| license: apache-2.0 | |
| base_model: | |
| - google/siglip-so400m-patch14-384 | |
| pipeline_tag: zero-shot-image-classification | |
| tags: | |
| - renesas | |
| - x5h | |
| - coming-soon | |
| - cv | |
| - vision | |
| # SigLIP-SO400M – Renesas X5H | |
| <p align="center"> | |
| <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/model_doc/siglip_table.jpeg" alt="SigLIP vs. CLIP zero-shot and retrieval benchmark comparison" width="640"> | |
| </p> | |
| <p align="center"><sub>SigLIP vs. CLIP evaluation results, from the original paper — as shown on the official <a href="https://huggingface.co/google/siglip-so400m-patch14-384">google/siglip-so400m-patch14-384</a> model card</sub></p> | |
| 🚧 **Coming soon.** This model is being optimized and validated for Renesas hardware — no fixed release date yet. | |
| ## Introduction | |
| **SigLIP-SO400M** is Google's Sigmoid-Loss vision-language embedding model (the "shape-optimized" 400M-parameter variant), commonly used as a vision encoder for zero-shot image classification and retrieval. Renesas is preparing an optimized deployment of this model for the R-Car Gen5 platform. | |
| - **Model Architecture:** SigLIP vision/text dual encoder (vision tower only is the focus for on-device deployment). | |
| - **Source Model:** [google/siglip-so400m-patch14-384](https://huggingface.co/google/siglip-so400m-patch14-384) | |
| *This page is a placeholder. Content is provisional and subject to change before the model is fully published.* | |