Spaces:
Runtime error
Runtime error
File size: 1,841 Bytes
be9690e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 |
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import Tuple
import torch.nn as nn
from torch.nn import functional as F
from cosyvoice.utils.mask import make_pad_mask
class InterpolateRegulator(nn.Module):
def __init__(
self,
channels: int,
sampling_ratios: Tuple,
out_channels: int = None,
groups: int = 1,
):
super().__init__()
self.sampling_ratios = sampling_ratios
out_channels = out_channels or channels
model = nn.ModuleList([])
if len(sampling_ratios) > 0:
for _ in sampling_ratios:
module = nn.Conv1d(channels, channels, 3, 1, 1)
norm = nn.GroupNorm(groups, channels)
act = nn.Mish()
model.extend([module, norm, act])
model.append(
nn.Conv1d(channels, out_channels, 1, 1)
)
self.model = nn.Sequential(*model)
def forward(self, x, ylens=None):
# x in (B, T, D)
mask = (~make_pad_mask(ylens)).to(x).unsqueeze(-1)
x = F.interpolate(x.transpose(1, 2).contiguous(), size=ylens.max(), mode='nearest')
out = self.model(x).transpose(1, 2).contiguous()
olens = ylens
return out * mask, olens
|