chore: vendor selva_core from jnwnlee/selva@d7d40a9
Pure PyTorch SelVA source for SelvaModelLoader/FeatureExtractor/Sampler nodes. Imports rewritten from selva.* to selva_core.*. mel_converter.py: replaced librosa.filters.mel with pure-numpy implementation to avoid librosa→numba→NumPy version incompatibility in some ComfyUI environments. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
@@ -0,0 +1,21 @@
|
||||
MIT License
|
||||
|
||||
Copyright (c) 2024 NVIDIA CORPORATION.
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
of this software and associated documentation files (the "Software"), to deal
|
||||
in the Software without restriction, including without limitation the rights
|
||||
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
||||
copies of the Software, and to permit persons to whom the Software is
|
||||
furnished to do so, subject to the following conditions:
|
||||
|
||||
The above copyright notice and this permission notice shall be included in all
|
||||
copies or substantial portions of the Software.
|
||||
|
||||
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
||||
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
||||
SOFTWARE.
|
||||
@@ -0,0 +1,126 @@
|
||||
# Implementation adapted from https://github.com/EdwardDixon/snake under the MIT license.
|
||||
# LICENSE is in incl_licenses directory.
|
||||
|
||||
import torch
|
||||
from torch import nn, sin, pow
|
||||
from torch.nn import Parameter
|
||||
|
||||
|
||||
class Snake(nn.Module):
|
||||
"""
|
||||
Implementation of a sine-based periodic activation function
|
||||
Shape:
|
||||
- Input: (B, C, T)
|
||||
- Output: (B, C, T), same shape as the input
|
||||
Parameters:
|
||||
- alpha - trainable parameter
|
||||
References:
|
||||
- This activation function is from this paper by Liu Ziyin, Tilman Hartwig, Masahito Ueda:
|
||||
https://arxiv.org/abs/2006.08195
|
||||
Examples:
|
||||
>>> a1 = snake(256)
|
||||
>>> x = torch.randn(256)
|
||||
>>> x = a1(x)
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self, in_features, alpha=1.0, alpha_trainable=True, alpha_logscale=False
|
||||
):
|
||||
"""
|
||||
Initialization.
|
||||
INPUT:
|
||||
- in_features: shape of the input
|
||||
- alpha: trainable parameter
|
||||
alpha is initialized to 1 by default, higher values = higher-frequency.
|
||||
alpha will be trained along with the rest of your model.
|
||||
"""
|
||||
super(Snake, self).__init__()
|
||||
self.in_features = in_features
|
||||
|
||||
# Initialize alpha
|
||||
self.alpha_logscale = alpha_logscale
|
||||
if self.alpha_logscale: # Log scale alphas initialized to zeros
|
||||
self.alpha = Parameter(torch.zeros(in_features) * alpha)
|
||||
else: # Linear scale alphas initialized to ones
|
||||
self.alpha = Parameter(torch.ones(in_features) * alpha)
|
||||
|
||||
self.alpha.requires_grad = alpha_trainable
|
||||
|
||||
self.no_div_by_zero = 0.000000001
|
||||
|
||||
def forward(self, x):
|
||||
"""
|
||||
Forward pass of the function.
|
||||
Applies the function to the input elementwise.
|
||||
Snake ∶= x + 1/a * sin^2 (xa)
|
||||
"""
|
||||
alpha = self.alpha.unsqueeze(0).unsqueeze(-1) # Line up with x to [B, C, T]
|
||||
if self.alpha_logscale:
|
||||
alpha = torch.exp(alpha)
|
||||
x = x + (1.0 / (alpha + self.no_div_by_zero)) * pow(sin(x * alpha), 2)
|
||||
|
||||
return x
|
||||
|
||||
|
||||
class SnakeBeta(nn.Module):
|
||||
"""
|
||||
A modified Snake function which uses separate parameters for the magnitude of the periodic components
|
||||
Shape:
|
||||
- Input: (B, C, T)
|
||||
- Output: (B, C, T), same shape as the input
|
||||
Parameters:
|
||||
- alpha - trainable parameter that controls frequency
|
||||
- beta - trainable parameter that controls magnitude
|
||||
References:
|
||||
- This activation function is a modified version based on this paper by Liu Ziyin, Tilman Hartwig, Masahito Ueda:
|
||||
https://arxiv.org/abs/2006.08195
|
||||
Examples:
|
||||
>>> a1 = snakebeta(256)
|
||||
>>> x = torch.randn(256)
|
||||
>>> x = a1(x)
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self, in_features, alpha=1.0, alpha_trainable=True, alpha_logscale=False
|
||||
):
|
||||
"""
|
||||
Initialization.
|
||||
INPUT:
|
||||
- in_features: shape of the input
|
||||
- alpha - trainable parameter that controls frequency
|
||||
- beta - trainable parameter that controls magnitude
|
||||
alpha is initialized to 1 by default, higher values = higher-frequency.
|
||||
beta is initialized to 1 by default, higher values = higher-magnitude.
|
||||
alpha will be trained along with the rest of your model.
|
||||
"""
|
||||
super(SnakeBeta, self).__init__()
|
||||
self.in_features = in_features
|
||||
|
||||
# Initialize alpha
|
||||
self.alpha_logscale = alpha_logscale
|
||||
if self.alpha_logscale: # Log scale alphas initialized to zeros
|
||||
self.alpha = Parameter(torch.zeros(in_features) * alpha)
|
||||
self.beta = Parameter(torch.zeros(in_features) * alpha)
|
||||
else: # Linear scale alphas initialized to ones
|
||||
self.alpha = Parameter(torch.ones(in_features) * alpha)
|
||||
self.beta = Parameter(torch.ones(in_features) * alpha)
|
||||
|
||||
self.alpha.requires_grad = alpha_trainable
|
||||
self.beta.requires_grad = alpha_trainable
|
||||
|
||||
self.no_div_by_zero = 0.000000001
|
||||
|
||||
def forward(self, x):
|
||||
"""
|
||||
Forward pass of the function.
|
||||
Applies the function to the input elementwise.
|
||||
SnakeBeta ∶= x + 1/b * sin^2 (xa)
|
||||
"""
|
||||
alpha = self.alpha.unsqueeze(0).unsqueeze(-1) # Line up with x to [B, C, T]
|
||||
beta = self.beta.unsqueeze(0).unsqueeze(-1)
|
||||
if self.alpha_logscale:
|
||||
alpha = torch.exp(alpha)
|
||||
beta = torch.exp(beta)
|
||||
x = x + (1.0 / (beta + self.no_div_by_zero)) * pow(sin(x * alpha), 2)
|
||||
|
||||
return x
|
||||
@@ -0,0 +1,77 @@
|
||||
# Copyright (c) 2024 NVIDIA CORPORATION.
|
||||
# Licensed under the MIT license.
|
||||
|
||||
import torch
|
||||
import torch.nn as nn
|
||||
from alias_free_activation.torch.resample import UpSample1d, DownSample1d
|
||||
|
||||
# load fused CUDA kernel: this enables importing anti_alias_activation_cuda
|
||||
from alias_free_activation.cuda import load
|
||||
|
||||
anti_alias_activation_cuda = load.load()
|
||||
|
||||
|
||||
class FusedAntiAliasActivation(torch.autograd.Function):
|
||||
"""
|
||||
Assumes filter size 12, replication padding on upsampling/downsampling, and logscale alpha/beta parameters as inputs.
|
||||
The hyperparameters are hard-coded in the kernel to maximize speed.
|
||||
NOTE: The fused kenrel is incorrect for Activation1d with different hyperparameters.
|
||||
"""
|
||||
|
||||
@staticmethod
|
||||
def forward(ctx, inputs, up_ftr, down_ftr, alpha, beta):
|
||||
activation_results = anti_alias_activation_cuda.forward(
|
||||
inputs, up_ftr, down_ftr, alpha, beta
|
||||
)
|
||||
|
||||
return activation_results
|
||||
|
||||
@staticmethod
|
||||
def backward(ctx, output_grads):
|
||||
raise NotImplementedError
|
||||
return output_grads, None, None
|
||||
|
||||
|
||||
class Activation1d(nn.Module):
|
||||
def __init__(
|
||||
self,
|
||||
activation,
|
||||
up_ratio: int = 2,
|
||||
down_ratio: int = 2,
|
||||
up_kernel_size: int = 12,
|
||||
down_kernel_size: int = 12,
|
||||
fused: bool = True,
|
||||
):
|
||||
super().__init__()
|
||||
self.up_ratio = up_ratio
|
||||
self.down_ratio = down_ratio
|
||||
self.act = activation
|
||||
self.upsample = UpSample1d(up_ratio, up_kernel_size)
|
||||
self.downsample = DownSample1d(down_ratio, down_kernel_size)
|
||||
|
||||
self.fused = fused # Whether to use fused CUDA kernel or not
|
||||
|
||||
def forward(self, x):
|
||||
if not self.fused:
|
||||
x = self.upsample(x)
|
||||
x = self.act(x)
|
||||
x = self.downsample(x)
|
||||
return x
|
||||
else:
|
||||
if self.act.__class__.__name__ == "Snake":
|
||||
beta = self.act.alpha.data # Snake uses same params for alpha and beta
|
||||
else:
|
||||
beta = (
|
||||
self.act.beta.data
|
||||
) # Snakebeta uses different params for alpha and beta
|
||||
alpha = self.act.alpha.data
|
||||
if (
|
||||
not self.act.alpha_logscale
|
||||
): # Exp baked into cuda kernel, cancel it out with a log
|
||||
alpha = torch.log(alpha)
|
||||
beta = torch.log(beta)
|
||||
|
||||
x = FusedAntiAliasActivation.apply(
|
||||
x, self.upsample.filter, self.downsample.lowpass.filter, alpha, beta
|
||||
)
|
||||
return x
|
||||
@@ -0,0 +1,23 @@
|
||||
/* coding=utf-8
|
||||
* Copyright (c) 2024, NVIDIA CORPORATION. All rights reserved.
