nn.model[layer].items()})
R.append({k: np.zeros_like(v) for k, v in nn.model[layer].items()})
M_l, R_l = [], []
for layer in range(nn.L):
M_l.append({k: np.zeros_like(v) for k, v in nn.model_l[layer].items()})
R_l.append({k: np.zeros_like(v) for k, v in nn.model_l[layer].items()})
beta1 = .99
beta2 = .999
state = nn.initial_state()
smooth_loss = 1.
eps = 1e-8
minibatches = get_minibatch(X_train, y_train, mb_size, shuffle=False)
for iter in range(1, n_iter + 1):
for idx in range(len(minibatches)):
X_mini, y_mini = minibatches[idx]
ys, caches = nn.train_forward(X_mini, state)
ys_l, caches_l = nn.train_forward_l(X_mini, state)
ys += ys_l
loss, dys = nn.loss_function(y_mini, ys)
_, grads = nn.train_backward(dys, caches)
_, grads_l = nn.train_backward_l(dys, caches_l)
nn.losses['train'].append(loss)
smooth_loss = (0.999 * smooth_loss) + (0.001 * loss)
nn.losses['smooth train'].append(smooth_loss)
for layer in range(nn.L):
for k in grads[layer].keys(): #key, value: items
M[layer][k] = l.exp_running_avg(M[layer][k], grads[layer][k], beta1)
R[layer][k] = l.exp_running_avg(R[layer][k], grads[layer][k]**2, beta2)
m_k_hat = M[layer][k] / (1. - (beta1**(iter)))
r_k_hat = R[layer][k] / (1. - (beta2**(iter)))
nn.model[layer][k] -= alpha * m_k_hat / (np.sqrt(r_k_hat) + eps)
for k in grads_l[layer].keys(): #key, value: items
M_l[layer][k] = l.exp_running_avg(M_l[layer][k], grads_l[layer][k], beta1)
R_l[layer][k] = l.exp_running_avg(R_l[layer][k], grads_l[layer][k]**2, beta2)
m_k_hat = M_l[layer][k] / (1. - (beta1**(iter)))
r_k_hat = R_l[layer][k] / (1. - (beta2**(iter)))
nn.model_l[layer][k] -= alpha * m_k_hat / (np.sqrt(r_k_hat) + eps)
# Print loss and test sample
if iter % print_after == 0:
print('Iter-{} loss: {:.4f}'.format(iter, loss))
# sample = nn.test(X_mini[0], state, 100)
# print(sample)
return nn
In [7]:
# Hyper-parameters
time_step = 100 # width, minibatch size and test sample size as well
num_layers = 2 # depth
n_iter = 300 # epochs
alpha = 1e-4 # learning_rate
print_after = 1 # n_iter//10 # print training loss, valid, and test
num_hidden_units = 64 # num_hidden_units in hidden layer
num_input_units = len(char_to_idx) # vocab_size = len(char_to_idx)
# Build the network and learning it or optimizing it using SGD
net = GRU(D=num_input_units, H=num_hidden_units, L=num_layers, char2idx=char_to_idx, idx2char=idx_to_char)
# Start learning using BP-SGD-ADAM
adam_rnn(nn=net, X_train=X, y_train=y, alpha=alpha, mb_size=time_step, n_iter=n_iter, print_after=print_after)
# # Display the learning curve and losses for training, validation, and testing
# %matplotlib inline
# %config InlineBackend.figure_format = 'retina'
import matplotlib.pyplot as plt
plt.plot(net.losses['train'], label='Train loss')
plt.plot(net.losses['smooth train'], label='Train smooth loss')
plt.legend()
plt.show()
In [5]:
# # Hyper-parameters
# time_step = 100 # width, minibatch size and test sample size as well
# num_layers = 1 # depth
# n_iter = 300 # epochs
# alpha = 1e-4 # learning_rate
# print_after = 1 # n_iter//10 # print training loss, valid, and test
# num_hidden_units = 64 # num_hidden_units in hidden layer
# num_input_units = len(char_to_idx) # vocab_size = len(char_to_idx)
# # Build the network and learning it or optimizing it using SGD
# net = GRU(D=num_input_units, H=num_hidden_units, L=num_layers, char2idx=char_to_idx, idx2char=idx_to_char)
# # Start learning using BP-SGD-ADAM
# adam_rnn(nn=net, X_train=X, y_train=y, alpha=alpha, mb_size=time_step, n_iter=n_iter, print_after=print_after)
# # # Display the learning curve and losses for training, validation, and testing
# # %matplotlib inline
# # %config InlineBackend.figure_format = 'retina'
# import matplotlib.pyplot as plt
# plt.plot(net.losses['train'], label='Train loss')
# plt.plot(net.losses['smooth train'], label='Train smooth loss')
# plt.legend()
# plt.show()
In [ ]: