Model weights for the paper "Data-Augmented Phrase-Level Alignment for Mitigating Object Hallucination"