SigLIP-base is a medium-sized multimodal model developed by Google, built on the SoViT (Shape-optimized Vision Transformer) architecture and trained using Sigmoid Loss instead of the contrastive loss used in CLIP. This training approach improves performance in small-batch settings and enhances robustness to negative samples. SigLIP-base achieves strong results in tasks such as image-text retrieval and zero-shot image classification. With solid inference efficiency and scalability, it is well-suited for multilingual and multitask vision-language applications.

Source model

• Input shape: [1x3x384x384], [1x64]
• Number of parameters: 88.86M, 105.16M
• Model size: 359.10M, 424.01M
• Output shape: [1x768], [1x768]

The source model can be found here

When the user has fine-tuned the source model, the model conversion process must be performed again.

Users can refer to either of the following two methods to complete the model conversion:

• Using AIMO for model conversion: Click Model Conversion Reference in the Performance Reference section on the right to view the conversion steps.

• Using Qualcomm QNN for model conversion: Please refer to the Qualcomm QNN Documentation.

The model performance benchmarks and example code provided by Model Farm are all implemented based on the APLUX AidLite SDK.

For models in .bin format, you can use either of the following two inference engines to run inference on Qualcomm chips:

Inference using APLUX AidLite: please refer to the APLUX AidLite Developer Documentation

Inference using Qualcomm QNN: Please refer to the Qualcomm QNN Documentation

Inference Example Code

The inference example code is implemented using the AidLite SDK.

Click Model & Code to download the model files and the inference code package. The file structure is as follows:

/model_farm_{model_name}_aidlite
    
    |__ models # folder where model files are stored

    |__ python # aidlite python model inference example

    |__ cpp # aidlite cpp model inference example

    |__ README.md