Allow to `finetune` GPU-trained models on CPU

[StpMax]

At the moment model that trained on GPU can be finetuned only on GPU. Will be good to be able make it on CPU.

Currently error is:

RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.

[yuhuishi-convect]

Seems like duplicates of https://github.com/mindsdb/mindsdb/issues/6949

[paxcema]

Closing (solved in mindsdb, here it is a duplicate of #1129)