{"id":998335,"date":"2024-12-27T09:32:08","date_gmt":"2024-12-27T01:32:08","guid":{"rendered":"https:\/\/docs.pingcode.com\/ask\/ask-ask\/998335.html"},"modified":"2024-12-27T09:32:14","modified_gmt":"2024-12-27T01:32:14","slug":"python%e5%a6%82%e4%bd%95%e6%98%be%e7%a4%ba%e7%bd%91%e7%bb%9c%e6%9d%83%e9%87%8d","status":"publish","type":"post","link":"https:\/\/docs.pingcode.com\/ask\/998335.html","title":{"rendered":"python\u5982\u4f55\u663e\u793a\u7f51\u7edc\u6743\u91cd"},"content":{"rendered":"

\"python\u5982\u4f55\u663e\u793a\u7f51\u7edc\u6743\u91cd\"<\/p>\n

\u5728Python\u4e2d\uff0c\u663e\u793a\u795e\u7ecf\u7f51\u7edc\u6743\u91cd\u7684\u65b9\u6cd5\u4e3b\u8981\u6709\uff1a\u4f7f\u7528\u6df1\u5ea6\u5b66\u4e60\u5e93\uff08\u5982TensorFlow\u6216PyTorch\uff09\u63d0\u53d6\u6a21\u578b\u7684\u6743\u91cd\u3001\u901a\u8fc7\u8bbf\u95ee\u6a21\u578b\u5c42\u7684\u53c2\u6570\u5c5e\u6027\u3001\u4f7f\u7528\u53ef\u89c6\u5316\u5de5\u5177\u5c06\u6743\u91cd\u56fe\u5f62\u5316\u663e\u793a\u3002<\/strong> \u4e0b\u9762\u6211\u5c06\u8be6\u7ec6\u63cf\u8ff0\u5982\u4f55\u5728PyTorch\u4e2d\u8bbf\u95ee\u548c\u663e\u793a\u7f51\u7edc\u6743\u91cd\u3002<\/p>\n<\/p>\n

\u5728\u6df1\u5ea6\u5b66\u4e60\u9879\u76ee\u4e2d\uff0c\u4e86\u89e3\u548c\u53ef\u89c6\u5316\u7f51\u7edc\u6743\u91cd\u53ef\u4ee5\u5e2e\u52a9\u6211\u4eec\u66f4\u597d\u5730\u7406\u89e3\u6a21\u578b\u7684\u5185\u90e8\u5de5\u4f5c\u673a\u5236\u3002PyTorch\u4f5c\u4e3a\u4e00\u4e2a\u7075\u6d3b\u4e14\u5f3a\u5927\u7684\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\uff0c\u63d0\u4f9b\u4e86\u591a\u79cd\u65b9\u6cd5\u6765\u8bbf\u95ee\u548c\u663e\u793a\u6a21\u578b\u7684\u6743\u91cd\u3002<\/p>\n<\/p>\n

\u4e00\u3001\u4f7f\u7528PYTORCH\u63d0\u53d6\u7f51\u7edc\u6743\u91cd<\/h3>\n<\/p>\n

\u5728PyTorch\u4e2d\uff0c\u6a21\u578b\u7684\u6743\u91cd\u5b58\u50a8\u5728\u6bcf\u4e00\u5c42\u7684\u53c2\u6570\u4e2d\u3002\u6211\u4eec\u53ef\u4ee5\u901a\u8fc7\u8fed\u4ee3\u6a21\u578b\u7684\u53c2\u6570\u6765\u63d0\u53d6\u8fd9\u4e9b\u6743\u91cd\u3002<\/p>\n<\/p>\n

1. \u83b7\u53d6\u6743\u91cd<\/h4>\n<\/p>\n

\u5728PyTorch\u4e2d\uff0c\u53ef\u4ee5\u4f7f\u7528model.parameters()<\/code>\u65b9\u6cd5\u6765\u904d\u5386\u6a21\u578b\u7684\u6240\u6709\u53c2\u6570\uff0c\u5305\u62ec\u6743\u91cd\u548c\u504f\u7f6e\u9879\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u7b80\u5355\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n

import torch<\/p>\n
import torch.nn as nn<\/p>\n
\u5b9a\u4e49\u4e00\u4e2a\u7b80\u5355\u7684\u795e\u7ecf\u7f51\u7edc<\/strong><\/h2>\n
class SimpleNet(nn.Module):<\/p>\n
    def __init__(self):<\/p>\n
