Fine-tuning example is broken

[candalfigomoro]

I tried to run the fine-tuning example in examples/finetune.ipynb (without any change) but when I call trainer.train() I get the following error for each training step:

Skipping iteration due to error: The tensor has a non-zero number of elements, but its data is not allocated yet. Caffe2 uses a lazy allocation, so you will need to call mutable_data() or raw_mutable_data() to actually allocate memory.

@urchade Can you help me?