Ive tried for about 3 hours to upload training data to Huggingface Autotrain and i used this video - Train tabular models automatically with Hugging Face AutoTrain - YouTube
It is now irrelevant as the autotrain is looking for some format that doesnt exist.
I have tried uploading a CSV and JSONL formatted dataset and I get this error- Error type: ValidationError
Details: Can’t split automatically: Invalid number of samples for label “10am EST”, “upcoming monday through Wednesday” in column ‘target’. Auto-splitting requires at least 2 samples for each label, but “10am EST”, “upcoming monday through Wednesday” only has 1 samples
Every target column has at least 2 examples in them, which is not mentioned in the above video. I have formatted that column about 6 different ways to no avail:
Please someone tell me, what needs to be done here? I have tried this target table formats:
- [tuesday at 5, tuesday 5pm]
- tuesday at 5, tuesday 5pm
- [“tuesday at 5”, “tuesday 5pm”]
- “tuesday at 5, tuesday 5pm”
- “tuesday at 5”, “tuesday 5pm”
i have also tried multiple target columns… must be glitchy with the auto-splitter?
Thanks - Jon