RemoveColumnsCollator is removing all keys from variable “features”. No matter what model I choose or which dataset I choose. The input to the function transformers.trainer_utils._remove_columns is a tokenized string which is a dictionary with “input_ids” and “labels” as keys. But output is always N…

RemoveColumnsCollator is removing all columns

sheaako August 12, 2024, 6:34pm 5

I had a similar issue.
They key for me was to add this to my TrainingArguments:

training_args = TrainingArguments (
    remove_unused_columns  = False,
     ...
     )

Topic		Replies	Views
Remove columns before training Beginners	0	845	January 4, 2023
Creating Trainer object is deleting my 'labels' feature Beginners	3	1460	January 21, 2021
Non-label features are not passed into data collator Beginners	6	58	January 1, 2025
Remove columns from streamable datasets doesn't work 🤗Datasets	3	6254	January 24, 2024
The model did not return a loss from the inputs, only the following keys: logits. For reference, the inputs it received are input_values Intermediate	23	41283	April 23, 2025