Ensure tokenizer is on GPU

Diagnostics

Unpack BatchEncoding

Missing 'from_pretrained'

config didn't need to be a property

There's a simpler way of doing this, I hope

Might be simpler to inherit from RobertaModel rather than PreTrainedModel

Removed a base model that was causing a loop in model initialisation

Further changes for compatibility with HuggingFace Pytorch implementation

PyTorch implementation of HugggingFace PreTrainedModel class does not allow direct setting of base_model. Rejig constructors accordingly

Converted QaracEncoderModel to use PyTorch

Making sure RoBERTa layers have all required arguments

Fixed indentation error

Use base model config in superclass __init__

Fix typos

More work on models

Encoder, Decoder and Trainer models (assuming RoBERTa base models)