Closed gumityolcu closed 1 month ago
1- select_idx device is set everywhere 2- models are sent to devices after training to prevent any errors. #176 is open to decide for a better fix 3- benchmark tutorials which uses all 3 initialization strategies for 3 different downstream tasks
1- select_idx device is set everywhere 2- models are sent to devices after training to prevent any errors. #176 is open to decide for a better fix 3- benchmark tutorials which uses all 3 initialization strategies for 3 different downstream tasks