|
||||
*
|
||||
* 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.
|
||||
*/
|
||||
|
||||
#include <torch/extension.h>
|
||||
|
||||
extern "C" torch::Tensor fwd_cuda(torch::Tensor const &input, torch::Tensor const &up_filter, torch::Tensor const &down_filter, torch::Tensor const &alpha, torch::Tensor const &beta);
|
||||
|
||||
PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) {
|
||||
m.def("forward", &fwd_cuda, "Anti-Alias Activation forward (CUDA)");
|
||||
}
|
||||
@@ -0,0 +1,246 @@
|
||||
/* coding=utf-8
|
||||
* Copyright (c) 2024, NVIDIA CORPORATION. All rights reserved.
|
||||
*
|
||||
* 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.
|
||||
*/
|
||||
|
||||
#include <ATen/ATen.h>
|
||||
#include <cuda.h>
|
||||
#include <cuda_runtime.h>
|
||||
#include <cuda_fp16.h>
|
||||
#include <cuda_profiler_api.h>
|
||||
#include <ATen/cuda/CUDAContext.h>
|
||||
#include <torch/extension.h>
|
||||
#include "type_shim.h"
|
||||
#include <assert.h>
|
||||
#include <cfloat>
|
||||
#include <limits>
|
||||
#include <stdint.h>
|
||||
#include <c10/macros/Macros.h>
|
||||
|
||||
namespace
|
||||
{
|
||||
// Hard-coded hyperparameters
|
||||
// WARP_SIZE and WARP_BATCH must match the return values batches_per_warp and
|
||||
constexpr int ELEMENTS_PER_LDG_STG = 1; //(WARP_ITERATIONS < 4) ? 1 : 4;
|
||||
constexpr int BUFFER_SIZE = 32;
|
||||
constexpr int FILTER_SIZE = 12;
|
||||
constexpr int HALF_FILTER_SIZE = 6;
|
||||
constexpr int UPSAMPLE_REPLICATION_PAD = 5; // 5 on each side, matching torch impl
|
||||
constexpr int DOWNSAMPLE_REPLICATION_PAD_LEFT = 5; // matching torch impl
|
||||
constexpr int DOWNSAMPLE_REPLICATION_PAD_RIGHT = 6; // matching torch impl
|
||||
|
||||
template <typename input_t, typename output_t, typename acc_t>
|
||||
__global__ void anti_alias_activation_forward(
|
||||
output_t *dst,
|
||||
const input_t *src,
|
||||
const input_t *up_ftr,
|
||||
const input_t *down_ftr,
|
||||
const input_t *alpha,
|
||||
const input_t *beta,
|
||||
int batch_size,
|
||||
int channels,
|
||||
int seq_len)
|
||||
{
|
||||
// Up and downsample filters
|
||||
input_t up_filter[FILTER_SIZE];
|
||||
input_t down_filter[FILTER_SIZE];
|
||||
|
||||
// Load data from global memory including extra indices reserved for replication paddings
|
||||
input_t elements[2 * FILTER_SIZE + 2 * BUFFER_SIZE + 2 * UPSAMPLE_REPLICATION_PAD] = {0};
|
||||
input_t intermediates[2 * FILTER_SIZE + 2 * BUFFER_SIZE + DOWNSAMPLE_REPLICATION_PAD_LEFT + DOWNSAMPLE_REPLICATION_PAD_RIGHT] = {0};
|
||||
|
||||
// Output stores downsampled output before writing to dst
|
||||
output_t output[BUFFER_SIZE];
|
||||
|
||||
// blockDim/threadIdx = (128, 1, 1)
|
||||
// gridDim/blockIdx = (seq_blocks, channels, batches)
|
||||
int block_offset = (blockIdx.x * 128 * BUFFER_SIZE + seq_len * (blockIdx.y + gridDim.y * blockIdx.z));
|
||||
int local_offset = threadIdx.x * BUFFER_SIZE;
|
||||
int seq_offset = blockIdx.x * 128 * BUFFER_SIZE + local_offset;
|
||||
|
||||
// intermediate have double the seq_len
|
||||
int intermediate_local_offset = threadIdx.x * BUFFER_SIZE * 2;
|
||||
int intermediate_seq_offset = blockIdx.x * 128 * BUFFER_SIZE * 2 + intermediate_local_offset;
|
||||
|
||||
// Get values needed for replication padding before moving pointer
|
||||
const input_t *right_most_pntr = src + (seq_len * (blockIdx.y + gridDim.y * blockIdx.z));
|
||||
input_t seq_left_most_value = right_most_pntr[0];
|
||||
input_t seq_right_most_value = right_most_pntr[seq_len - 1];
|
||||
|
||||
// Move src and dst pointers
|
||||
src += block_offset + local_offset;
|
||||
dst += block_offset + local_offset;
|
||||
|
||||
// Alpha and beta values for snake activatons. Applies exp by default
|
||||
alpha = alpha + blockIdx.y;
|
||||
input_t alpha_val = expf(alpha[0]);
|
||||
beta = beta + blockIdx.y;
|
||||
input_t beta_val = expf(beta[0]);
|
||||
|
||||
#pragma unroll
|
||||
for (int it = 0; it < FILTER_SIZE; it += 1)
|
||||
{
|
||||
up_filter[it] = up_ftr[it];
|
||||
down_filter[it] = down_ftr[it];
|
||||
}
|
||||
|
||||
// Apply replication padding for upsampling, matching torch impl
|
||||
#pragma unroll
|
||||
for (int it = -HALF_FILTER_SIZE; it < BUFFER_SIZE + HALF_FILTER_SIZE; it += 1)
|
||||
{
|
||||
int element_index = seq_offset + it; // index for element
|
||||
if ((element_index < 0) && (element_index >= -UPSAMPLE_REPLICATION_PAD))
|
||||
{
|
||||
elements[2 * (HALF_FILTER_SIZE + it)] = 2 * seq_left_most_value;
|
||||
}
|
||||
if ((element_index >= seq_len) && (element_index < seq_len + UPSAMPLE_REPLICATION_PAD))
|
||||
{
|
||||
elements[2 * (HALF_FILTER_SIZE + it)] = 2 * seq_right_most_value;
|
||||
}
|
||||
if ((element_index >= 0) && (element_index < seq_len))
|
||||
{
|
||||
elements[2 * (HALF_FILTER_SIZE + it)] = 2 * src[it];
|
||||
}
|
||||
}
|
||||
|
||||
// Apply upsampling strided convolution and write to intermediates. It reserves DOWNSAMPLE_REPLICATION_PAD_LEFT for replication padding of the downsampilng conv later
|
||||
#pragma unroll
|
||||
for (int it = 0; it < (2 * BUFFER_SIZE + 2 * FILTER_SIZE); it += 1)
|
||||
{
|
||||
input_t acc = 0.0;
|
||||
int element_index = intermediate_seq_offset + it; // index for intermediate
|
||||
#pragma unroll
|
||||
for (int f_idx = 0; f_idx < FILTER_SIZE; f_idx += 1)
|
||||
{
|
||||
if ((element_index + f_idx) >= 0)
|
||||
{
|
||||
acc += up_filter[f_idx] * elements[it + f_idx];
|
||||
}
|
||||
}
|
||||
intermediates[it + DOWNSAMPLE_REPLICATION_PAD_LEFT] = acc;
|
||||
}
|
||||
|
||||
// Apply activation function. It reserves DOWNSAMPLE_REPLICATION_PAD_LEFT and DOWNSAMPLE_REPLICATION_PAD_RIGHT for replication padding of the downsampilng conv later
|
||||
double no_div_by_zero = 0.000000001;
|
||||
#pragma unroll
|
||||
for (int it = 0; it < 2 * BUFFER_SIZE + 2 * FILTER_SIZE; it += 1)
|
||||
{
|
||||
intermediates[it + DOWNSAMPLE_REPLICATION_PAD_LEFT] += (1.0 / (beta_val + no_div_by_zero)) * sinf(intermediates[it + DOWNSAMPLE_REPLICATION_PAD_LEFT] * alpha_val) * sinf(intermediates[it + DOWNSAMPLE_REPLICATION_PAD_LEFT] * alpha_val);
|
||||
}
|
||||
|
||||
// Apply replication padding before downsampling conv from intermediates
|
||||
#pragma unroll
|
||||
for (int it = 0; it < DOWNSAMPLE_REPLICATION_PAD_LEFT; it += 1)
|
||||
{
|
||||
intermediates[it] = intermediates[DOWNSAMPLE_REPLICATION_PAD_LEFT];
|
||||
}
|
||||
#pragma unroll
|
||||
for (int it = DOWNSAMPLE_REPLICATION_PAD_LEFT + 2 * BUFFER_SIZE + 2 * FILTER_SIZE; it < DOWNSAMPLE_REPLICATION_PAD_LEFT + 2 * BUFFER_SIZE + 2 * FILTER_SIZE + DOWNSAMPLE_REPLICATION_PAD_RIGHT; it += 1)
|
||||
{
|
||||
intermediates[it] = intermediates[DOWNSAMPLE_REPLICATION_PAD_LEFT + 2 * BUFFER_SIZE + 2 * FILTER_SIZE - 1];
|
||||
}
|
||||
|
||||
// Apply downsample strided convolution (assuming stride=2) from intermediates
|
||||
#pragma unroll
|
||||
for (int it = 0; it < BUFFER_SIZE; it += 1)
|
||||
{
|
||||
