def test_ctc_lambda_loss(self):
''' test ctc loss '''
with self.cached_session():
label_lens = np.expand_dims(np.asarray([5, 4]), 1)
input_lens = np.expand_dims(np.asarray([5, 5]), 1) # number of timesteps
loss_log_probs = [7.2771974, 8.057934]
# dimensions are batch x time x categories
labels = np.asarray([[0, 1, 2, 1, 0], [0, 1, 1, 0, -1]])
inputs = np.asarray(
[[[0.633766, 0.221185, 0.0917319, 0.0129757, 0.0142857, 0.0260553],
[0.111121, 0.588392, 0.278779, 0.0055756, 0.00569609, 0.010436],
[0.0357786, 0.633813, 0.321418, 0.00249248, 0.00272882, 0.0037688],
[0.0663296, 0.643849, 0.280111, 0.00283995, 0.0035545, 0.00331533],
[0.458235, 0.396634, 0.123377, 0.00648837, 0.00903441, 0.00623107]],
[[0.30176, 0.28562, 0.0831517, 0.0862751, 0.0816851, 0.161508],
[0.24082, 0.397533, 0.0557226, 0.0546814, 0.0557528, 0.19549],
[0.230246, 0.450868, 0.0389607, 0.038309, 0.0391602, 0.202456],
[0.280884, 0.429522, 0.0326593, 0.0339046, 0.0326856, 0.190345],
[0.423286, 0.315517, 0.0338439, 0.0393744, 0.0339315, 0.154046]]],
dtype=np.float32)
loss = loss_utils.ctc_lambda_loss(
labels=tf.constant(labels),
logits=tf.constant(inputs),
input_length=tf.constant(input_lens),
label_length=tf.constant(label_lens),
blank_index=0)
self.assertEqual(loss.eval().shape[0], inputs.shape[0])
self.assertAllClose(loss.eval(), loss_log_probs, atol=1e-05)
self.assertAllClose(
np.mean(loss.eval()), np.mean(loss_log_probs), atol=1e-05)
# test when batch_size = 1, that is, one sample only
ref = [7.277198]
input_lens = np.asarray([5])
label_lens = np.asarray([5])
labels = np.asarray([[0, 1, 2, 1, 0]])
inputs = np.asarray(
[[[0.633766, 0.221185, 0.0917319, 0.0129757, 0.0142857, 0.0260553],
[0.111121, 0.588392, 0.278779, 0.0055756, 0.00569609, 0.010436],
[0.0357786, 0.633813, 0.321418, 0.00249248, 0.00272882, 0.0037688],
[0.0663296, 0.643849, 0.280111, 0.00283995, 0.0035545, 0.00331533],
[0.458235, 0.396634, 0.123377, 0.00648837, 0.00903441, 0.00623107]]
],
dtype=np.float32)
loss = loss_utils.ctc_lambda_loss(
labels=tf.constant(labels),
logits=tf.constant(inputs),
input_length=tf.constant(input_lens),
label_length=tf.constant(label_lens),
blank_index=0)
self.assertAllClose(loss.eval(), ref, atol=1e-05)
self.assertAllClose(np.mean(loss.eval()), np.mean(ref), atol=1e-05)
def call(self,
logits=None,
input_length=None,
labels=None,
label_length=None,
soft_labels=None):
del soft_labels
return ctc_lambda_loss(logits=logits,
input_length=input_length,
labels=labels,
label_length=label_length)
def call(self,
logits=None,
input_length=None,
labels=None,
label_length=None,
**kwargs):
blank_index = kwargs.get('blank_index', 0)
return ctc_lambda_loss(logits=logits,
input_length=input_length,
labels=labels,
label_length=label_length,
blank_index=blank_index)
