File size: 3,145 Bytes
679c1f9 |
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 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 |
# Copyright (c) 2025 SparkAudio
# 2025 Xinsheng Wang ([email protected])
#
# 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.
import torch
import torch.nn as nn
from typing import List
from sparktts.modules.blocks.vocos import VocosBackbone
from sparktts.modules.blocks.samper import SamplingBlock
class Encoder(nn.Module):
"""Encoder module with convnext and downsampling blocks"""
def __init__(
self,
input_channels: int,
vocos_dim: int,
vocos_intermediate_dim: int,
vocos_num_layers: int,
out_channels: int,
sample_ratios: List[int] = [1, 1],
):
super().__init__()
"""
Encoder module with VocosBackbone and sampling blocks.
Args:
sample_ratios (List[int]): sample ratios
example: [2, 2] means downsample by 2x and then upsample by 2x
"""
self.encoder = VocosBackbone(
input_channels=input_channels,
dim=vocos_dim,
intermediate_dim=vocos_intermediate_dim,
num_layers=vocos_num_layers,
condition_dim=None,
)
modules = [
nn.Sequential(
SamplingBlock(
dim=vocos_dim,
groups=vocos_dim,
downsample_scale=ratio,
),
VocosBackbone(
input_channels=vocos_dim,
dim=vocos_dim,
intermediate_dim=vocos_intermediate_dim,
num_layers=2,
condition_dim=None,
),
)
for ratio in sample_ratios
]
self.downsample = nn.Sequential(*modules)
self.project = nn.Linear(vocos_dim, out_channels)
def forward(self, x: torch.Tensor, *args):
"""
Args:
x (torch.Tensor): (batch_size, input_channels, length)
Returns:
x (torch.Tensor): (batch_size, encode_channels, length)
"""
x = self.encoder(x)
x = self.downsample(x)
x = self.project(x)
return x.transpose(1, 2)
# test
if __name__ == "__main__":
test_input = torch.randn(8, 1024, 50) # Batch size = 8, 1024 channels, length = 50
encoder = Encoder(
input_channels=1024,
vocos_dim=384,
vocos_intermediate_dim=2048,
vocos_num_layers=12,
out_channels=256,
sample_ratios=[2, 2],
)
output = encoder(test_input)
print(output.shape) # torch.Size([8, 256, 12])
if output.shape == torch.Size([8, 256, 12]):
print("test successful")
|