Improve model card

README.md

@@ -1,3 +1,23 @@
----
-license: mit
----
+---
+license: mit
+pipeline_tag: zero-shot-image-classification
+---
+
+# SVL: Spike-based Vision-language Pretraining for Efficient 3D Open-world Understanding
+
+This repository contains the model weights for **SVL**, a Spike-based Vision-Language pretraining framework that empowers Spiking Neural Networks (SNNs) with open-world 3D understanding while maintaining spike-driven efficiency.
+
+- **Paper:** [SVL: Spike-based Vision-language Pretraining for Efficient 3D Open-world Understanding](https://huggingface.co/papers/2505.17674)
+- **Code:** [https://github.com/bollossom/SVL](https://github.com/bollossom/SVL)
+
+## Introduction
+
+SVL introduces two key components:
+1. **Multi-scale Triple Alignment (MTA)** for label-free triplet-based contrastive learning across 3D, image, and text modalities.
+2. **Re-parameterizable Vision-Language Integration (Rep-VLI)** to enable lightweight inference without relying on large text encoders.
+
+The framework effectively bridges the performance gap between SNNs and Artificial Neural Networks (ANNs) in complex tasks such as zero-shot 3D classification, multimodal question answering, and 3D detection/segmentation.
+
+## News
+- **Jun. 29, 2025**: Weights made available on Hugging Face.
+- **Jun. 24, 2025**: Release of code for training and testing.