input_t acc = 0.0;
|
||||
#pragma unroll
|
||||
for (int f_idx = 0; f_idx < FILTER_SIZE; f_idx += 1)
|
||||
{
|
||||
// Add constant DOWNSAMPLE_REPLICATION_PAD_RIGHT to match torch implementation
|
||||
acc += down_filter[f_idx] * intermediates[it * 2 + f_idx + DOWNSAMPLE_REPLICATION_PAD_RIGHT];
|
||||
}
|
||||
output[it] = acc;
|
||||
}
|
||||
|
||||
// Write output to dst
|
||||
#pragma unroll
|
||||
for (int it = 0; it < BUFFER_SIZE; it += ELEMENTS_PER_LDG_STG)
|
||||
{
|
||||
int element_index = seq_offset + it;
|
||||
if (element_index < seq_len)
|
||||
{
|
||||
dst[it] = output[it];
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
template <typename input_t, typename output_t, typename acc_t>
|
||||
void dispatch_anti_alias_activation_forward(
|
||||
output_t *dst,
|
||||
const input_t *src,
|
||||
const input_t *up_ftr,
|
||||
const input_t *down_ftr,
|
||||
const input_t *alpha,
|
||||
const input_t *beta,
|
||||
int batch_size,
|
||||
int channels,
|
||||
int seq_len)
|
||||
{
|
||||
if (seq_len == 0)
|
||||
{
|
||||
return;
|
||||
}
|
||||
else
|
||||
{
|
||||
// Use 128 threads per block to maximimize gpu utilization
|
||||
constexpr int threads_per_block = 128;
|
||||
constexpr int seq_len_per_block = 4096;
|
||||
int blocks_per_seq_len = (seq_len + seq_len_per_block - 1) / seq_len_per_block;
|
||||
dim3 blocks(blocks_per_seq_len, channels, batch_size);
|
||||
dim3 threads(threads_per_block, 1, 1);
|
||||
|
||||
anti_alias_activation_forward<input_t, output_t, acc_t>
|
||||
<<<blocks, threads, 0, at::cuda::getCurrentCUDAStream()>>>(dst, src, up_ftr, down_ftr, alpha, beta, batch_size, channels, seq_len);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
extern "C" torch::Tensor fwd_cuda(torch::Tensor const &input, torch::Tensor const &up_filter, torch::Tensor const &down_filter, torch::Tensor const &alpha, torch::Tensor const &beta)
|
||||
{
|
||||
// Input is a 3d tensor with dimensions [batches, channels, seq_len]
|
||||
const int batches = input.size(0);
|
||||
const int channels = input.size(1);
|
||||
const int seq_len = input.size(2);
|
||||
|
||||
// Output
|
||||
auto act_options = input.options().requires_grad(false);
|
||||
|
||||
torch::Tensor anti_alias_activation_results =
|
||||
torch::empty({batches, channels, seq_len}, act_options);
|
||||
|
||||
void *input_ptr = static_cast<void *>(input.data_ptr());
|
||||
void *up_filter_ptr = static_cast<void *>(up_filter.data_ptr());
|
||||
void *down_filter_ptr = static_cast<void *>(down_filter.data_ptr());
|
||||
void *alpha_ptr = static_cast<void *>(alpha.data_ptr());
|
||||
void *beta_ptr = static_cast<void *>(beta.data_ptr());
|
||||
void *anti_alias_activation_results_ptr = static_cast<void *>(anti_alias_activation_results.data_ptr());
|
||||
|
||||
DISPATCH_FLOAT_HALF_AND_BFLOAT(
|
||||
input.scalar_type(),
|
||||
"dispatch anti alias activation_forward",
|
||||
dispatch_anti_alias_activation_forward<scalar_t, scalar_t, float>(
|
||||
reinterpret_cast<scalar_t *>(anti_alias_activation_results_ptr),
|
||||
reinterpret_cast<const scalar_t *>(input_ptr),
|
||||
reinterpret_cast<const scalar_t *>(up_filter_ptr),
|
||||
reinterpret_cast<const scalar_t *>(down_filter_ptr),
|
||||
reinterpret_cast<const scalar_t *>(alpha_ptr),
|
||||
reinterpret_cast<const scalar_t *>(beta_ptr),
|
||||
batches,
|
||||
channels,
|
||||
seq_len););
|
||||
return anti_alias_activation_results;
|
||||
}
|
||||
@@ -0,0 +1,29 @@
|
||||
/* coding=utf-8
|
||||
* Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.
|
||||
*
|
||||
* 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.
|
||||
*/
|
||||
|
||||
/*This code is copied fron NVIDIA apex:
|
||||
* https://github.com/NVIDIA/apex
|
||||
* with minor changes. */
|
||||
|
||||
#ifndef TORCH_CHECK
|
||||
#define TORCH_CHECK AT_CHECK
|
||||
#endif
|
||||
|
||||
#ifdef VERSION_GE_1_3
|
||||
#define DATA_PTR data_ptr
|
||||
#else
|
||||
#define DATA_PTR data
|
||||
#endif
|
||||
@@ -0,0 +1,86 @@
|
||||
# Copyright (c) 2024 NVIDIA CORPORATION.
|
||||
# Licensed under the MIT license.
|
||||
|
||||
import os
|
||||
import pathlib
|
||||
import subprocess
|
||||
|
||||
from torch.utils import cpp_extension
|
||||
|
||||
"""
|
||||
Setting this param to a list has a problem of generating different compilation commands (with diferent order of architectures) and leading to recompilation of fused kernels.
|
||||
Set it to empty stringo avoid recompilation and assign arch flags explicity in extra_cuda_cflags below
|
||||
"""
|
||||
os.environ["TORCH_CUDA_ARCH_LIST"] = ""
|
||||
|
||||
|
||||
def load():
|
||||
# Check if cuda 11 is installed for compute capability 8.0
|
||||
cc_flag = []
|
||||
_, bare_metal_major, _ = _get_cuda_bare_metal_version(cpp_extension.CUDA_HOME)
|
||||
if int(bare_metal_major) >= 11:
|
||||
cc_flag.append("-gencode")
|
||||
cc_flag.append("arch=compute_80,code=sm_80")
|
||||
|
||||
# Build path
|
||||
srcpath = pathlib.Path(__file__).parent.absolute()
|
||||
buildpath = srcpath / "build"
|
||||
_create_build_dir(buildpath)
|
||||
|
||||
# Helper function to build the kernels.
|
||||
def _cpp_extention_load_helper(name, sources, extra_cuda_flags):
|
||||
return cpp_extension.load(
|
||||
name=name,
|
||||
sources=sources,
|
||||
build_directory=buildpath,
|
||||
extra_cflags=[
|
||||
"-O3",
|
||||
],
|
||||
extra_cuda_cflags=[
|
||||
"-O3",
|
||||
"-gencode",
|
||||
"arch=compute_70,code=sm_70",
|
||||
"--use_fast_math",
|
||||
]
|
||||
+ extra_cuda_flags
|
||||
+ cc_flag,
|
||||
verbose=True,
|
||||
)
|
||||
|
||||
extra_cuda_flags = [
|
||||
"-U__CUDA_NO_HALF_OPERATORS__",
|
||||
"-U__CUDA_NO_HALF_CONVERSIONS__",
|
||||
"--expt-relaxed-constexpr",
|
||||
"--expt-extended-lambda",
|
||||
]
|
||||
|
||||
sources = [
|
||||
srcpath / "anti_alias_activation.cpp",
|
||||
srcpath / "anti_alias_activation_cuda.cu",
|
||||
]
|
||||
anti_alias_activation_cuda = _cpp_extention_load_helper(
|
||||
"anti_alias_activation_cuda", sources, extra_cuda_flags
|
||||
)
|
||||
|
||||
return anti_alias_activation_cuda
|
||||
|
||||
|
||||
def _get_cuda_bare_metal_version(cuda_dir):
|
||||
raw_output = subprocess.check_output(
|
||||
[cuda_dir + "/bin/nvcc", "-V"], universal_newlines=True
|
||||
)
|
||||
output = raw_output.split()
|
||||
release_idx = output.index("release") + 1
|
||||
release = output[release_idx].split(".")
|
||||
bare_metal_major = release[0]
|
||||
bare_metal_minor = release[1][0]
|
||||
|
||||
return raw_output, bare_metal_major, bare_metal_minor
|
||||
|
||||
|
||||
def _create_build_dir(buildpath):
|
||||
try:
|
||||
os.mkdir(buildpath)
|
||||
except OSError:
|
||||
if not os.path.isdir(buildpath):
|
||||
print(f"Creation of the build directory {buildpath} failed")
|
||||
@@ -0,0 +1,92 @@
|
||||
/* coding=utf-8
|
||||
* Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.
|
||||
*
|
||||
* 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.
