有没有办法将预训练模型从 PyTorch 转换为 ONNX?

Question

我在自定义数据集上训练了StarGAN模型。 我需要将此模型从.pth(Pytorch(转换为.pb，以便在Android studio上使用。 我搜索了很多，我找到了一些转换的方法。 但是，所有解决方案都不适用于我的情况。

我尝试了仅包含一个nn的小型网络。线性层。 在这个网络上，解决方案运行良好！

我认为，我的网络包括 Conv2D 层和 MaxPooling2D 层，因此转换处理不起作用。

首先，这是我的网络(StarGAN(。

import torch
import torch.nn as nn
import numpy as np

class ResidualBlock(nn.Module):
def __init__(self, dim_in, dim_out):
super(ResidualBlock, self).__init__()
self.main = nn.Sequential(
nn.Conv2d(dim_in, dim_out, kernel_size=3, stride=1, padding=1, bias=False),
nn.InstanceNorm2d(dim_out, affine=True, track_running_stats=True),
nn.ReLU(inplace=True),
nn.Conv2d(dim_out, dim_out, kernel_size=3, stride=1, padding=1, bias=False),
nn.InstanceNorm2d(dim_out, affine=True, track_running_stats=True))
def forward(self, x):
return x + self.main(x)

class Generator(nn.Module):
def __init__(self, conv_dim=64, c_dim=5, repeat_num=6):
super(Generator, self).__init__()
layers = []
layers.append(nn.Conv2d(3 + c_dim, conv_dim, kernel_size=7, stride=1, padding=3, bias=False))
layers.append(nn.InstanceNorm2d(conv_dim, affine=True, track_running_stats=True))
layers.append(nn.ReLU(inplace=True))
curr_dim = conv_dim
for _ in range(2):
layers.append(nn.Conv2d(curr_dim, curr_dim * 2, kernel_size=4, stride=2, padding=1, bias=False))
layers.append(nn.InstanceNorm2d(curr_dim * 2, affine=True, track_running_stats=True))
layers.append(nn.ReLU(inplace=True))
curr_dim = curr_dim * 2
for _ in range(repeat_num):
layers.append(ResidualBlock(dim_in=curr_dim, dim_out=curr_dim))
for _ in range(2):
layers.append(nn.ConvTranspose2d(curr_dim, curr_dim // 2, kernel_size=4, stride=2, padding=1, bias=False))
layers.append(nn.InstanceNorm2d(curr_dim // 2, affine=True, track_running_stats=True))
layers.append(nn.ReLU(inplace=True))
curr_dim = curr_dim // 2
layers.append(nn.Conv2d(curr_dim, 3, kernel_size=7, stride=1, padding=3, bias=False))
layers.append(nn.Tanh())
self.main = nn.Sequential(*layers)
def forward(self, x, c):
c = c.view(c.size(0), c.size(1), 1, 1)
c = c.repeat(1, 1, x.size(2), x.size(3))
x = torch.cat([x, c], dim=1)
return self.main(x)

class Discriminator(nn.Module):
def __init__(self, image_size=128, conv_dim=64, c_dim=5, repeat_num=6):
super(Discriminator, self).__init__()
layers = []
layers.append(nn.Conv2d(3, conv_dim, kernel_size=4, stride=2, padding=1))
layers.append(nn.LeakyReLU(0.01))
curr_dim = conv_dim
for _ in range(1, repeat_num):
layers.append(nn.Conv2d(curr_dim, curr_dim * 2, kernel_size=4, stride=2, padding=1))
layers.append(nn.LeakyReLU(0.01))
curr_dim = curr_dim * 2
kernel_size = int(image_size / np.power(2, repeat_num))
self.main = nn.Sequential(*layers)
self.conv1 = nn.Conv2d(curr_dim, 1, kernel_size=3, stride=1, padding=1, bias=False)
self.conv2 = nn.Conv2d(curr_dim, c_dim, kernel_size=kernel_size, bias=False)
def forward(self, x):
h = self.main(x)
out_src = self.conv1(h)
out_cls = self.conv2(h)
return out_src, out_cls.view(out_cls.size(0), out_cls.size(1))

这是错误消息。

TypeError: object of type 'torch._C.Value' has no len() (occurred when translating repeat)

有什么方法可以转换吗？帮帮我。

Answer 1

我在尝试使用 TensorboardX 生成模型图时遇到了同样的问题。

我相信错误来自运营商torch.onnx目前支持的内容。您可以查看此链接：
https://pytorch.org/docs/stable/onnx.html
在支持的运算符部分下，您将看到未列出repeat。

为了回答您的问题，您目前似乎无法使用带有torch.onnx的repeat转换模型。