def __init__(self, index, token_name, classes):
super(OneHotEncoding, self).__init__()
vocab = Vocab(token_name)
vocab.init()
for c in classes[index]:
vocab.add(c)
num_class = len(vocab)
self.num_class = num_class
one_hot_encoding = torch.eye(num_class)
self.one_hots = nn.Parameter(one_hot_encoding, requires_grad=False)
def __init__(
self,
index,
token_name,
classes,
dropout=0,
embed_dim=15,
trainable=True,
padding_idx=None,
max_norm=None,
norm_type=2,
scale_grad_by_freq=False,
sparse=False,
):
super(SparseToEmbedding, self).__init__()
self.embed_dim = embed_dim
vocab = Vocab(token_name)
vocab.init()
for c in classes[index]:
vocab.add(c)
embedding_params = {
"vocab": vocab,
"dropout": dropout,
"embed_dim": embed_dim,
"trainable": trainable,
"padding_idx": padding_idx,
"max_norm": max_norm,
"norm_type": norm_type,
"scale_grad_by_freq": scale_grad_by_freq,
"sparse": sparse,
}
self.embedding = WordEmbedding(**embedding_params)