|
||||
*/
|
||||
|
||||
#include <ATen/ATen.h>
|
||||
#include "compat.h"
|
||||
|
||||
#define DISPATCH_FLOAT_HALF_AND_BFLOAT(TYPE, NAME, ...) \
|
||||
switch (TYPE) \
|
||||
{ \
|
||||
case at::ScalarType::Float: \
|
||||
{ \
|
||||
using scalar_t = float; \
|
||||
__VA_ARGS__; \
|
||||
break; \
|
||||
} \
|
||||
case at::ScalarType::Half: \
|
||||
{ \
|
||||
using scalar_t = at::Half; \
|
||||
__VA_ARGS__; \
|
||||
break; \
|
||||
} \
|
||||
case at::ScalarType::BFloat16: \
|
||||
{ \
|
||||
using scalar_t = at::BFloat16; \
|
||||
__VA_ARGS__; \
|
||||
break; \
|
||||
} \
|
||||
default: \
|
||||
AT_ERROR(#NAME, " not implemented for '", toString(TYPE), "'"); \
|
||||
}
|
||||
|
||||
#define DISPATCH_FLOAT_HALF_AND_BFLOAT_INOUT_TYPES(TYPEIN, TYPEOUT, NAME, ...) \
|
||||
switch (TYPEIN) \
|
||||
{ \
|
||||
case at::ScalarType::Float: \
|
||||
{ \
|
||||
using scalar_t_in = float; \
|
||||
switch (TYPEOUT) \
|
||||
{ \
|
||||
case at::ScalarType::Float: \
|
||||
{ \
|
||||
using scalar_t_out = float; \
|
||||
__VA_ARGS__; \
|
||||
break; \
|
||||
} \
|
||||
case at::ScalarType::Half: \
|
||||
{ \
|
||||
using scalar_t_out = at::Half; \
|
||||
__VA_ARGS__; \
|
||||
break; \
|
||||
} \
|
||||
case at::ScalarType::BFloat16: \
|
||||
{ \
|
||||
using scalar_t_out = at::BFloat16; \
|
||||
__VA_ARGS__; \
|
||||
break; \
|
||||
} \
|
||||
default: \
|
||||
AT_ERROR(#NAME, " not implemented for '", toString(TYPEOUT), "'"); \
|
||||
} \
|
||||
break; \
|
||||
} \
|
||||
case at::ScalarType::Half: \
|
||||
{ \
|
||||
using scalar_t_in = at::Half; \
|
||||
using scalar_t_out = at::Half; \
|
||||
__VA_ARGS__; \
|
||||
break; \
|
||||
} \
|
||||
case at::ScalarType::BFloat16: \
|
||||
{ \
|
||||
using scalar_t_in = at::BFloat16; \
|
||||
using scalar_t_out = at::BFloat16; \
|
||||
__VA_ARGS__; \
|
||||
break; \
|
||||
} \
|
||||
default: \
|
||||
AT_ERROR(#NAME, " not implemented for '", toString(TYPEIN), "'"); \
|
||||
}
|
||||
@@ -0,0 +1,6 @@
|
||||
# Adapted from https://github.com/junjun3518/alias-free-torch under the Apache License 2.0
|
||||
# LICENSE is in incl_licenses directory.
|
||||
|
||||
from .filter import *
|
||||
from .resample import *
|
||||
from .act import *
|
||||
@@ -0,0 +1,32 @@
|
||||
# Adapted from https://github.com/junjun3518/alias-free-torch under the Apache License 2.0
|
||||
# LICENSE is in incl_licenses directory.
|
||||
|
||||
import torch.nn as nn
|
||||
|
||||
from selva_core.ext.bigvgan_v2.alias_free_activation.torch.resample import (DownSample1d, UpSample1d)
|
||||
|
||||
|
||||
class Activation1d(nn.Module):
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
activation,
|
||||
up_ratio: int = 2,
|
||||
down_ratio: int = 2,
|
||||
up_kernel_size: int = 12,
|
||||
down_kernel_size: int = 12,
|
||||
):
|
||||
super().__init__()
|
||||
self.up_ratio = up_ratio
|
||||
self.down_ratio = down_ratio
|
||||
self.act = activation
|
||||
self.upsample = UpSample1d(up_ratio, up_kernel_size)
|
||||
self.downsample = DownSample1d(down_ratio, down_kernel_size)
|
||||
|
||||
# x: [B,C,T]
|
||||
def forward(self, x):
|
||||
x = self.upsample(x)
|
||||
x = self.act(x)
|
||||
x = self.downsample(x)
|
||||
|
||||
return x
|
||||
@@ -0,0 +1,101 @@
|
||||
# Adapted from https://github.com/junjun3518/alias-free-torch under the Apache License 2.0
|
||||
# LICENSE is in incl_licenses directory.
|
||||
|
||||
import torch
|
||||
import torch.nn as nn
|
||||
import torch.nn.functional as F
|
||||
import math
|
||||
|
||||
if "sinc" in dir(torch):
|
||||
sinc = torch.sinc
|
||||
else:
|
||||
# This code is adopted from adefossez's julius.core.sinc under the MIT License
|
||||
# https://adefossez.github.io/julius/julius/core.html
|
||||
# LICENSE is in incl_licenses directory.
|
||||
def sinc(x: torch.Tensor):
|
||||
"""
|
||||
Implementation of sinc, i.e. sin(pi * x) / (pi * x)
|
||||
__Warning__: Different to julius.sinc, the input is multiplied by `pi`!
|
||||
"""
|
||||
return torch.where(
|
||||
x == 0,
|
||||
torch.tensor(1.0, device=x.device, dtype=x.dtype),
|
||||
torch.sin(math.pi * x) / math.pi / x,
|
||||
)
|
||||
|
||||
|
||||
# This code is adopted from adefossez's julius.lowpass.LowPassFilters under the MIT License
|
||||
# https://adefossez.github.io/julius/julius/lowpass.html
|
||||
# LICENSE is in incl_licenses directory.
|
||||
def kaiser_sinc_filter1d(
|
||||
cutoff, half_width, kernel_size
|
||||
): # return filter [1,1,kernel_size]
|
||||
even = kernel_size % 2 == 0
|
||||
half_size = kernel_size // 2
|
||||
|
||||
# For kaiser window
|
||||
delta_f = 4 * half_width
|
||||
A = 2.285 * (half_size - 1) * math.pi * delta_f + 7.95
|
||||
if A > 50.0:
|
||||
beta = 0.1102 * (A - 8.7)
|
||||
elif A >= 21.0:
|
||||
beta = 0.5842 * (A - 21) ** 0.4 + 0.07886 * (A - 21.0)
|
||||
else:
|
||||
beta = 0.0
|
||||
window = torch.kaiser_window(kernel_size, beta=beta, periodic=False)
|
||||
|
||||
# ratio = 0.5/cutoff -> 2 * cutoff = 1 / ratio
|
||||
if even:
|
||||
time = torch.arange(-half_size, half_size) + 0.5
|
||||
else:
|
||||
time = torch.arange(kernel_size) - half_size
|
||||
if cutoff == 0:
|
||||
filter_ = torch.zeros_like(time)
|
||||
else:
|
||||
filter_ = 2 * cutoff * window * sinc(2 * cutoff * time)
|
||||
"""
|
||||
Normalize filter to have sum = 1, otherwise we will have a small leakage of the constant component in the input signal.
|
||||
"""
|
||||
filter_ /= filter_.sum()
|
||||
filter = filter_.view(1, 1, kernel_size)
|
||||
|
||||
return filter
|
||||
|
||||
|
||||
class LowPassFilter1d(nn.Module):
|
||||
def __init__(
|
||||
self,
|
||||
cutoff=0.5,
|
||||
half_width=0.6,
|
||||
stride: int = 1,
|
||||
padding: bool = True,
|
||||
padding_mode: str = "replicate",
|
||||
kernel_size: int = 12,
|
||||
):
|
||||
"""
|
||||
kernel_size should be even number for stylegan3 setup, in this implementation, odd number is also possible.
|
||||
"""
|
||||
super().__init__()
|
||||
if cutoff < -0.0:
|
||||
raise ValueError("Minimum cutoff must be larger than zero.")
|
||||
if cutoff > 0.5:
|
||||
raise ValueError("A cutoff above 0.5 does not make sense.")
