- 计算预测值
model的输出output是一个(256, 10)的数据,需要计算预测类别的最大值
outputs = model(images)
_, predicted = torch.max(output.data, 1)
- 计算精度TP
total += (predicted == labels).sum().item()
num_samples += labels.size(0)
return total / num_samples
- 构造零矩阵
onezero = torch.zeros(outputs.shape[0], outputs.shape[1])
- 将零矩阵对应位置改为1
onezero[partialY > 0] = 1
- 每个位置做 1-x 操作
counter_onezero = 1- onezero
- 修改示例:
x = torch.tensor([[1, -1, -1], [2, -2, 3]])
y = 0.5 * torch.ones(x.shape[0], x.shape[1])
print(x)
print(y)
print(x[x < 0])
print(x[x < 0].shape)
print(y[x < 0])
y[x < 0] = 0
print(y)
示例输出: