Hi,
I had raised a similar issue on Github. Got the following solution. They’d suggested to use config_name:split_name as the train/validation split. Linking the specific comment below:
opened 10:05AM - 16 May 24 UTC
closed 10:28AM - 21 May 24 UTC
bug
### Prerequisites
- [X] I have read the [documentation](https://hf.co/docs/auto… train).
- [X] 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