|
||||
self.kernel_size = kernel_size
|
||||
self.even = kernel_size % 2 == 0
|
||||
self.pad_left = kernel_size // 2 - int(self.even)
|
||||
self.pad_right = kernel_size // 2
|
||||
self.stride = stride
|
||||
self.padding = padding
|
||||
self.padding_mode = padding_mode
|
||||
filter = kaiser_sinc_filter1d(cutoff, half_width, kernel_size)
|
||||
self.register_buffer("filter", filter)
|
||||
|
||||
# Input [B, C, T]
|
||||
def forward(self, x):
|
||||
_, C, _ = x.shape
|
||||
|
||||
if self.padding:
|
||||
x = F.pad(x, (self.pad_left, self.pad_right), mode=self.padding_mode)
|
||||
out = F.conv1d(x, self.filter.expand(C, -1, -1), stride=self.stride, groups=C)
|
||||
|
||||
return out
|
||||
@@ -0,0 +1,54 @@
|
||||
# Adapted from https://github.com/junjun3518/alias-free-torch under the Apache License 2.0
|
||||
# LICENSE is in incl_licenses directory.
|
||||
|
||||
import torch.nn as nn
|
||||
from torch.nn import functional as F
|
||||
|
||||
from selva_core.ext.bigvgan_v2.alias_free_activation.torch.filter import (LowPassFilter1d,
|
||||
kaiser_sinc_filter1d)
|
||||
|
||||
|
||||
class UpSample1d(nn.Module):
|
||||
|
||||
def __init__(self, ratio=2, kernel_size=None):
|
||||
super().__init__()
|
||||
self.ratio = ratio
|
||||
self.kernel_size = (int(6 * ratio // 2) * 2 if kernel_size is None else kernel_size)
|
||||
self.stride = ratio
|
||||
self.pad = self.kernel_size // ratio - 1
|
||||
self.pad_left = self.pad * self.stride + (self.kernel_size - self.stride) // 2
|
||||
self.pad_right = (self.pad * self.stride + (self.kernel_size - self.stride + 1) // 2)
|
||||
filter = kaiser_sinc_filter1d(cutoff=0.5 / ratio,
|
||||
half_width=0.6 / ratio,
|
||||
kernel_size=self.kernel_size)
|
||||
self.register_buffer("filter", filter)
|
||||
|
||||
# x: [B, C, T]
|
||||
def forward(self, x):
|
||||
_, C, _ = x.shape
|
||||
|
||||
x = F.pad(x, (self.pad, self.pad), mode="replicate")
|
||||
x = self.ratio * F.conv_transpose1d(
|
||||
x, self.filter.expand(C, -1, -1), stride=self.stride, groups=C)
|
||||
x = x[..., self.pad_left:-self.pad_right]
|
||||
|
||||
return x
|
||||
|
||||
|
||||
class DownSample1d(nn.Module):
|
||||
|
||||
def __init__(self, ratio=2, kernel_size=None):
|
||||
super().__init__()
|
||||
self.ratio = ratio
|
||||
self.kernel_size = (int(6 * ratio // 2) * 2 if kernel_size is None else kernel_size)
|
||||
self.lowpass = LowPassFilter1d(
|
||||
cutoff=0.5 / ratio,
|
||||
half_width=0.6 / ratio,
|
||||
stride=ratio,
|
||||
kernel_size=self.kernel_size,
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
xx = self.lowpass(x)
|
||||
|
||||
return xx
|
||||
@@ -0,0 +1,439 @@
|
||||
# Copyright (c) 2024 NVIDIA CORPORATION.
|
||||
# Licensed under the MIT license.
|
||||
|
||||
# Adapted from https://github.com/jik876/hifi-gan under the MIT license.
|
||||
# LICENSE is in incl_licenses directory.
|
||||
|
||||
import json
|
||||
import os
|
||||
from pathlib import Path
|
||||
from typing import Dict, Optional, Union
|
||||
|
||||
import torch
|
||||
import torch.nn as nn
|
||||
from huggingface_hub import PyTorchModelHubMixin, hf_hub_download
|
||||
from torch.nn import Conv1d, ConvTranspose1d
|
||||
from torch.nn.utils.parametrizations import weight_norm
|
||||
from torch.nn.utils.parametrize import remove_parametrizations
|
||||
|
||||
from selva_core.ext.bigvgan_v2 import activations
|
||||
from selva_core.ext.bigvgan_v2.alias_free_activation.torch.act import \
|
||||
Activation1d as TorchActivation1d
|
||||
from selva_core.ext.bigvgan_v2.env import AttrDict
|
||||
from selva_core.ext.bigvgan_v2.utils import get_padding, init_weights
|
||||
|
||||
|
||||
def load_hparams_from_json(path) -> AttrDict:
|
||||
with open(path) as f:
|
||||
data = f.read()
|
||||
return AttrDict(json.loads(data))
|
||||
|
||||
|
||||
class AMPBlock1(torch.nn.Module):
|
||||
"""
|
||||
AMPBlock applies Snake / SnakeBeta activation functions with trainable parameters that control periodicity, defined for each layer.
|
||||
AMPBlock1 has additional self.convs2 that contains additional Conv1d layers with a fixed dilation=1 followed by each layer in self.convs1
|
||||
|
||||
Args:
|
||||
h (AttrDict): Hyperparameters.
|
||||
channels (int): Number of convolution channels.
|
||||
kernel_size (int): Size of the convolution kernel. Default is 3.
|
||||
dilation (tuple): Dilation rates for the convolutions. Each dilation layer has two convolutions. Default is (1, 3, 5).
|
||||
activation (str): Activation function type. Should be either 'snake' or 'snakebeta'. Default is None.
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
h: AttrDict,
|
||||
channels: int,
|
||||
kernel_size: int = 3,
|
||||
dilation: tuple = (1, 3, 5),
|
||||
activation: str = None,
|
||||
):
|
||||
super().__init__()
|
||||
|
||||
self.h = h
|
||||
|
||||
self.convs1 = nn.ModuleList([
|
||||
weight_norm(
|
||||
Conv1d(
|
||||
channels,
|
||||
channels,
|
||||
kernel_size,
|
||||
stride=1,
|
||||
dilation=d,
|
||||
padding=get_padding(kernel_size, d),
|
||||
)) for d in dilation
|
||||
])
|
||||
self.convs1.apply(init_weights)
|
||||
|
||||
self.convs2 = nn.ModuleList([
|
||||
weight_norm(
|
||||
Conv1d(
|
||||
channels,
|
||||
channels,
|
||||
kernel_size,
|
||||
stride=1,
|
||||
dilation=1,
|
||||
padding=get_padding(kernel_size, 1),
|
||||
)) for _ in range(len(dilation))
|
||||
])
|
||||
self.convs2.apply(init_weights)
|
||||
|
||||
self.num_layers = len(self.convs1) + len(self.convs2) # Total number of conv layers
|
||||
|
||||
# Select which Activation1d, lazy-load cuda version to ensure backward compatibility
|
||||
if self.h.get("use_cuda_kernel", False):
|
||||
from alias_free_activation.cuda.activation1d import \
|
||||
Activation1d as CudaActivation1d
|
||||
|
||||
Activation1d = CudaActivation1d
|
||||
else:
|
||||
Activation1d = TorchActivation1d
|
||||
|
||||
# Activation functions
|
||||
if activation == "snake":
|
||||
self.activations = nn.ModuleList([
|
||||
Activation1d(
|
||||
activation=activations.Snake(channels, alpha_logscale=h.snake_logscale))
|
||||
for _ in range(self.num_layers)
|
||||
])
|
||||
elif activation == "snakebeta":
|
||||
self.activations = nn.ModuleList([
|
||||
Activation1d(
|
||||
activation=activations.SnakeBeta(channels, alpha_logscale=h.snake_logscale))
|
||||
for _ in range(self.num_layers)
|
||||
])
|
||||
else:
|
||||
raise NotImplementedError(
|
||||
"activation incorrectly specified. check the config file and look for 'activation'."
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
acts1, acts2 = self.activations[::2], self.activations[1::2]
|
||||
for c1, c2, a1, a2 in zip(self.convs1, self.convs2, acts1, acts2):
|
||||
xt = a1(x)
|
||||
xt = c1(xt)
|
||||
xt = a2(xt)
|
||||
xt = c2(xt)
|
||||
x = xt + x
|
||||
|
||||
return x
|
||||
|
||||
def remove_weight_norm(self):
|
||||
for l in self.convs1:
|
||||
remove_parametrizations(l, 'weight')
|
||||
for l in self.convs2:
|
||||
remove_parametrizations(l, 'weight')
|
||||
|
||||
|
||||
class AMPBlock2(torch.nn.Module):
|
||||
"""
|
||||
AMPBlock applies Snake / SnakeBeta activation functions with trainable parameters that control periodicity, defined for each layer.
|
||||
Unlike AMPBlock1, AMPBlock2 does not contain extra Conv1d layers with fixed dilation=1
|
||||
|
||||
Args:
|
||||
h (AttrDict): Hyperparameters.
|
||||
channels (int): Number of convolution channels.
|
||||
kernel_size (int): Size of the convolution kernel. Default is 3.
|
||||
dilation (tuple): Dilation rates for the convolutions. Each dilation layer has two convolutions. Default is (1, 3, 5).
|
||||
activation (str): Activation function type. Should be either 'snake' or 'snakebeta'. Default is None.
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
h: AttrDict,
|
||||
channels: int,
|
||||
kernel_size: int = 3,
|
||||
dilation: tuple = (1, 3, 5),
|
||||
activation: str = None,
|
||||
):
|
||||
super().__init__()
|
||||
|
||||
self.h = h
|
||||
|
||||
self.convs = nn.ModuleList([
|
||||
weight_norm(
|
||||
Conv1d(
|
||||
channels,
|
||||
channels,
|
||||
kernel_size,
|
||||
stride=1,
|
||||
dilation=d,
|
||||
padding=get_padding(kernel_size, d),
|
||||
)) for d in dilation
|
||||
])
|
||||
self.convs.apply(init_weights)
|
||||
|
||||
self.num_layers = len(self.convs) # Total number of conv layers
|
||||
|
||||
# Select which Activation1d, lazy-load cuda version to ensure backward compatibility
|
||||
if self.h.get("use_cuda_kernel", False):
|
||||
from alias_free_activation.cuda.activation1d import \
|
||||
Activation1d as CudaActivation1d
|
||||
|
||||
Activation1d = CudaActivation1d
|
||||
else:
|
||||
Activation1d = TorchActivation1d
|
||||
|
||||
# Activation functions
|
||||
if activation == "snake":
|
||||
self.activations = nn.ModuleList([
|
||||
Activation1d(
|
||||
activation=activations.Snake(channels, alpha_logscale=h.snake_logscale))
|
||||
for _ in range(self.num_layers)
|
||||
])
|
||||
elif activation == "snakebeta":
|
||||
self.activations = nn.ModuleList([
|
||||
Activation1d(
|
||||
activation=activations.SnakeBeta(channels, alpha_logscale=h.snake_logscale))
|
||||
for _ in range(self.num_layers)
|
||||
])
|
||||
else:
|
||||
raise NotImplementedError(
|
||||
"activation incorrectly specified. check the config file and look for 'activation'."
