Image Captioning with LSTMs

In the previous exercise you implemented a vanilla RNN and applied it to image captioning. In this notebook you will implement the LSTM update rule and use it for image captioning.


In [1]:
# As usual, a bit of setup

import time, os, json
import numpy as np
import matplotlib.pyplot as plt

from skynet.utils.gradient_check import eval_numerical_gradient, eval_numerical_gradient_array
from skynet.neural_network.rnn.rnn_layers import *
from skynet.solvers.captioning_solver import CaptioningSolver
from skynet.neural_network.classifiers.rnn import CaptioningRNN
from skynet.utils.coco_utils import load_coco_data, sample_coco_minibatch, decode_captions
from skynet.utils.image_utils import image_from_url

%matplotlib inline
plt.rcParams['figure.figsize'] = (10.0, 8.0) # set default size of plots
plt.rcParams['image.interpolation'] = 'nearest'
plt.rcParams['image.cmap'] = 'gray'

# for auto-reloading external modules
# see http://stackoverflow.com/questions/1907993/autoreload-of-modules-in-ipython
%load_ext autoreload
%autoreload 2

def rel_error(x, y):
  """ returns relative error """
  return np.max(np.abs(x - y) / (np.maximum(1e-8, np.abs(x) + np.abs(y))))

Load MS-COCO data

As in the previous notebook, we will use the Microsoft COCO dataset for captioning.


In [2]:
# Load COCO data from disk; this returns a dictionary
# We'll work with dimensionality-reduced features for this notebook, but feel
# free to experiment with the original features by changing the flag below.
data = load_coco_data(pca_features=True)

# Print out all the keys and values from the data dictionary
for k, v in data.items():
  if type(v) == np.ndarray:
    print(k, type(v), v.shape, v.dtype)
  else:
    print(k, type(v), len(v))


idx_to_word <type 'list'> 1004
train_captions <type 'numpy.ndarray'> (400135, 17) int32
val_captions <type 'numpy.ndarray'> (195954, 17) int32
train_image_idxs <type 'numpy.ndarray'> (400135,) int32
val_features <type 'numpy.ndarray'> (40504, 512) float32
val_image_idxs <type 'numpy.ndarray'> (195954,) int32
train_features <type 'numpy.ndarray'> (82783, 512) float32
train_urls <type 'numpy.ndarray'> (82783,) |S63
val_urls <type 'numpy.ndarray'> (40504,) |S63
word_to_idx <type 'dict'> 1004

LSTM

If you read recent papers, you'll see that many people use a variant on the vanialla RNN called Long-Short Term Memory (LSTM) RNNs. Vanilla RNNs can be tough to train on long sequences due to vanishing and exploding gradiants caused by repeated matrix multiplication. LSTMs solve this problem by replacing the simple update rule of the vanilla RNN with a gating mechanism as follows.

Similar to the vanilla RNN, at each timestep we receive an input $x_t\in\mathbb{R}^D$ and the previous hidden state $h_{t-1}\in\mathbb{R}^H$; the LSTM also maintains an $H$-dimensional cell state, so we also receive the previous cell state $c_{t-1}\in\mathbb{R}^H$. The learnable parameters of the LSTM are an input-to-hidden matrix $W_x\in\mathbb{R}^{4H\times D}$, a hidden-to-hidden matrix $W_h\in\mathbb{R}^{4H\times H}$ and a bias vector $b\in\mathbb{R}^{4H}$.

At each timestep we first compute an activation vector $a\in\mathbb{R}^{4H}$ as $a=W_xx_t + W_hh_{t-1}+b$. We then divide this into four vectors $a_i,a_f,a_o,a_g\in\mathbb{R}^H$ where $a_i$ consists of the first $H$ elements of $a$, $a_f$ is the next $H$ elements of $a$, etc. We then compute the input gate $g\in\mathbb{R}^H$, forget gate $f\in\mathbb{R}^H$, output gate $o\in\mathbb{R}^H$ and block input $g\in\mathbb{R}^H$ as

$$ \begin{align*} i = \sigma(a_i) \hspace{2pc} f = \sigma(a_f) \hspace{2pc} o = \sigma(a_o) \hspace{2pc} g = \tanh(a_g) \end{align*} $$

where $\sigma$ is the sigmoid function and $\tanh$ is the hyperbolic tangent, both applied elementwise.