        super(SimpleNet, self).__init__()<\/p>\n
        self.fc1 = nn.Linear(10, 5)<\/p>\n
        self.fc2 = nn.Linear(5, 2)<\/p>\n
    def forward(self, x):<\/p>\n
        x = self.fc1(x)<\/p>\n
        x = self.fc2(x)<\/p>\n
        return x<\/p>\n
\u521b\u5efa\u4e00\u4e2a\u7f51\u7edc\u5b9e\u4f8b<\/strong><\/h2>\n
model = SimpleNet()<\/p>\n
\u83b7\u53d6\u6a21\u578b\u7684\u6743\u91cd<\/strong><\/h2>\n
for param in model.parameters():<\/p>\n
    print(param)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u5b9a\u4e49\u4e86\u4e00\u4e2a\u7b80\u5355\u7684\u795e\u7ecf\u7f51\u7edcSimpleNet<\/code>\uff0c\u5b83\u5305\u542b\u4e24\u4e2a\u5168\u8fde\u63a5\u5c42\u3002\u901a\u8fc7\u8fed\u4ee3model.parameters()<\/code>\uff0c\u6211\u4eec\u53ef\u4ee5\u8bbf\u95ee\u5e76\u6253\u5370\u51fa\u6bcf\u4e00\u5c42\u7684\u6743\u91cd\u548c\u504f\u7f6e\u3002<\/p>\n<\/p>\n
2. \u8bbf\u95ee\u7279\u5b9a\u5c42\u7684\u6743\u91cd<\/h4>\n<\/p>\n
\u5982\u679c\u6211\u4eec\u53ea\u60f3\u8bbf\u95ee\u6a21\u578b\u4e2d\u7279\u5b9a\u5c42\u7684\u6743\u91cd\uff0c\u53ef\u4ee5\u76f4\u63a5\u8bbf\u95ee\u8be5\u5c42\u7684\u6743\u91cd\u5c5e\u6027\u3002\u4f8b\u5982\uff1a<\/p>\n<\/p>\n
# \u8bbf\u95ee\u7b2c\u4e00\u4e2a\u5168\u8fde\u63a5\u5c42\u7684\u6743\u91cd<\/p>\n
fc1_weights = model.fc1.weight.data<\/p>\n
print(fc1_weights)<\/p>\n
\u8bbf\u95ee\u7b2c\u4e00\u4e2a\u5168\u8fde\u63a5\u5c42\u7684\u504f\u7f6e<\/strong><\/h2>\n
fc1_bias = model.fc1.bias.data<\/p>\n
print(fc1_bias)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u901a\u8fc7\u8fd9\u79cd\u65b9\u6cd5\uff0c\u6211\u4eec\u53ef\u4ee5\u76f4\u63a5\u83b7\u53d6\u7279\u5b9a\u5c42\u7684\u6743\u91cd\u548c\u504f\u7f6e\uff0c\u5e76\u5bf9\u5176\u8fdb\u884c\u8fdb\u4e00\u6b65\u7684\u5206\u6790\u6216\u5904\u7406\u3002<\/p>\n<\/p>\n
\u4e8c\u3001\u4f7f\u7528\u53ef\u89c6\u5316\u5de5\u5177\u663e\u793a\u6743\u91cd<\/h3>\n<\/p>\n
\u9664\u4e86\u76f4\u63a5\u8bbf\u95ee\u548c\u6253\u5370\u6743\u91cd\u5916\uff0c\u6211\u4eec\u8fd8\u53ef\u4ee5\u4f7f\u7528\u53ef\u89c6\u5316\u5de5\u5177\u5c06\u6743\u91cd\u56fe\u5f62\u5316\u663e\u793a\u3002\u8fd9\u6709\u52a9\u4e8e\u6211\u4eec\u66f4\u76f4\u89c2\u5730\u7406\u89e3\u7f51\u7edc\u7684\u5de5\u4f5c\u673a\u5236\u3002<\/p>\n<\/p>\n
1. \u4f7f\u7528MATPLOTLIB\u53ef\u89c6\u5316\u6743\u91cd<\/h4>\n<\/p>\n
matplotlib<\/code>\u662f\u4e00\u4e2a\u5f3a\u5927\u7684\u7ed8\u56fe\u5e93\uff0c\u53ef\u4ee5\u7528\u4e8e\u5c06\u6743\u91cd\u6570\u636e\u53ef\u89c6\u5316\u3002<\/p>\n<\/p>\n
import matplotlib.pyplot as plt<\/p>\n
\u7ed8\u5236\u7b2c\u4e00\u4e2a\u5168\u8fde\u63a5\u5c42\u7684\u6743\u91cd<\/strong><\/h2>\n
