def __init__(self, embedding_tokens):
super(Net, self).__init__()
question_features = 1024
vision_features = config.output_features
glimpses = 2
objects = 10
self.text = TextProcessor(
embedding_tokens=len(embedding_tokens) + 1,
embedding_features=300,
lstm_features=question_features,
drop=0.5,
)
self.attention = Attention(
v_features=vision_features,
q_features=question_features,
mid_features=512,
glimpses=glimpses,
drop=0.5,
)
self.classifier = Classifier(
in_features=(glimpses * vision_features, question_features),
mid_features=1024,
out_features=config.max_answers,
count_features=objects + 1,
drop=0.5,
)
self.counter = counting.Counter(objects)
for m in self.modules():
if isinstance(m, nn.Linear) or isinstance(m, nn.Conv2d):
init.xavier_uniform_(m.weight)
if m.bias is not None:
m.bias.data.zero_()
def __init__(self, cf, confidence=None):
super(Baseline, self).__init__()
self.cf = cf
self.classifier = nn.Linear(cf + 1, cf + 1)
self.dummy = counting.Counter(cf,
already_sigmoided=True,
confidence=confidence)
init.eye_(self.classifier.weight)
def __init__(self, embedding_tokens):
super(Net, self).__init__()
question_features = 1024
vision_features = config.output_features
glimpses = 2
self.num_models = 2
objects = 10
tree_hidden_size = 1024
self.text = TextProcessor(
embedding_tokens=embedding_tokens,
embedding_features=300,
lstm_features=question_features,
drop=0.5,
)
self.attention = Attention(
v_features=vision_features,
q_features=question_features,
mid_features=512,
glimpses=glimpses,
drop=0.5,
)
self.tree_attention = Attention(
v_features=tree_hidden_size,
q_features=question_features,
mid_features=512,
glimpses=glimpses,
drop=0.5,
)
self.model_attention = ModelAttention(
q_features=question_features,
mid_features=1024,
q_type_num=65,
num_models=self.num_models,
drop=0.5,
)
self.classifier = Classifier(
in_features=(glimpses * vision_features, question_features),
mid_features=1024,
num_module=self.num_models,
out_features=config.max_answers,
tree_features=tree_hidden_size * glimpses,
count_features=objects + 1,
drop=0.5,
)
self.counter = counting.Counter(objects)
self.tree_lstm = tree_feature.TreeFeature(objects, vision_features,
tree_hidden_size)
for m in self.modules():
if isinstance(m, nn.Linear) or isinstance(m, nn.Conv2d):
init.xavier_uniform_(m.weight)
if m.bias is not None:
m.bias.data.zero_()
def __init__(self, cf):
super(Baseline, self).__init__()
self.cf = cf
self.classifier = nn.Linear(cf + 1, cf + 1)
self.dummy = counting.Counter(cf)
init.eye_(self.classifier.weight)
def __init__(self, cf):
super(Net, self).__init__()
self.cf = cf
self.counter = counting.Counter(cf)
self.classifier = nn.Linear(cf + 1, cf + 1)
init.eye_(self.classifier.weight)
def __init__(self, cf):
super(Net, self).__init__()
self.cf = cf
self.counter = counting.Counter(cf, already_sigmoided=True)
self.classifier = nn.Linear(cf + 1, cf + 1)
init.eye(self.classifier.weight)