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
for c, a in zip(self.convs, self.activations):
|
||||
xt = a(x)
|
||||
xt = c(xt)
|
||||
x = xt + x
|
||||
return x
|
||||
|
||||
def remove_weight_norm(self):
|
||||
for l in self.convs:
|
||||
remove_weight_norm(l)
|
||||
|
||||
|
||||
class BigVGAN(
|
||||
torch.nn.Module,
|
||||
PyTorchModelHubMixin,
|
||||
library_name="bigvgan",
|
||||
repo_url="https://github.com/NVIDIA/BigVGAN",
|
||||
docs_url="https://github.com/NVIDIA/BigVGAN/blob/main/README.md",
|
||||
pipeline_tag="audio-to-audio",
|
||||
license="mit",
|
||||
tags=["neural-vocoder", "audio-generation", "arxiv:2206.04658"],
|
||||
):
|
||||
"""
|
||||
BigVGAN is a neural vocoder model that applies anti-aliased periodic activation for residual blocks (resblocks).
|
||||
New in BigVGAN-v2: it can optionally use optimized CUDA kernels for AMP (anti-aliased multi-periodicity) blocks.
|
||||
|
||||
Args:
|
||||
h (AttrDict): Hyperparameters.
|
||||
use_cuda_kernel (bool): If set to True, loads optimized CUDA kernels for AMP. This should be used for inference only, as training is not supported with CUDA kernels.
|
||||
|
||||
Note:
|
||||
- The `use_cuda_kernel` parameter should be used for inference only, as training with CUDA kernels is not supported.
|
||||
- Ensure that the activation function is correctly specified in the hyperparameters (h.activation).
|
||||
"""
|
||||
|
||||
def __init__(self, h: AttrDict, use_cuda_kernel: bool = False):
|
||||
super().__init__()
|
||||
self.h = h
|
||||
self.h["use_cuda_kernel"] = use_cuda_kernel
|
||||
|
||||
# Select which Activation1d, lazy-load cuda version to ensure backward compatibility
|
||||
if self.h.get("use_cuda_kernel", False):
|
||||
from alias_free_activation.cuda.activation1d import \
|
||||
Activation1d as CudaActivation1d
|
||||
|
||||
Activation1d = CudaActivation1d
|
||||
else:
|
||||
Activation1d = TorchActivation1d
|
||||
|
||||
self.num_kernels = len(h.resblock_kernel_sizes)
|
||||
self.num_upsamples = len(h.upsample_rates)
|
||||
|
||||
# Pre-conv
|
||||
self.conv_pre = weight_norm(Conv1d(h.num_mels, h.upsample_initial_channel, 7, 1, padding=3))
|
||||
|
||||
# Define which AMPBlock to use. BigVGAN uses AMPBlock1 as default
|
||||
if h.resblock == "1":
|
||||
resblock_class = AMPBlock1
|
||||
elif h.resblock == "2":
|
||||
resblock_class = AMPBlock2
|
||||
else:
|
||||
raise ValueError(
|
||||
f"Incorrect resblock class specified in hyperparameters. Got {h.resblock}")
|
||||
|
||||
# Transposed conv-based upsamplers. does not apply anti-aliasing
|
||||
self.ups = nn.ModuleList()
|
||||
for i, (u, k) in enumerate(zip(h.upsample_rates, h.upsample_kernel_sizes)):
|
||||
self.ups.append(
|
||||
nn.ModuleList([
|
||||
weight_norm(
|
||||
ConvTranspose1d(
|
||||
h.upsample_initial_channel // (2**i),
|
||||
h.upsample_initial_channel // (2**(i + 1)),
|
||||
k,
|
||||
u,
|
||||
padding=(k - u) // 2,
|
||||
))
|
||||
]))
|
||||
|
||||
# Residual blocks using anti-aliased multi-periodicity composition modules (AMP)
|
||||
self.resblocks = nn.ModuleList()
|
||||
for i in range(len(self.ups)):
|
||||
ch = h.upsample_initial_channel // (2**(i + 1))
|
||||
for j, (k, d) in enumerate(zip(h.resblock_kernel_sizes, h.resblock_dilation_sizes)):
|
||||
self.resblocks.append(resblock_class(h, ch, k, d, activation=h.activation))
|
||||
|
||||
# Post-conv
|
||||
activation_post = (activations.Snake(ch, alpha_logscale=h.snake_logscale)
|
||||
if h.activation == "snake" else
|
||||
(activations.SnakeBeta(ch, alpha_logscale=h.snake_logscale)
|
||||
if h.activation == "snakebeta" else None))
|
||||
if activation_post is None:
|
||||
raise NotImplementedError(
|
||||
"activation incorrectly specified. check the config file and look for 'activation'."
|
||||
)
|
||||
|
||||
self.activation_post = Activation1d(activation=activation_post)
|
||||
|
||||
# Whether to use bias for the final conv_post. Default to True for backward compatibility
|
||||
self.use_bias_at_final = h.get("use_bias_at_final", True)
|
||||
self.conv_post = weight_norm(Conv1d(ch, 1, 7, 1, padding=3, bias=self.use_bias_at_final))
|
||||
|
||||
# Weight initialization
|
||||
for i in range(len(self.ups)):
|
||||
self.ups[i].apply(init_weights)
|
||||
self.conv_post.apply(init_weights)
|
||||
|
||||
# Final tanh activation. Defaults to True for backward compatibility
|
||||
self.use_tanh_at_final = h.get("use_tanh_at_final", True)
|
||||
|
||||
def forward(self, x):
|
||||
# Pre-conv
|
||||
x = self.conv_pre(x)
|
||||
|
||||
for i in range(self.num_upsamples):
|
||||
# Upsampling
|
||||
for i_up in range(len(self.ups[i])):
|
||||
x = self.ups[i][i_up](x)
|
||||
# AMP blocks
|
||||
xs = None
|
||||
for j in range(self.num_kernels):
|
||||
if xs is None:
|
||||
xs = self.resblocks[i * self.num_kernels + j](x)
|
||||
else:
|
||||
xs += self.resblocks[i * self.num_kernels + j](x)
|
||||
x = xs / self.num_kernels
|
||||
|
||||
# Post-conv
|
||||
x = self.activation_post(x)
|
||||
x = self.conv_post(x)
|
||||
# Final tanh activation
|
||||
if self.use_tanh_at_final:
|
||||
x = torch.tanh(x)
|
||||
else:
|
||||
x = torch.clamp(x, min=-1.0, max=1.0) # Bound the output to [-1, 1]
|
||||
|
||||
return x
|
||||
|
||||
def remove_weight_norm(self):
|
||||
try:
|
||||
print("Removing weight norm...")
|
||||
for l in self.ups:
|
||||
for l_i in l:
|
||||
remove_parametrizations(l_i, 'weight')
|
||||
for l in self.resblocks:
|
||||
l.remove_weight_norm()
|
||||
remove_parametrizations(self.conv_pre, 'weight')
|
||||
remove_parametrizations(self.conv_post, 'weight')
|
||||
except ValueError:
|
||||
print("[INFO] Model already removed weight norm. Skipping!")
|
||||
pass
|
||||
|
||||
# Additional methods for huggingface_hub support
|
||||
def _save_pretrained(self, save_directory: Path) -> None:
|
||||
"""Save weights and config.json from a Pytorch model to a local directory."""
|
||||
|
||||
model_path = save_directory / "bigvgan_generator.pt"
|
||||
torch.save({"generator": self.state_dict()}, model_path)
|
||||
|
||||
config_path = save_directory / "config.json"
|
||||
with open(config_path, "w") as config_file:
|
||||
json.dump(self.h, config_file, indent=4)
|
||||
|
||||
@classmethod
|
||||
def _from_pretrained(
|
||||
cls,
|
||||
*,
|
||||
model_id: str,
|
||||
revision: str,
|
||||
cache_dir: str,
|
||||
force_download: bool,
|
||||
proxies: Optional[Dict],
|
||||
resume_download: bool,
|
||||
local_files_only: bool,
|
||||
token: Union[str, bool, None],
|
||||
map_location: str = "cpu", # Additional argument
|
||||
strict: bool = False, # Additional argument
|
||||
use_cuda_kernel: bool = False,
|
||||
**model_kwargs,
|
||||
):
|
||||
"""Load Pytorch pretrained weights and return the loaded model."""
