| | |
|---|
| | | from funasr.modules.nets_utils import make_pad_mask
|
|---|
| | |
|
|---|
| | | class CifPredictor(nn.Module):
|
|---|
| | | def __init__(self, idim, l_order, r_order, threshold=1.0, dropout=0.1, smooth_factor=1.0, noise_threshold=0):
|
|---|
| | | def __init__(self, idim, l_order, r_order, threshold=1.0, dropout=0.1, smooth_factor=1.0, noise_threshold=0, tail_threshold=0.45):
|
|---|
| | | super(CifPredictor, self).__init__()
|
|---|
| | |
|
|---|
| | | self.pad = nn.ConstantPad1d((l_order, r_order), 0)
|
|---|
| | |
|---|
| | | b, t, d = hidden.size()
|
|---|
| | | tail_threshold = self.tail_threshold
|
|---|
| | | tail_threshold = torch.tensor([tail_threshold], dtype=alphas.dtype).to(alphas.device)
|
|---|
| | | tail_threshold = torch.reshape(tail_threshold, (1, 1))
|
|---|
| | | tail_threshold = tail_threshold.unsqueeze(0).repeat(b, 1)
|
|---|
| | | alphas = torch.cat([alphas, tail_threshold], dim=1)
|
|---|
| | | zeros = torch.zeros((b, 1, d), dtype=hidden.dtype).to(hidden.device)
|
|---|
| | | hidden = torch.cat([hidden, zeros], dim=1)
|
|---|
