v1.1.0

更新

• 支持OCRBench, OCRBench-v2, DocVQA, InfoVQA, ChartQA, BLINK 等图文多模态评测基准，所有支持的数据集请参考
• 编写Qwen3-Omni和Qwen3-VL模型评测最佳实践
• 支持pyproject.toml安装

Update

• The platform now supports OCRBench, OCRBench-v2, DocVQA, InfoVQA, ChartQA, BLINK, and other multimodal evaluation benchmarks. For a comprehensive list of supported datasets, please refer.
• Developed best practice guidelines for evaluating models with Qwen3-Omni and Qwen3-VL.
• Installation via pyproject.toml is now supported.

What's Changed

• [Doc] Add qwen omni doc by @Yunnglin in #854
• [Fix] Fix bfcl_v3 validation by @Yunnglin in #858
• [Feature] Add pyproject.toml by @Yunnglin in #857
• [Benchmark] Add ChartQA and BLINK by @Yunnglin in #861
• [Benchmark] Add DocVQA and InfoVQA by @Yunnglin in #862
• [Fix] transformers import by @Yunnglin in #865
• [Benchmark] Add OCRBench and OCRBench-v2 by @Yunnglin in #869
• [Fix] None string error by @Yunnglin in #871

Full Changelog: v1.0.2...v1.1.0