|
||||
|
||||
# Download and load hyperparameters (h) used by BigVGAN
|
||||
if os.path.isdir(model_id):
|
||||
print("Loading config.json from local directory")
|
||||
config_file = os.path.join(model_id, "config.json")
|
||||
else:
|
||||
config_file = hf_hub_download(
|
||||
repo_id=model_id,
|
||||
filename="config.json",
|
||||
revision=revision,
|
||||
cache_dir=cache_dir,
|
||||
force_download=force_download,
|
||||
proxies=proxies,
|
||||
resume_download=resume_download,
|
||||
token=token,
|
||||
local_files_only=local_files_only,
|
||||
)
|
||||
h = load_hparams_from_json(config_file)
|
||||
|
||||
# instantiate BigVGAN using h
|
||||
if use_cuda_kernel:
|
||||
print(
|
||||
f"[WARNING] You have specified use_cuda_kernel=True during BigVGAN.from_pretrained(). Only inference is supported (training is not implemented)!"
|
||||
)
|
||||
print(
|
||||
f"[WARNING] You need nvcc and ninja installed in your system that matches your PyTorch build is using to build the kernel. If not, the model will fail to initialize or generate incorrect waveform!"
|
||||
)
|
||||
print(
|
||||
f"[WARNING] For detail, see the official GitHub repository: https://github.com/NVIDIA/BigVGAN?tab=readme-ov-file#using-custom-cuda-kernel-for-synthesis"
|
||||
)
|
||||
model = cls(h, use_cuda_kernel=use_cuda_kernel)
|
||||
|
||||
# Download and load pretrained generator weight
|
||||
if os.path.isdir(model_id):
|
||||
print("Loading weights from local directory")
|
||||
model_file = os.path.join(model_id, "bigvgan_generator.pt")
|
||||
else:
|
||||
print(f"Loading weights from {model_id}")
|
||||
model_file = hf_hub_download(
|
||||
repo_id=model_id,
|
||||
filename="bigvgan_generator.pt",
|
||||
revision=revision,
|
||||
cache_dir=cache_dir,
|
||||
force_download=force_download,
|
||||
proxies=proxies,
|
||||
resume_download=resume_download,
|
||||
token=token,
|
||||
local_files_only=local_files_only,
|
||||
)
|
||||
|
||||
checkpoint_dict = torch.load(model_file, map_location=map_location, weights_only=True)
|
||||
|
||||
try:
|
||||
model.load_state_dict(checkpoint_dict["generator"])
|
||||
except RuntimeError:
|
||||
print(
|
||||
f"[INFO] the pretrained checkpoint does not contain weight norm. Loading the checkpoint after removing weight norm!"
|
||||
)
|
||||
model.remove_weight_norm()
|
||||
model.load_state_dict(checkpoint_dict["generator"])
|
||||
|
||||
return model
|
||||
@@ -0,0 +1,18 @@
|
||||
# Adapted from https://github.com/jik876/hifi-gan under the MIT license.
|
||||
# LICENSE is in incl_licenses directory.
|
||||
|
||||
import os
|
||||
import shutil
|
||||
|
||||
|
||||
class AttrDict(dict):
|
||||
def __init__(self, *args, **kwargs):
|
||||
super(AttrDict, self).__init__(*args, **kwargs)
|
||||
self.__dict__ = self
|
||||
|
||||
|
||||
def build_env(config, config_name, path):
|
||||
t_path = os.path.join(path, config_name)
|
||||
if config != t_path:
|
||||
os.makedirs(path, exist_ok=True)
|
||||
shutil.copyfile(config, os.path.join(path, config_name))
|
||||
@@ -0,0 +1,21 @@
|
||||
MIT License
|
||||
|
||||
Copyright (c) 2020 Jungil Kong
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
of this software and associated documentation files (the "Software"), to deal
|
||||
in the Software without restriction, including without limitation the rights
|
||||
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
||||
copies of the Software, and to permit persons to whom the Software is
|
||||
furnished to do so, subject to the following conditions:
|
||||
|
||||
The above copyright notice and this permission notice shall be included in all
|
||||
copies or substantial portions of the Software.
|
||||
|
||||
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
||||
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
||||
SOFTWARE.
|
||||
@@ -0,0 +1,21 @@
|
||||
MIT License
|
||||
|
||||
Copyright (c) 2020 Edward Dixon
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
of this software and associated documentation files (the "Software"), to deal
|
||||
in the Software without restriction, including without limitation the rights
|
||||
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
||||
copies of the Software, and to permit persons to whom the Software is
|
||||
furnished to do so, subject to the following conditions:
|
||||
|
||||
The above copyright notice and this permission notice shall be included in all
|
||||
copies or substantial portions of the Software.
|
||||
|
||||
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
||||
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
||||
SOFTWARE.
|
||||
@@ -0,0 +1,201 @@
|
||||
Apache License
|
||||
Version 2.0, January 2004
|
||||
http://www.apache.org/licenses/
|
||||
|
||||
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
||||
|
||||
1. Definitions.
|
||||
|
||||
"License" shall mean the terms and conditions for use, reproduction,
|
||||
and distribution as defined by Sections 1 through 9 of this document.
|
||||
|
||||
"Licensor" shall mean the copyright owner or entity authorized by
|
||||
the copyright owner that is granting the License.
|
||||
|
||||
"Legal Entity" shall mean the union of the acting entity and all
|
||||
other entities that control, are controlled by, or are under common
|
||||
control with that entity. For the purposes of this definition,
|
||||
"control" means (i) the power, direct or indirect, to cause the
|
||||
direction or management of such entity, whether by contract or
|
||||
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
||||
outstanding shares, or (iii) beneficial ownership of such entity.
|
||||
|
||||
"You" (or "Your") shall mean an individual or Legal Entity
|
||||
exercising permissions granted by this License.
|
||||
|
||||
"Source" form shall mean the preferred form for making modifications,
|
||||
including but not limited to software source code, documentation
|
||||
source, and configuration files.
|
||||
|
||||
"Object" form shall mean any form resulting from mechanical
|
||||
transformation or translation of a Source form, including but
|
||||
not limited to compiled object code, generated documentation,
|
||||
and conversions to other media types.
|
||||
|
||||
"Work" shall mean the work of authorship, whether in Source or
|
||||
Object form, made available under the License, as indicated by a
|
||||
copyright notice that is included in or attached to the work
|
||||
(an example is provided in the Appendix below).
|
||||
|
||||
"Derivative Works" shall mean any work, whether in Source or Object
|
||||
form, that is based on (or derived from) the Work and for which the
|
||||
editorial revisions, annotations, elaborations, or other modifications
|
||||
represent, as a whole, an original work of authorship. For the purposes
|
||||
of this License, Derivative Works shall not include works that remain
|
||||
separable from, or merely link (or bind by name) to the interfaces of,
|
||||
the Work and Derivative Works thereof.
|
||||
|
||||
"Contribution" shall mean any work of authorship, including
|
||||
the original version of the Work and any modifications or additions
|
||||
to that Work or Derivative Works thereof, that is intentionally
|
||||
submitted to Licensor for inclusion in the Work by the copyright owner
|
||||
or by an individual or Legal Entity authorized to submit on behalf of
|
||||
the copyright owner. For the purposes of this definition, "submitted"
|
||||
means any form of electronic, verbal, or written communication sent
|
||||
to the Licensor or its representatives, including but not limited to
|
||||
communication on electronic mailing lists, source code control systems,
|
||||
and issue tracking systems that are managed by, or on behalf of, the
|
||||
Licensor for the purpose of discussing and improving the Work, but
|
||||
excluding communication that is conspicuously marked or otherwise
|
||||
designated in writing by the copyright owner as "Not a Contribution."
|
||||
|
||||
"Contributor" shall mean Licensor and any individual or Legal Entity
|
||||
on behalf of whom a Contribution has been received by Licensor and
|
||||
subsequently incorporated within the Work.
|
||||
|
||||
2. Grant of Copyright License. Subject to the terms and conditions of
|
||||
this License, each Contributor hereby grants to You a perpetual,
|
||||
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
||||
copyright license to reproduce, prepare Derivative Works of,
|
||||
publicly display, publicly perform, sublicense, and distribute the
|
||||
Work and such Derivative Works in Source or Object form.
|
||||
|
||||
3. Grant of Patent License. Subject to the terms and conditions of
|
||||
this License, each Contributor hereby grants to You a perpetual,
|
||||
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
||||
(except as stated in this section) patent license to make, have made,
|
||||
use, offer to sell, sell, import, and otherwise transfer the Work,
|
||||
where such license applies only to those patent claims licensable
|
||||
by such Contributor that are necessarily infringed by their
|
||||
Contribution(s) alone or by combination of their Contribution(s)
|
||||
with the Work to which such Contribution(s) was submitted. If You
|
||||
institute patent litigation against any entity (including a
|
||||
cross-claim or counterclaim in a lawsuit) alleging that the Work
|
||||
or a Contribution incorporated within the Work constitutes direct
|
||||
or contributory patent infringement, then any patent licenses
|
||||
granted to You under this License for that Work shall terminate
|
||||
as of the date such litigation is filed.