plt.imshow(fc1_weights, cmap='viridis', aspect='auto')<\/p>\n
plt.colorbar()<\/p>\n
plt.title('FC1 Weights')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u4f7f\u7528imshow<\/code>\u51fd\u6570\u5c06\u7b2c\u4e00\u4e2a\u5168\u8fde\u63a5\u5c42\u7684\u6743\u91cd\u53ef\u89c6\u5316\uff0c\u5e76\u4f7f\u7528colorbar<\/code>\u6dfb\u52a0\u4e00\u4e2a\u989c\u8272\u6761\u4ee5\u663e\u793a\u6743\u91cd\u7684\u6570\u503c\u8303\u56f4\u3002<\/p>\n<\/p>\n
2. \u4f7f\u7528SEABORN\u8fdb\u884c\u66f4\u9ad8\u7ea7\u7684\u53ef\u89c6\u5316<\/h4>\n<\/p>\n
Seaborn\u662f\u57fa\u4e8ematplotlib\u7684\u9ad8\u7ea7\u53ef\u89c6\u5316\u5e93\uff0c\u53ef\u4ee5\u7528\u4e8e\u521b\u5efa\u66f4\u7f8e\u89c2\u7684\u56fe\u5f62\u3002<\/p>\n<\/p>\n
import seaborn as sns<\/p>\n
\u4f7f\u7528seaborn\u53ef\u89c6\u5316\u6743\u91cd<\/strong><\/h2>\n
sns.heatmap(fc1_weights, annot=True, cmap='coolwarm', fmt=".2f")<\/p>\n
plt.title('FC1 Weights Heatmap')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u901a\u8fc7\u8fd9\u79cd\u65b9\u6cd5\uff0c\u6211\u4eec\u53ef\u4ee5\u521b\u5efa\u66f4\u52a0\u7f8e\u89c2\u4e14\u4fe1\u606f\u4e30\u5bcc\u7684\u6743\u91cd\u70ed\u56fe\u3002<\/p>\n<\/p>\n
\u4e09\u3001\u5728\u5927\u578b\u6a21\u578b\u4e2d\u7684\u5e94\u7528<\/h3>\n<\/p>\n
\u5bf9\u4e8e\u5927\u578b\u6a21\u578b\uff0c\u5982\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\u548c\u6df1\u5ea6\u6b8b\u5dee\u7f51\u7edc\uff0c\u6743\u91cd\u7684\u6570\u91cf\u548c\u7ed3\u6784\u66f4\u52a0\u590d\u6742\u3002\u6211\u4eec\u53ef\u4ee5\u901a\u8fc7\u7c7b\u4f3c\u7684\u65b9\u6cd5\u8bbf\u95ee\u7279\u5b9a\u5c42\u7684\u6743\u91cd\uff0c\u5e76\u9009\u62e9\u6027\u5730\u8fdb\u884c\u53ef\u89c6\u5316\u3002<\/p>\n<\/p>\n
1. \u8bbf\u95ee\u5377\u79ef\u5c42\u7684\u6743\u91cd<\/h4>\n<\/p>\n
\u5bf9\u4e8e\u5377\u79ef\u5c42\uff0c\u6743\u91cd\u901a\u5e38\u662f\u56db\u7ef4\u7684\uff08\u8f93\u51fa\u901a\u9053\u6570\uff0c\u8f93\u5165\u901a\u9053\u6570\uff0c\u9ad8\u5ea6\uff0c\u5bbd\u5ea6\uff09\u3002\u6211\u4eec\u53ef\u4ee5\u9009\u62e9\u6027\u5730\u53ef\u89c6\u5316\u67d0\u4e2a\u8f93\u51fa\u901a\u9053\u7684\u6743\u91cd\u3002<\/p>\n<\/p>\n
class ConvNet(nn.Module):<\/p>\n
    def __init__(self):<\/p>\n
        super(ConvNet, self).__init__()<\/p>\n
        self.conv1 = nn.Conv2d(1, 6, 3)<\/p>\n
    def forward(self, x):<\/p>\n
        x = self.conv1(x)<\/p>\n
        return x<\/p>\n
conv_model = ConvNet()<\/p>\n
\u83b7\u53d6\u5377\u79ef\u5c42\u7684\u6743\u91cd<\/strong><\/h2>\n
conv1_weights = conv_model.conv1.weight.data<\/p>\n
\u53ef\u89c6\u5316\u7b2c\u4e00\u4e2a\u8f93\u51fa\u901a\u9053\u7684\u6743\u91cd<\/strong><\/h2>\n
