add CUB-200-2011 to CoOp, display error causes an error to run

[KevinLi-167]

add CUB-200-2011 to CoOp, display error causes an error to run

[KevinLi-167]

The CUB200 dataset I generated, the category number is 1-200 instead of 0-199.

And after reading a func train, val, test = OxfordPets.read_split(self.split_path, self.image_dir) that was not used in cal101.

I realized that it might be because my CUB.split.json' catyID index didn't start at 0.

So I tried Simply replace the category id of 200 with 0. Problem solved.

---

To locate this problem, I looked for the content of the print in two projects (CoOp and Dassl).

At CoOp.train.py.main ,'build_trainer'
trainer = build_trainer(cfg)
In this sentence, some functions of Dassl are called

I found in Dassl.pytorch\dassl\engine\trainer.py,
class SimpleTrainer(TrainerBase).__init__(self, cfg):
Someone called his function and passed in the wrong cfg,
and the content in the cfg identifies the number of my category IDs as 201

DataManager.show_dataset_summary:
    table.append(["# classes", f"{self.num_classes:,}"])

Although after a len() function, the output of my category length is 200.

---

To conclusion, I did ignore that cal101's json starts with categoryId=0 which given by CoOp.
But the toolbox's method of identifying lengths is not very smart ( COOP calculates len instead of reading json).