model-epoch-50
results.parquet
image → text

prediction
Judge the following image on a few different dimensions. Be very critical. Each of judgements should be one of three values: * "poor" if the image does not match the description * "good" if the image matches the description, but could be better * "great" if there is nothing that could be improved about the image Return the judgements in xml format. The xml should contain with the following field names: <reasoning> Step by step reasoning of why the image is good or not </reasoning> <style> Is the character in the style specified? Is the fur pure white with a nice texture? Is the character friendly? </style> <task> How well is the task being portrayed? Are all the items and actions present? </task> <quality> What is the overall quality of the character? Are there any defects such as too many legs, eyes not open, etc? </quality> Prompt: {prompt} Image: {image} Reason through your thoughts step by step before responding. Put your thoughts in the <reasoning></reasoning> tags.
May 29, 2025, 6:08 PM UTC
May 29, 2025, 6:09 PM UTC
5 row sample
8358 tokens$ 0.0042
5 rows processed, 8358 tokens used ($0.0042)
Estimated cost for all 50 rows: $0.0416Sample Results completed
3 columns, 1-5 of 50 rows