Question

我正在尝试检索训练的均方误差。在基于TensorFlow的原始代码中，我将这段代码移至PyTorch（出于研究原因）。

原始TensorFlow代码：

    print("Calculating threshold")
    x_opt_predictions = model.predict(x_opt)
    print("Calculating MSE on optimization set...")
    mse = np.mean(np.power(x_opt - x_opt_predictions, 2), axis=1)
    print("mean is %.5f" % mse.mean())
    print("min is %.5f" % mse.min())
    print("max is %.5f" % mse.max())
    print("std is %.5f" % mse.std())
    tr = mse.mean() + mse.std()

火炬的训练方法：

def train(net, x_train, x_opt, BATCH_SIZE, EPOCHS, input_dim):
    outputs = 0
    mse = 0
    optimizer = optim.SGD(net.parameters(), lr=0.001)
    loss_function = nn.MSELoss()
    loss = 0
    for epoch in range(EPOCHS):
        for i in tqdm(range(0, len(x_train), BATCH_SIZE)):
            batch_y = x_opt[i:i + BATCH_SIZE]
            
            net.zero_grad()
            
            outputs = net(batch_y)
            

            loss = loss_function(outputs, batch_y)
            loss.backward()
            optimizer.step()

        print(f"Epoch: {epoch}. Loss: {loss}")
        print("opt", x_opt.size(), "output", outputs.__sizeof__())

    # VVVVVVVVVVVVVVVVVVVVVVVVVVVVVV
    return np.mean(np.power(x_opt - outputs, 2), axis=1)
    # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

不是预测的数字数组，获取等效项以生成阈值
如果还有其他（改善或缺失）的方法来获取此价值，请提前进行升值。

Answer 1

变量输出是一个pytorch张量，将其转换为numpy，您需要更改的只是将代码行return np.mean(np.power(x_opt - outputs, 2), axis=1)更改为return np.mean(np.power(x_opt - outputs.cpu().data.numpy(), 2), axis=1)即可将张量转换为numpy数组。如果您的网络中没有使用cuda，则不需要.cpu（）部分。

Tensorflow到PyTorch-model.predict等效

1 个答案: