README.md

Prot_bert finetuned on GPCR_train dataset of Drug Target prediction 

Trainig paramenters:
  overwrite_output_dir=True,
  evaluation_strategy="epoch",
  learning_rate=1e-3,
  weight_decay=0.001,
  per_device_train_batch_size=batch_size,
  per_device_eval_batch_size=batch_size,
  push_to_hub=True,
  fp16=True,
  logging_steps=logging_steps,
  save_strategy='epoch',
  num_train_epochs=2