README.md

metadata

license: apache-2.0
datasets:
  - Dauka-transformers/Compact_VLM_filter_data
language:
  - en
base_model:
  - Qwen/Qwen2-VL-2B-Instruct

Qwen2VL Fine-Tuned for Filtration Tasks

This model is a fine-tuned version of Qwen/Qwen2-VL-2B-Instruct trained to perform filtration-oriented image-text evaluation, based on our custom dataset.

🔍 Intended Use

The model is designed to:

• Evaluate alignment of image and caption
• Provide justification scores for noisy web-scale data
• Support local deployment for cost-efficient filtering

🏋️ Training Details

• Base model: Qwen/Qwen2-VL-2B-Instruct
• Fine-tuning objective: in-context scoring + justification
• Dataset: ~4.8K samples with score, justification, text, and image

📁 Files

• model.safetensors – fine-tuned weights
• processor – image and text processor
• README.md – this card

🤝 Acknowledgements

Thanks to the Qwen team for open-sourcing their VLM models, which serve as the foundation for our filtration-oriented model.

📜 License

Licensed under the Apache License 2.0.