learn2learn integration

[Rusteam]

Hi there,

I'm trying to adapt learn2learn integration examplt to a fiftyone dataset

# data
import fiftyone as fo
dataset = fo.load_dataset("quickstart")
dataset = dataset.filter_labels("resnet50", fo.ViewField("label").is_in(["parachute", "sandbar", "breakwater", "umbrella", "zebra"]))
datamodule = ImageClassificationData.from_fiftyone(
    train_dataset=dataset,
    label_field="resnet50",
    val_dataset=dataset,
    batch_size=4,
    # train_transform=transform,
)
# model
model = ImageClassifier(
    backbone="resnet18",
    training_strategy="prototypicalnetworks",
    training_strategy_kwargs={
        "epoch_length": 2,
        "meta_batch_size": 4,
        "num_tasks": 200,
        "test_num_tasks": 5,
        "ways": datamodule.num_classes,
        "shots": 1,
        "queries": 1,
        "test_ways": 5,
        "test_shots": 1,
        "test_queries": 1,
    },
    optimizer=torch.optim.Adam,
    learning_rate=0.001,
)
# trainer
trainer = flash.Trainer(
    max_epochs=2,
    gpus=1,
    precision=16,
)

if __name__ == '__main__':
    trainer.finetune(model, datamodule=datamodule, strategy="no_freeze")

however, I'm getting this error:

 File "/home/rustynail/miniconda3/envs/finegrained/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 570, in _next_data
    data = self._dataset_fetcher.fetch(index)  # may raise StopIteration
  File "/home/rustynail/miniconda3/envs/finegrained/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 32, in fetch
    data.append(next(self.dataset_iter))
  File "/home/rustynail/miniconda3/envs/finegrained/lib/python3.9/site-packages/flash/image/classification/integrations/learn2learn.py", line 69, in __next__
    for item in self.tasks.sample():
  File "learn2learn/data/task_dataset.pyx", line 158, in learn2learn.data.task_dataset.CythonTaskDataset.sample
  File "learn2learn/data/task_dataset.pyx", line 169, in learn2learn.data.task_dataset.CythonTaskDataset.__getitem__
  File "learn2learn/data/task_dataset.pyx", line 142, in learn2learn.data.task_dataset.CythonTaskDataset.get_task
  File "/home/rustynail/miniconda3/envs/finegrained/lib/python3.9/site-packages/flash/image/classification/adapters.py", line 55, in remap
    data[DataKeys.TARGET] = mapping(data[DataKeys.TARGET])
  File "learn2learn/data/transforms.pyx", line 213, in learn2learn.data.transforms.RemapLabels.__call__.mapping
ValueError: 0 is not in list

I've checked that RemapLabels class but I can't figure out how mapping function works considering that fact that self.dataset object contains all required mappings. Checked this but there's no sufficient details on this.
My gut feeling tells me that it's because labels are str unlike the example here where labels are integers