Hi, Can you please provide more info on the model that was used to finetune and the metrics achieved. Also is there any paper published for this? Thanks, Gokul Raj