RuntimeError: Trying to backward through the graph a second time (or directly access saved tensors after they have already been freed). Saved intermediate values of the graph are freed when you call .backward() or autograd.grad()

vision
2

This line of code:

run_loss_disc+=loss_disc

looks wrong as you are accumulating the current discriminator loss including the entire computation graph.
If you want to accumulate the loss value only for printing purposes, use run_loss_dict += loss_dict.item() or .detach() the tensor.