Oleg Serikov

I propose to allow for model_config, model, and tokenizer to be optional arguments to the experiment class, rather than setting them post-factum. Like, you either simply pass the model's name,...

А почему в одном месте `device` https://github.com/AIRI-Institute/Probing_framework/blob/a3d382811b49e9a858ed13c552404f481559b6fc/scripts/probing_experiment.py#L29 , а в другом `classifier_device` https://github.com/AIRI-Institute/Probing_framework/blob/a3d382811b49e9a858ed13c552404f481559b6fc/scripts/big_bloom_multiGPU.py#L71

https://github.com/deeppavlov/dream/blob/c779fb9aa61ec48953760c8c6269b625da6360d2/annotators/vidchapters_service/server.py#L103 Let's move this out of the function