Re-using intermediate features without second forward pass

I have a setup with two networks, a classifier and a predictor. The classifier has two outputs, the class predictions and the intermediate features, cls, f = classifier(x). The predictor simply outputs a scalar based on those intermediate features pred = predictor(f).

Now, I would like to train both networks on the following objective: obj = CE + MSE, where CE is the cross-entropy of the classifier and MSE is the mean squared error of the predictor. The training is adversarial, i.e. the classifier attempts to both lower the CE and increase the MSE, while the predictor attempts to lower the MSE. The relevant snippet of the training procedure:

cls, f = classifier(x)

with torch.no_grad():
    pred = predictor(f)
classifier_loss = ce_loss(cls, class_target) - mse_loss(pred, prediction_target)

pred = predictor(f)
predictor_loss = mse_loss(pred, prediction_target)

Needless to say, this requires double computation of pred = predictor(f). GAN-style detaching does not seem to be applicable here.

Is there any way to perform these two backward passes without recomputing pred = predictor(f) or is it strictly necessary to construct two computation graphs and this part cannot be circumvented?