Search results

Script to calculate Correlation between automated metrics and human preference rankings. (Table 3 of paper)

Status: Open.

#33 In MizzenAI/HPSv3;

· ritikajha opened on Feb 6, 2026

HPSv3 ignore prompt

Status: Open.

#32 In MizzenAI/HPSv3;

· weathon opened on Jan 13, 2026

Emotion Bias

Status: Open.

#31 In MizzenAI/HPSv3;

· weathon opened on Dec 18, 2025

Is it only evaluating the image without any prompts?

Status: Open.

#30 In MizzenAI/HPSv3;

· cyy-1234 opened on Dec 18, 2025

How is the confidence of each annotator obtained?

Status: Open.

#28 In MizzenAI/HPSv3;

· g-h-chen opened on Dec 1, 2025

release train code use hpsv3 as reward for rl-training

Status: Open.

#26 In MizzenAI/HPSv3;

· akk-123 opened on Oct 29, 2025

How to get the returned uncertainty?

Status: Open.

#25 In MizzenAI/HPSv3;

· weathon opened on Oct 25, 2025

Famous Painting being rated very low

Status: Open.

#24 In MizzenAI/HPSv3;

· weathon opened on Oct 24, 2025

Validation set accuracy

Status: Open.

#23 In MizzenAI/HPSv3;

· chenxxc opened on Oct 15, 2025

The download pretrain model script used a default mirror that is not working

Status: Open.

#22 In MizzenAI/HPSv3;

· weathon opened on Oct 14, 2025

Access Request for HPDv3 Test Benchmark

Status: Open.

#21 In MizzenAI/HPSv3;

· jaxfield opened on Oct 3, 2025

Please check the PartialEmbeddingUpdateCallback

Status: Open.

#20 In MizzenAI/HPSv3;

· gaoyixuan1025 opened on Sep 25, 2025