metadata
license: apache-2.0
datasets:
- SPUH/FRABench
language:
- en
metrics:
- accuracy
base_model:
- Qwen/Qwen2-VL-7B-Instruct
UFEval-7B
Model Summary
UFEval-7b is the first unified fine-grained evaluator with task and aspect generalization. Built on the foundation of Qwen2-VL-7B-Instruct, it has been finetuned on FRABench dataset.
For further details, please refer to the following resources:
- 📰 Paper: https://arxiv.org/abs/2505.12795
- 🪐 Project Page: https://github.com/ALEX-nlp/UFEval
- 📦 Datasets: https://huggingface.co/datasets/SPUH/FRABench