|
||||
|
||||
4. Redistribution. You may reproduce and distribute copies of the
|
||||
Work or Derivative Works thereof in any medium, with or without
|
||||
modifications, and in Source or Object form, provided that You
|
||||
meet the following conditions:
|
||||
|
||||
(a) You must give any other recipients of the Work or
|
||||
Derivative Works a copy of this License; and
|
||||
|
||||
(b) You must cause any modified files to carry prominent notices
|
||||
stating that You changed the files; and
|
||||
|
||||
(c) You must retain, in the Source form of any Derivative Works
|
||||
that You distribute, all copyright, patent, trademark, and
|
||||
attribution notices from the Source form of the Work,
|
||||
excluding those notices that do not pertain to any part of
|
||||
the Derivative Works; and
|
||||
|
||||
(d) If the Work includes a "NOTICE" text file as part of its
|
||||
distribution, then any Derivative Works that You distribute must
|
||||
include a readable copy of the attribution notices contained
|
||||
within such NOTICE file, excluding those notices that do not
|
||||
pertain to any part of the Derivative Works, in at least one
|
||||
of the following places: within a NOTICE text file distributed
|
||||
as part of the Derivative Works; within the Source form or
|
||||
documentation, if provided along with the Derivative Works; or,
|
||||
within a display generated by the Derivative Works, if and
|
||||
wherever such third-party notices normally appear. The contents
|
||||
of the NOTICE file are for informational purposes only and
|
||||
do not modify the License. You may add Your own attribution
|
||||
notices within Derivative Works that You distribute, alongside
|
||||
or as an addendum to the NOTICE text from the Work, provided
|
||||
that such additional attribution notices cannot be construed
|
||||
as modifying the License.
|
||||
|
||||
You may add Your own copyright statement to Your modifications and
|
||||
may provide additional or different license terms and conditions
|
||||
for use, reproduction, or distribution of Your modifications, or
|
||||
for any such Derivative Works as a whole, provided Your use,
|
||||
reproduction, and distribution of the Work otherwise complies with
|
||||
the conditions stated in this License.
|
||||
|
||||
5. Submission of Contributions. Unless You explicitly state otherwise,
|
||||
any Contribution intentionally submitted for inclusion in the Work
|
||||
by You to the Licensor shall be under the terms and conditions of
|
||||
this License, without any additional terms or conditions.
|
||||
Notwithstanding the above, nothing herein shall supersede or modify
|
||||
the terms of any separate license agreement you may have executed
|
||||
with Licensor regarding such Contributions.
|
||||
|
||||
6. Trademarks. This License does not grant permission to use the trade
|
||||
names, trademarks, service marks, or product names of the Licensor,
|
||||
except as required for reasonable and customary use in describing the
|
||||
origin of the Work and reproducing the content of the NOTICE file.
|
||||
|
||||
7. Disclaimer of Warranty. Unless required by applicable law or
|
||||
agreed to in writing, Licensor provides the Work (and each
|
||||
Contributor provides its Contributions) on an "AS IS" BASIS,
|
||||
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
||||
implied, including, without limitation, any warranties or conditions
|
||||
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
||||
PARTICULAR PURPOSE. You are solely responsible for determining the
|
||||
appropriateness of using or redistributing the Work and assume any
|
||||
risks associated with Your exercise of permissions under this License.
|
||||
|
||||
8. Limitation of Liability. In no event and under no legal theory,
|
||||
whether in tort (including negligence), contract, or otherwise,
|
||||
unless required by applicable law (such as deliberate and grossly
|
||||
negligent acts) or agreed to in writing, shall any Contributor be
|
||||
liable to You for damages, including any direct, indirect, special,
|
||||
incidental, or consequential damages of any character arising as a
|
||||
result of this License or out of the use or inability to use the
|
||||
Work (including but not limited to damages for loss of goodwill,
|
||||
work stoppage, computer failure or malfunction, or any and all
|
||||
other commercial damages or losses), even if such Contributor
|
||||
has been advised of the possibility of such damages.
|
||||
|
||||
9. Accepting Warranty or Additional Liability. While redistributing
|
||||
the Work or Derivative Works thereof, You may choose to offer,
|
||||
and charge a fee for, acceptance of support, warranty, indemnity,
|
||||
or other liability obligations and/or rights consistent with this
|
||||
License. However, in accepting such obligations, You may act only
|
||||
on Your own behalf and on Your sole responsibility, not on behalf
|
||||
of any other Contributor, and only if You agree to indemnify,
|
||||
defend, and hold each Contributor harmless for any liability
|
||||
incurred by, or claims asserted against, such Contributor by reason
|
||||
of your accepting any such warranty or additional liability.
|
||||
|
||||
END OF TERMS AND CONDITIONS
|
||||
|
||||
APPENDIX: How to apply the Apache License to your work.
|
||||
|
||||
To apply the Apache License to your work, attach the following
|
||||
boilerplate notice, with the fields enclosed by brackets "[]"
|
||||
replaced with your own identifying information. (Don't include
|
||||
the brackets!) The text should be enclosed in the appropriate
|
||||
comment syntax for the file format. We also recommend that a
|
||||
file or class name and description of purpose be included on the
|
||||
same "printed page" as the copyright notice for easier
|
||||
identification within third-party archives.
|
||||
|
||||
Copyright [yyyy] [name of copyright owner]
|
||||
|
||||
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.
|
||||
@@ -0,0 +1,29 @@
|
||||
BSD 3-Clause License
|
||||
|
||||
Copyright (c) 2019, Seungwon Park 박승원
|
||||
All rights reserved.
|
||||
|
||||
Redistribution and use in source and binary forms, with or without
|
||||
modification, are permitted provided that the following conditions are met:
|
||||
|
||||
1. Redistributions of source code must retain the above copyright notice, this
|
||||
list of conditions and the following disclaimer.
|
||||
|
||||
2. Redistributions in binary form must reproduce the above copyright notice,
|
||||
this list of conditions and the following disclaimer in the documentation
|
||||
and/or other materials provided with the distribution.
|
||||
|
||||
3. Neither the name of the copyright holder nor the names of its
|
||||
contributors may be used to endorse or promote products derived from
|
||||
this software without specific prior written permission.
|
||||
|
||||
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
||||
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
||||
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
|
||||
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
|
||||
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
||||
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
|
||||
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
|
||||
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
|
||||
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
||||
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
||||
@@ -0,0 +1,16 @@
|
||||
Copyright 2020 Alexandre Défossez
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and
|
||||
associated documentation files (the "Software"), to deal in the Software without restriction,
|
||||
including without limitation the rights to use, copy, modify, merge, publish, distribute,
|
||||
sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is
|
||||
furnished to do so, subject to the following conditions:
|
||||
|
||||
The above copyright notice and this permission notice shall be included in all copies or
|
||||
substantial portions of the Software.
|
||||
|
||||
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT
|
||||
NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
|
||||
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
|
||||
DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
||||
@@ -0,0 +1,21 @@
|
||||
MIT License
|
||||
|
||||
Copyright (c) 2023-present, Descript
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
of this software and associated documentation files (the "Software"), to deal
|
||||
in the Software without restriction, including without limitation the rights
|
||||
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
||||
copies of the Software, and to permit persons to whom the Software is
|
||||
furnished to do so, subject to the following conditions:
|
||||
|
||||
The above copyright notice and this permission notice shall be included in all
|
||||
copies or substantial portions of the Software.
|
||||
|
||||
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
||||
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
||||
SOFTWARE.
|
||||
@@ -0,0 +1,21 @@
|
||||
MIT License
|
||||
|
||||
Copyright (c) 2023 Charactr Inc.
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
of this software and associated documentation files (the "Software"), to deal
|
||||
in the Software without restriction, including without limitation the rights
|
||||
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
||||
copies of the Software, and to permit persons to whom the Software is
|
||||
furnished to do so, subject to the following conditions:
|
||||
|
||||
The above copyright notice and this permission notice shall be included in all
|
||||
copies or substantial portions of the Software.
|
||||
|
||||
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
||||
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
||||
SOFTWARE.
|
||||
@@ -0,0 +1,21 @@
|
||||
MIT License
|
||||
|
||||
Copyright (c) 2023 Amphion
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
of this software and associated documentation files (the "Software"), to deal
|
||||
in the Software without restriction, including without limitation the rights
|
||||
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
||||
copies of the Software, and to permit persons to whom the Software is
|
||||
furnished to do so, subject to the following conditions:
|
||||
|
||||
The above copyright notice and this permission notice shall be included in all
|
||||
copies or substantial portions of the Software.
|
||||
|
||||
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
||||
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
||||
SOFTWARE.
|
||||
@@ -0,0 +1,31 @@
|
||||
# Adapted from https://github.com/jik876/hifi-gan under the MIT license.
|
||||
# LICENSE is in incl_licenses directory.
|
||||
|
||||
import os
|
||||
|
||||
import torch
|
||||
from torch.nn.utils import weight_norm
|
||||
|
||||
|
||||
def init_weights(m, mean=0.0, std=0.01):
|
||||
classname = m.__class__.__name__
|
||||
if classname.find("Conv") != -1:
|
||||
m.weight.data.normal_(mean, std)
|
||||
|
||||
|
||||
def apply_weight_norm(m):
|
||||
classname = m.__class__.__name__
|
||||
if classname.find("Conv") != -1:
|
||||
weight_norm(m)
|
||||
|
||||
|
||||
def get_padding(kernel_size, dilation=1):
|
||||
return int((kernel_size * dilation - dilation) / 2)
|
||||
|
||||
|
||||
def load_checkpoint(filepath, device):
|
||||
assert os.path.isfile(filepath)
|
||||
print(f"Loading '{filepath}'")
|
||||
checkpoint_dict = torch.load(filepath, map_location=device)
|
||||
print("Complete.")
|
||||
return checkpoint_dict
|
||||
Reference in New Issue
Block a user