Add model card

README.md

@@ -0,0 +1,38 @@
+---
+pipeline_tag: image-text-to-text
+---
+
+# BARD-VL
+
+BARD (Bridging AutoRegressive and Diffusion) is a framework designed to convert pretrained autoregressive Vision-Language Models (VLMs) into decoding-efficient diffusion VLMs (dVLMs). By combining progressive supervised block merging with stage-wise distillation, BARD-VL achieves significant speedups in decoding throughput while maintaining strong multimodal capabilities.
+
+[**Project Page**](https://fudan-generative-vision.github.io/Bard-VL) | [**Paper**](https://huggingface.co/papers/2604.16514) | [**GitHub**](https://github.com/fudan-generative-vision/Bard-VL)
+
+## Model Description
+BARD-VL addresses the inference bottleneck of token-by-token autoregressive decoding by enabling a parallel decoding paradigm. It transfers multimodal capabilities from existing large-scale VLMs (like Qwen3-VL) to a large-block diffusion VLM. Experimental results show that BARD-VL establishes a new state-of-the-art among comparable-scale open dVLMs while achieving up to 3$\times$ decoding throughput speedup compared to the source model.
+
+## Inference
+To use BARD-VL, please follow the installation instructions in the [official GitHub repository](https://github.com/fudan-generative-vision/Bard-VL). You can run inference for image or video understanding using the provided `inference.py` script:
+
+```bash
+python3 inference.py \
+  --model_id fudan-generative-ai/Bard-VL-B4-Mask-4B-Instruct \
+  --block_size 4 \
+  --denoising_steps 4 \
+  --confidence_threshold 0.6
+```
+
+## Citation
+If you find BARD-VL useful in your research, please cite the following paper:
+
+```bibtex
+@article{chen2026bard,
+  title={BARD: Bridging AutoRegressive and Diffusion Vision-Language Models Via Highly Efficient Progressive Block Merging and Stage-Wise Distillation},
+  author={Chen, Baoyou and Xia, Hanchen and Tu, Peng ...},
+  journal={arXiv preprint arXiv:2604.16514},
+  year={2026}
+}
+```
+
+## Acknowledgements
+This repository builds on top of [NVIDIA NeMo AutoModel](https://github.com/NVIDIA-NeMo/Automodel).