plt.imshow(conv1_weights[0, 0, :, :], cmap='gray')<\/p>\n
plt.title('Conv1 Channel 1 Weights')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
2. \u5728\u9884\u8bad\u7ec3\u6a21\u578b\u4e2d\u63d0\u53d6\u6743\u91cd<\/h4>\n<\/p>\n
\u5bf9\u4e8e\u4f7f\u7528\u9884\u8bad\u7ec3\u6a21\u578b\uff08\u5982ResNet\u3001VGG\uff09\u7684\u60c5\u51b5\uff0c\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528torchvision.models<\/code>\u6a21\u5757\u52a0\u8f7d\u6a21\u578b\uff0c\u5e76\u63d0\u53d6\u5176\u6743\u91cd\u3002<\/p>\n<\/p>\n
from torchvision import models<\/p>\n
\u52a0\u8f7d\u9884\u8bad\u7ec3\u7684ResNet18<\/strong><\/h2>\n
resnet18 = models.resnet18(pretrAI<\/a>ned=True)<\/p>\n
\u83b7\u53d6\u6a21\u578b\u7684\u7b2c\u4e00\u4e2a\u5377\u79ef\u5c42\u7684\u6743\u91cd<\/strong><\/h2>\n
first_conv_weights = resnet18.conv1.weight.data<\/p>\n
\u53ef\u89c6\u5316\u7b2c\u4e00\u4e2a\u5377\u79ef\u6838\u7684\u6743\u91cd<\/strong><\/h2>\n
plt.imshow(first_conv_weights[0, 0, :, :], cmap='gray')<\/p>\n
plt.title('ResNet18 First Conv Kernel 1 Weights')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u901a\u8fc7\u8fd9\u79cd\u65b9\u5f0f\uff0c\u6211\u4eec\u53ef\u4ee5\u8f7b\u677e\u5730\u8bbf\u95ee\u548c\u5206\u6790\u590d\u6742\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\u4e2d\u7684\u6743\u91cd\u3002<\/p>\n<\/p>\n
\u56db\u3001\u603b\u7ed3\u4e0e\u7ecf\u9a8c\u5206\u4eab<\/h3>\n<\/p>\n
\u5728\u5b9e\u8df5\u4e2d\uff0c\u5206\u6790\u548c\u53ef\u89c6\u5316\u795e\u7ecf\u7f51\u7edc\u7684\u6743\u91cd\u53ef\u4ee5\u5e2e\u52a9\u6211\u4eec\u66f4\u597d\u5730\u7406\u89e3\u6a21\u578b\u7684\u8bad\u7ec3\u8fc7\u7a0b\u548c\u6027\u80fd\u8868\u73b0\u3002\u4ee5\u4e0b\u662f\u4e00\u4e9b\u7ecf\u9a8c\u5206\u4eab\uff1a<\/p>\n<\/p>\n
    \n
  1. \n
    \u6743\u91cd\u5206\u6790\u7684\u91cd\u8981\u6027<\/strong>\uff1a\u901a\u8fc7\u5206\u6790\u6743\u91cd\uff0c\u6211\u4eec\u53ef\u4ee5\u53d1\u73b0\u6a21\u578b\u5728\u8bad\u7ec3\u8fc7\u7a0b\u4e2d\u662f\u5426\u5b58\u5728\u8fc7\u62df\u5408\u3001\u6b20\u62df\u5408\u7b49\u95ee\u9898\u3002\u4f8b\u5982\uff0c\u6743\u91cd\u8fc7\u5927\u53ef\u80fd\u5bfc\u81f4\u6a21\u578b\u8fc7\u62df\u5408\uff0c\u800c\u6743\u91cd\u8fc7\u5c0f\u53ef\u80fd\u5bfc\u81f4\u6a21\u578b\u96be\u4ee5\u5b66\u4e60\u3002<\/p>\n<\/p>\n<\/li>\n
  2. \n
    \u53ef\u89c6\u5316\u7684\u4ef7\u503c<\/strong>\uff1a\u53ef\u89c6\u5316\u6743\u91cd\u53ef\u4ee5\u5e2e\u52a9\u6211\u4eec\u66f4\u76f4\u89c2\u5730\u7406\u89e3\u6a21\u578b\u7684\u5de5\u4f5c\u673a\u5236\uff0c\u5c24\u5176\u662f\u5728\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\u4e2d\uff0c\u901a\u8fc7\u53ef\u89c6\u5316\u5377\u79ef\u6838\u7684\u6743\u91cd\uff0c\u6211\u4eec\u53ef\u4ee5\u89c2\u5bdf\u5230\u6a21\u578b\u5b66\u4e60\u5230\u7684\u7279\u5f81\u3002<\/p>\n<\/p>\n<\/li>\n
