ControlNetLatentDiffusionInferer errors out when conditioning is not False

[virginiafdez]

Describe the bug
ControlNets are essentially copies of the diffusion model encoders. However, on the ControlNetLatentDiffusionInferer, the context (conditioning) isn't passed to the ControlNet even though the DiffusionModel might take it in. If ControlNet has been defined with conditioning, this causes an error.

To Reproduce
Steps to reproduce the behaviour:

1. Create DiffusionModel and ControlNet with conditioning = True.
2. Create ControlNetDiffusionInferer
3. Run the call() function on the inferer passing dummy random tensors.

Expected behaviour
ControlNet should take in the conditioning tensor.