PyTorch 免费服务 完全指南

import torch

# 创建张量
x = torch.tensor([[1.0, 2.0], [3.0, 4.0]])
y = torch.tensor([[5.0, 6.0], [7.0, 8.0]])

# 张量加法
z = x + y
print(z)

# 矩阵乘法
w = torch.matmul(x, y)
print(w)

import torch
import torch.nn as nn

# 定义神经网络
class SimpleNN(nn.Module):
  def __init__(self):
    super(SimpleNN, self).__init__()
    self.fc1 = nn.Linear(10, 50)
    self.fc2 = nn.Linear(50, 1)
    self.relu = nn.ReLU()

  def forward(self, x):
    x = self.fc1(x)
    x = self.relu(x)
    x = self.fc2(x)
    return x

# 创建模型实例
model = SimpleNN()
print(model)

from torchvision import models
import torch

# 加载预训练的ResNet50模型
model = models.resnet50(pretrained=True)
model.eval() # 设置为评估模式

# 创建示例输入
input_tensor = torch.randn(1, 3, 224, 224)

# 进行预测
with torch.no_grad():
  output = model(input_tensor)

print(output.shape) # 输出预测结果的形状

# 首先安装transformers库
# pip install transformers

from transformers import BertTokenizer, BertModel

# 加载预训练的BERT模型和分词器
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
model = BertModel.from_pretrained('bert-base-uncased')

# 处理文本
text = "Hello, PyTorch!"
inputs = tokenizer(text, return_tensors="pt")

# 获取模型输出
outputs = model(**inputs)
print(outputs.last_hidden_state.shape)