def __init__(self):
melt.PredictorBase.__init__(self)
ShowAndTell.__init__(self, is_training=False, is_predict=True)
if FLAGS.pre_calc_image_feature:
self.image_feature_len = IMAGE_FEATURE_LEN
self.image_feature_feed = tf.placeholder(
tf.float32, [None, self.image_feature_len],
name='image_feature')
else:
self.image_feature_feed = tf.placeholder(tf.string, [
None,
],
name='image_feature')
tf.add_to_collection('feed', self.image_feature_feed)
tf.add_to_collection('lfeed', self.image_feature_feed)
self.text_feed = tf.placeholder(tf.int64, [None, TEXT_MAX_WORDS],
name='text')
tf.add_to_collection('rfeed', self.text_feed)
self.beam_text = None
self.beam_text_score = None
self.image_model = None
def __init__(self):
#super(ShowAndTellPredictor, self).__init__()
melt.PredictorBase.__init__(self)
ShowAndTell.__init__(self, is_training=False, is_predict=True)
self.text_list = []
self.image_feature_place = tf.placeholder(tf.float32, [None, IMAGE_FEATURE_LEN], name='image_feature')
self.text_place = tf.placeholder(tf.int64, [None, TEXT_MAX_WORDS], name='text')
def __init__(self):
melt.PredictorBase.__init__(self)
ShowAndTell.__init__(self, is_training=False, is_predict=True)
if FLAGS.pre_calc_image_feature:
self.image_feature_len = FLAGS.image_feature_len or IMAGE_FEATURE_LEN
#TODO for rl, need use feed dict, so predict will introduce ... need to feed, how to use with_default?
#self.image_feature_feed = tf.placeholder(tf.float32, [None, self.image_feature_len], name='image_feature')
self.image_feature_feed = tf.placeholder_with_default(
[[0.] * self.image_feature_len],
[None, self.image_feature_len],
name='image_feature')
else:
#self.image_feature_feed = tf.placeholder(tf.string, [None,], name='image_feature')
# TODO HACK for nasnet... need this due to using average decay
if os.path.exists('./test.jpg'):
test_image = melt.read_image('./test.jpg')
elif os.path.exists('/tmp/test.jpg'):
test_image = melt.read_image('/tmp/test.jpg')
else:
test_image = None
if test_image is not None:
self.image_feature_feed = tf.placeholder_with_default(
tf.constant([test_image]), [
None,
], name='image_feature')
else:
assert not FLAGS.image_model_name.startswith(
'nasnet'
), 'HACK for nasnet you need one test.jpg in current path or /tmp/ path'
self.image_feature_feed = tf.placeholder(tf.string, [
None,
],
name='image_feature')
tf.add_to_collection('feed', self.image_feature_feed)
tf.add_to_collection('lfeed', self.image_feature_feed)
self.text_feed = tf.placeholder(tf.int64, [None, TEXT_MAX_WORDS],
name='text')
tf.add_to_collection('rfeed', self.text_feed)
self.text = None
self.text_score = None
self.beam_text = None
self.beam_text_score = None
self.image_model = None
self.logprobs_history = False
self.alignment_history = False
self.feed_dict = {}
def _gen_builder(algo, is_predict=True):
"""
Args:
is_predict: set to False if only train, no need for predict/eval
"""
if is_predict:
if algo == Algos.bow:
return BowPredictor()
elif algo == Algos.show_and_tell:
return ShowAndTellPredictor()
elif algo == Algos.rnn:
return RnnPredictor()
elif algo == Algos.seq2seq:
return Seq2seqPredictor()
else:
raise ValueError('Unsupported algo %s'%algo)
else:
if algo == Algos.bow:
return Bow()
elif algo == Algos.show_and_tell:
return ShowAndTell()
elif algo == Algos.rnn:
return Rnn()
elif algo == Algos.seq2seq:
return Seq2seq()
else:
raise ValueError('Unsupported algo %s'%algo)
def __init__(self):
#super(ShowAndTellPredictor, self).__init__()
melt.PredictorBase.__init__(self)
ShowAndTell.__init__(self, is_training=False, is_predict=True)
if FLAGS.pre_calc_image_feature:
self.image_feature_len = IMAGE_FEATURE_LEN
self.image_feature_feed = tf.placeholder(
tf.float32, [None, self.image_feature_len],
name='image_feature')
else:
self.image_feature_len = 2048
self.image_feature_feed = tf.placeholder(tf.string, [
None,
],
name='image_feature')
self.text_feed = tf.placeholder(tf.int64, [None, TEXT_MAX_WORDS],
name='text')
def _gen_trainer(algo):
if algo == Algos.bow:
return Bow()
elif algo == Algos.show_and_tell:
return ShowAndTell()
elif algo == Algos.rnn:
return Rnn()
elif algo == Algos.pooling:
return Pooling()
elif algo == Algos.seq2seq:
return Seq2seq()
elif algo == Algos.imtxt2txt:
return Imtxt2txt()
else:
raise ValueError('Unsupported algo %s' % algo)
def _gen_trainer(algo):
if algo == Algos.bow:
return DiscriminantTrainer('bow')
elif algo == Algos.rnn:
return DiscriminantTrainer('rnn')
elif algo == Algos.cnn:
return DiscriminantTrainer('cnn')
elif algo == Algos.seq2seq:
return Seq2seq()
elif algo == Algos.show_and_tell:
return ShowAndTell()
elif algo == Algos.imtxt2txt:
return Imtxt2txt()
elif algo == Algos.dual_bow:
return DualTextsim('bow')
elif algo == Algos.dual_rnn:
return DualTextsim('rnn')
elif algo == Algos.dual_cnn:
return DualTextsim('cnn')
else:
raise ValueError('Unsupported algo %s' % algo)