[BUG] Config name is missing for Datasets with no default config

[Aryan-401]

Prerequisites

• I have read the documentation.
• I have checked other issues for similar problems.

Backend

Other cloud providers

Interface Used

UI

CLI Command

No response

UI Screenshots & Parameters

Error Logs

ERROR | 2024-05-16 09:52:25 | autotrain.trainers.common:wrapper:120 - train has failed due to an exception: Traceback (most recent call last):
File "/opt/conda/lib/python3.10/site-packages/autotrain/trainers/common.py", line 117, in wrapper
return func(*args, **kwargs)
File "/opt/conda/lib/python3.10/site-packages/autotrain/trainers/object_detection/main.py", line 49, in train
train_data = load_dataset(
File "/opt/conda/lib/python3.10/site-packages/datasets/load.py", line 2587, in load_dataset
builder_instance = load_dataset_builder(
File "/opt/conda/lib/python3.10/site-packages/datasets/load.py", line 2296, in load_dataset_builder
builder_instance: DatasetBuilder = builder_cls(
File "/opt/conda/lib/python3.10/site-packages/datasets/builder.py", line 374, in init
self.config, self.config_id = self._create_builder_config(
File "/opt/conda/lib/python3.10/site-packages/datasets/builder.py", line 584, in _create_builder_config
raise ValueError(
ValueError: Config name is missing.
Please pick one among the available configs: ['full', 'mini']
Example of usage:
load_dataset('license-plate-object-detection', 'full')

ERROR | 2024-05-16 09:52:25 | autotrain.trainers.common:wrapper:121 - Config name is missing.
Please pick one among the available configs: ['full', 'mini']
Example of usage:
load_dataset('license-plate-object-detection', 'full')

Additional Information

Seems like the Problem is due to Non-Default configs in the dataset I'm trying to use. Next logical question is this by design? If yes, how can I use datasets without any default to autotrain?

Links to support Theory:

• https://discuss.huggingface.co/t/need-help-setting-up-autotrain-locally/84225
• https://discuss.huggingface.co/t/config-default-not-found-for-dataset/49752

[abhishekkrthakur]

it was by design but ive added a workaround.
for the dataset you mention, you can use config_name:split_name as split.

for example: full:train in train split
and full:test or full:validation in valid split

make sure you are on version 0.7.99 or above.