Finally we compute the next cell state $c_t$ and next hidden state $h_t$ as

$$ c_{t} = f\odot c_{t-1} + i\odot g \hspace{4pc} h_t = o\odot\tanh(c_t) $$

where $\odot$ is the elementwise product of vectors.

In the rest of the notebook we will implement the LSTM update rule and apply it to the image captioning task.

LSTM: step forward

Implement the forward pass for a single timestep of an LSTM in the lstm_step_forward function in the file cs231n/rnn_layers.py. This should be similar to the rnn_step_forward function that you implemented above, but using the LSTM update rule instead.

Once you are done, run the following to perform a simple test of your implementation. You should see errors around 1e-8 or less.


In [3]:
N, D, H = 3, 4, 5
x = np.linspace(-0.4, 1.2, num=N*D).reshape(N, D)
prev_h = np.linspace(-0.3, 0.7, num=N*H).reshape(N, H)
prev_c = np.linspace(-0.4, 0.9, num=N*H).reshape(N, H)
Wx = np.linspace(-2.1, 1.3, num=4*D*H).reshape(D, 4 * H)
Wh = np.linspace(-0.7, 2.2, num=4*H*H).reshape(H, 4 * H)
b = np.linspace(0.3, 0.7, num=4*H)

next_h, next_c, cache = lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)

expected_next_h = np.asarray([
    [ 0.24635157,  0.28610883,  0.32240467,  0.35525807,  0.38474904],
    [ 0.49223563,  0.55611431,  0.61507696,  0.66844003,  0.7159181 ],
    [ 0.56735664,  0.66310127,  0.74419266,  0.80889665,  0.858299  ]])
expected_next_c = np.asarray([
    [ 0.32986176,  0.39145139,  0.451556,    0.51014116,  0.56717407],
    [ 0.66382255,  0.76674007,  0.87195994,  0.97902709,  1.08751345],
    [ 0.74192008,  0.90592151,  1.07717006,  1.25120233,  1.42395676]])

print('next_h error: ', rel_error(expected_next_h, next_h))
print('next_c error: ', rel_error(expected_next_c, next_c))


next_h error:  5.70541304045e-09
next_c error:  5.81431230888e-09

LSTM: step backward

Implement the backward pass for a single LSTM timestep in the function lstm_step_backward in the file cs231n/rnn_layers.py. Once you are done, run the following to perform numeric gradient checking on your implementation. You should see errors around 1e-8 or less.


In [4]:
N, D, H = 4, 5, 6
x = np.random.randn(N, D)
prev_h = np.random.randn(N, H)
prev_c = np.random.randn(N, H)
Wx = np.random.randn(D, 4 * H)
Wh = np.random.randn(H, 4 * H)
b = np.random.randn(4 * H)

next_h, next_c, cache = lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)

dnext_h = np.random.randn(*next_h.shape)
dnext_c = np.random.randn(*next_c.shape)

fx_h = lambda x: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[0]
fh_h = lambda h: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[0]
fc_h = lambda c: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[0]
fWx_h = lambda Wx: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[0]
fWh_h = lambda Wh: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[0]
fb_h = lambda b: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[0]

fx_c = lambda x: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[1]
fh_c = lambda h: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[1]
fc_c = lambda c: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[1]
fWx_c = lambda Wx: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[1]
fWh_c = lambda Wh: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[1]
fb_c = lambda b: lstm_step_forward(x, prev_h, prev_c, Wx, Wh, b)[1]

num_grad = eval_numerical_gradient_array

dx_num = num_grad(fx_h, x, dnext_h) + num_grad(fx_c, x, dnext_c)
dh_num = num_grad(fh_h, prev_h, dnext_h) + num_grad(fh_c, prev_h, dnext_c)
dc_num = num_grad(fc_h, prev_c, dnext_h) + num_grad(fc_c, prev_c, dnext_c)
dWx_num = num_grad(fWx_h, Wx, dnext_h) + num_grad(fWx_c, Wx, dnext_c)
dWh_num = num_grad(fWh_h, Wh, dnext_h) + num_grad(fWh_c, Wh, dnext_c)
db_num = num_grad(fb_h, b, dnext_h) + num_grad(fb_c, b, dnext_c)