  3. \n
    \u5de5\u5177\u7684\u9009\u62e9<\/strong>\uff1a\u5728\u8fdb\u884c\u53ef\u89c6\u5316\u65f6\uff0c\u6839\u636e\u6570\u636e\u7684\u6027\u8d28\u9009\u62e9\u5408\u9002\u7684\u5de5\u5177\u548c\u65b9\u6cd5\uff0c\u4f8b\u5982\uff0c\u5bf9\u4e8e\u4e8c\u7ef4\u6570\u636e\u53ef\u4ee5\u4f7f\u7528\u70ed\u56fe\uff0c\u5bf9\u4e8e\u9ad8\u7ef4\u6570\u636e\u53ef\u4ee5\u4f7f\u7528\u964d\u7ef4\u6280\u672f\u8fdb\u884c\u53ef\u89c6\u5316\u3002<\/p>\n<\/p>\n<\/li>\n
  4. \n
    \u6301\u7eed\u5b66\u4e60\u4e0e\u63a2\u7d22<\/strong>\uff1a\u968f\u7740\u6df1\u5ea6\u5b66\u4e60\u6280\u672f\u7684\u53d1\u5c55\uff0c\u65b0\u7684\u7f51\u7edc\u7ed3\u6784\u548c\u8bad\u7ec3\u65b9\u6cd5\u4e0d\u65ad\u6d8c\u73b0\u3002\u5728\u5206\u6790\u548c\u53ef\u89c6\u5316\u6743\u91cd\u65f6\uff0c\u6211\u4eec\u9700\u8981\u6301\u7eed\u5b66\u4e60\u548c\u63a2\u7d22\uff0c\u4ee5\u4fbf\u66f4\u597d\u5730\u7406\u89e3\u548c\u5e94\u7528\u8fd9\u4e9b\u65b0\u6280\u672f\u3002<\/p>\n<\/p>\n<\/li>\n<\/ol>\n
    \u901a\u8fc7\u4e0a\u8ff0\u65b9\u6cd5\u548c\u6280\u5de7\uff0c\u6211\u4eec\u53ef\u4ee5\u5728\u5b9e\u8df5\u4e2d\u66f4\u597d\u5730\u5206\u6790\u548c\u7406\u89e3\u795e\u7ecf\u7f51\u7edc\u7684\u6743\u91cd\uff0c\u4ece\u800c\u63d0\u5347\u6a21\u578b\u7684\u6027\u80fd\u548c\u7a33\u5b9a\u6027\u3002<\/p>\n<\/p>\n
    \u76f8\u5173\u95ee\u7b54FAQs\uff1a<\/strong><\/h2>\n
     \u5982\u4f55\u5728Python\u4e2d\u63d0\u53d6\u548c\u663e\u793a\u795e\u7ecf\u7f51\u7edc\u7684\u6743\u91cd\uff1f<\/strong>
    \u5728Python\u4e2d\uff0c\u53ef\u4ee5\u4f7f\u7528\u6df1\u5ea6\u5b66\u4e60\u5e93\u5982TensorFlow\u6216PyTorch\u6765\u63d0\u53d6\u795e\u7ecf\u7f51\u7edc\u7684\u6743\u91cd\u3002\u4ee5PyTorch\u4e3a\u4f8b\uff0c\u53ef\u4ee5\u901a\u8fc7\u8bbf\u95ee\u6a21\u578b\u7684\u53c2\u6570\u6765\u83b7\u53d6\u6743\u91cd\uff0c\u4f7f\u7528model.parameters()<\/code>\u65b9\u6cd5\u53ef\u4ee5\u904d\u5386\u6bcf\u4e00\u5c42\u7684\u6743\u91cd\u548c\u504f\u7f6e\u3002\u5bf9\u4e8eTensorFlow\uff0c\u4f7f\u7528model.weights<\/code>\u53ef\u4ee5\u8f7b\u677e\u83b7\u53d6\u6240\u6709\u6743\u91cd\u3002\u5728\u63d0\u53d6\u540e\uff0c\u5229\u7528NumPy\u6216Matplotlib\u7b49\u5e93\u53ef\u4ee5\u5c06\u8fd9\u4e9b\u6743\u91cd\u4ee5\u56fe\u5f62\u5316\u7684\u65b9\u5f0f\u5c55\u793a\u3002<\/p>\n
    \u663e\u793a\u6743\u91cd\u7684\u6700\u4f73\u5b9e\u8df5\u6709\u54ea\u4e9b\uff1f<\/strong>
