- class pytorch_lightning.loops.optimization.ManualOptimization¶
A special loop implementing what is known in Lightning as Manual Optimization where the optimization happens entirely in the
training_step()and therefore the user is responsible for back-propagating gradients and making calls to the optimizers.
This loop is a trivial case because it performs only a single iteration (calling directly into the module’s
training_step()) and passing through the output(s).
- advance(batch, batch_idx)¶
Performs the training step for manual optimization.
Returns the result of this loop, i.e., the post-processed outputs from the training step.
Resets the internal state of the loop at the beginning of each call to
def reset(self): # reset your internal state or add custom logic # if you expect run() to be called multiple times self.current_iteration = 0 self.outputs = 
- Return type