dx, dh, dc, dWx, dWh, db = lstm_step_backward(dnext_h, dnext_c, cache)

print('dx error: ', rel_error(dx_num, dx))
print('dh error: ', rel_error(dh_num, dh))
print('dc error: ', rel_error(dc_num, dc))
print('dWx error: ', rel_error(dWx_num, dWx))
print('dWh error: ', rel_error(dWh_num, dWh))
print('db error: ', rel_error(db_num, db))


dx error:  4.35289746272e-11
dh error:  7.69222255966e-11
dc error:  9.16414475384e-11
dWx error:  6.47028050073e-09
dWh error:  7.27723061557e-09
db error:  1.21500480123e-09

LSTM: forward

In the function lstm_forward in the file cs231n/rnn_layers.py, implement the lstm_forward function to run an LSTM forward on an entire timeseries of data.

When you are done run the following to check your implementation. You should see an error around 1e-7.


In [5]:
N, D, H, T = 2, 5, 4, 3
x = np.linspace(-0.4, 0.6, num=N*T*D).reshape(N, T, D)
h0 = np.linspace(-0.4, 0.8, num=N*H).reshape(N, H)
Wx = np.linspace(-0.2, 0.9, num=4*D*H).reshape(D, 4 * H)
Wh = np.linspace(-0.3, 0.6, num=4*H*H).reshape(H, 4 * H)
b = np.linspace(0.2, 0.7, num=4*H)

h, cache = lstm_forward(x, h0, Wx, Wh, b)

expected_h = np.asarray([
 [[ 0.01764008,  0.01823233,  0.01882671,  0.0194232 ],
  [ 0.11287491,  0.12146228,  0.13018446,  0.13902939],
  [ 0.31358768,  0.33338627,  0.35304453,  0.37250975]],
 [[ 0.45767879,  0.4761092,   0.4936887,   0.51041945],
  [ 0.6704845,   0.69350089,  0.71486014,  0.7346449 ],
  [ 0.81733511,  0.83677871,  0.85403753,  0.86935314]]])

print('h error: ', rel_error(expected_h, h))


h error:  8.61053745211e-08

LSTM: backward

Implement the backward pass for an LSTM over an entire timeseries of data in the function lstm_backward in the file cs231n/rnn_layers.py. When you are done run the following to perform numeric gradient checking on your implementation. You should see errors around 1e-8 or less.


In [6]:
from skynet.neural_network.rnn.rnn_layers import lstm_forward, lstm_backward

N, D, T, H = 2, 3, 10, 6

x = np.random.randn(N, T, D)
h0 = np.random.randn(N, H)
Wx = np.random.randn(D, 4 * H)
Wh = np.random.randn(H, 4 * H)
b = np.random.randn(4 * H)

out, cache = lstm_forward(x, h0, Wx, Wh, b)

dout = np.random.randn(*out.shape)

dx, dh0, dWx, dWh, db = lstm_backward(dout, cache)

fx = lambda x: lstm_forward(x, h0, Wx, Wh, b)[0]
fh0 = lambda h0: lstm_forward(x, h0, Wx, Wh, b)[0]
fWx = lambda Wx: lstm_forward(x, h0, Wx, Wh, b)[0]
fWh = lambda Wh: lstm_forward(x, h0, Wx, Wh, b)[0]
fb = lambda b: lstm_forward(x, h0, Wx, Wh, b)[0]

dx_num = eval_numerical_gradient_array(fx, x, dout)
dh0_num = eval_numerical_gradient_array(fh0, h0, dout)
dWx_num = eval_numerical_gradient_array(fWx, Wx, dout)
dWh_num = eval_numerical_gradient_array(fWh, Wh, dout)
db_num = eval_numerical_gradient_array(fb, b, dout)

print('dx error: ', rel_error(dx_num, dx))
print('dh0 error: ', rel_error(dh0_num, dh0))
print('dWx error: ', rel_error(dWx_num, dWx))
print('dWh error: ', rel_error(dWh_num, dWh))
print('db error: ', rel_error(db_num, db))


dx error:  5.59216900497e-09
dh0 error:  1.04667018058e-09
dWx error:  2.58274807977e-09
dWh error:  1.9476093782e-07
db error:  2.08157359332e-09