    \u5728\u5c55\u793a\u795e\u7ecf\u7f51\u7edc\u6743\u91cd\u65f6\uff0c\u901a\u5e38\u63a8\u8350\u4f7f\u7528\u70ed\u56fe\uff08heatmap\uff09\u6765\u76f4\u89c2\u5730\u8868\u73b0\u6743\u91cd\u7684\u5206\u5e03\u548c\u5f3a\u5ea6\u3002\u901a\u8fc7Matplotlib\u7684imshow<\/code>\u51fd\u6570\uff0c\u53ef\u4ee5\u5c06\u6743\u91cd\u77e9\u9635\u53ef\u89c6\u5316\uff0c\u5e2e\u52a9\u7406\u89e3\u6a21\u578b\u7684\u5b66\u4e60\u60c5\u51b5\u3002\u6b64\u5916\uff0c\u4f7f\u7528\u6807\u51c6\u5316\u7684\u6743\u91cd\u503c\uff08\u5982\u5f52\u4e00\u5316\u52300-1\u4e4b\u95f4\uff09\u80fd\u4f7f\u56fe\u50cf\u66f4\u5177\u53ef\u8bfb\u6027\u3002<\/p>\n
    \u5982\u4f55\u5206\u6790\u7f51\u7edc\u6743\u91cd\u5bf9\u6a21\u578b\u6027\u80fd\u7684\u5f71\u54cd\uff1f<\/strong>
    \u5206\u6790\u7f51\u7edc\u6743\u91cd\u53ef\u4ee5\u901a\u8fc7\u89c2\u5bdf\u6743\u91cd\u7684\u5206\u5e03\u548c\u53d8\u5316\u8d8b\u52bf\u6765\u8fdb\u884c\u3002\u4f8b\u5982\uff0c\u89c2\u5bdf\u6743\u91cd\u7684\u7a00\u758f\u6027\u53ef\u4ee5\u63ed\u793a\u6a21\u578b\u7684\u7b80\u7ea6\u6027\u6216\u590d\u6742\u6027\u3002\u6b64\u5916\uff0c\u5229\u7528\u6b63\u5219\u5316\u6280\u672f\u53ef\u4ee5\u5e2e\u52a9\u63a7\u5236\u6743\u91cd\u7684\u5927\u5c0f\uff0c\u4ece\u800c\u9632\u6b62\u8fc7\u62df\u5408\u3002\u901a\u8fc7\u53ef\u89c6\u5316\u5de5\u5177\uff08\u5982TensorBoard\uff09\u76d1\u63a7\u6743\u91cd\u53d8\u5316\uff0c\u53ef\u4ee5\u66f4\u6df1\u5165\u5730\u7406\u89e3\u6a21\u578b\u8bad\u7ec3\u8fc7\u7a0b\u4e2d\u7684\u8868\u73b0\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"\u5728Python\u4e2d\uff0c\u663e\u793a\u795e\u7ecf\u7f51\u7edc\u6743\u91cd\u7684\u65b9\u6cd5\u4e3b\u8981\u6709\uff1a\u4f7f\u7528\u6df1\u5ea6\u5b66\u4e60\u5e93\uff08\u5982TensorFlow\u6216PyTorch\uff09\u63d0\u53d6\u6a21 […]","protected":false},"author":3,"featured_media":998348,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[37],"tags":[],"acf":[],"_links":{"self":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/998335"}],"collection":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/comments?post=998335"}],"version-history":[{"count":"1","href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/998335\/revisions"}],"predecessor-version":[{"id":998351,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/998335\/revisions\/998351"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media\/998348"}],"wp:attachment":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media?parent=998335"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/categories?post=998335"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/tags?post=998335"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}