LSTM captioning model

Now that you have implemented an LSTM, update the implementation of the loss method of the CaptioningRNN class in the file neural_network/classifiers/rnn.py to handle the case where self.cell_type is lstm. This should require adding less than 10 lines of code.

Once you have done so, run the following to check your implementation. You should see a difference of less than 1e-10.


In [7]:
N, D, W, H = 10, 20, 30, 40
word_to_idx = {'<NULL>': 0, 'cat': 2, 'dog': 3}
V = len(word_to_idx)
T = 13

model = CaptioningRNN(word_to_idx,
          input_dim=D,
          wordvec_dim=W,
          hidden_dim=H,
          cell_type='lstm',
          dtype=np.float64)

# Set all model parameters to fixed values
for k, v in model.params.items():
  model.params[k] = np.linspace(-1.4, 1.3, num=v.size).reshape(*v.shape)

features = np.linspace(-0.5, 1.7, num=N*D).reshape(N, D)
captions = (np.arange(N * T) % V).reshape(N, T)

loss, grads = model.loss(features, captions)
expected_loss = 9.82445935443

print('loss: ', loss)
print('expected loss: ', expected_loss)
print('difference: ', abs(loss - expected_loss))


loss:  9.82445935443
expected loss:  9.82445935443
difference:  2.26485497024e-12

Overfit LSTM captioning model

Run the following to overfit an LSTM captioning model on the same small dataset as we used for the RNN above.


In [8]:
small_data = load_coco_data(max_train=50)

small_lstm_model = CaptioningRNN(
          cell_type='lstm',
          word_to_idx=data['word_to_idx'],
          input_dim=data['train_features'].shape[1],
          hidden_dim=512,
          wordvec_dim=256,
          dtype=np.float32,
        )

small_lstm_solver = CaptioningSolver(small_lstm_model, small_data,
           update_rule='adam',
           num_epochs=50,
           batch_size=25,
           optim_config={
             'learning_rate': 5e-3,
           },
           lr_decay=0.995,
           verbose=True, print_every=10,
         )

small_lstm_solver.train()

# Plot the training losses
plt.plot(small_lstm_solver.loss_history)
plt.xlabel('Iteration')
plt.ylabel('Loss')
plt.title('Training loss history')
plt.show()


(Iteration 1 / 100) loss: 77.462204
(Iteration 11 / 100) loss: 43.101875
(Iteration 21 / 100) loss: 21.793349
(Iteration 31 / 100) loss: 16.058987
(Iteration 41 / 100) loss: 5.733616
(Iteration 51 / 100) loss: 1.802905
(Iteration 61 / 100) loss: 0.796254
(Iteration 71 / 100) loss: 0.404397
(Iteration 81 / 100) loss: 0.228571
(Iteration 91 / 100) loss: 0.136641

LSTM test-time sampling

Modify the sample method of the CaptioningRNN class to handle the case where self.cell_type is lstm. This should take fewer than 10 lines of code.

When you are done run the following to sample from your overfit LSTM model on some training and validation set samples.


In [9]:
for split in ['train', 'val']:
  minibatch = sample_coco_minibatch(small_data, split=split, batch_size=2)
  gt_captions, features, urls = minibatch
  gt_captions = decode_captions(gt_captions, data['idx_to_word'])

  sample_captions = small_lstm_model.sample(features)
  sample_captions = decode_captions(sample_captions, data['idx_to_word'])

  for gt_caption, sample_caption, url in zip(gt_captions, sample_captions, urls):
    plt.imshow(image_from_url(url))
    plt.title('%s\n%s\nGT:%s' % (split, sample_caption, gt_caption))
    plt.axis('off')
    plt.show()