Spaces:
Running
on
Zero
Apply for community grant: Academic project (gpu)
Dear Hugging Face Team,
I am writing to request a Zero GPU grant for my Hugging Face Space, which hosts an interactive demo of our recent research paper, ZIP: Scalable Crowd Counting via Zero-Inflated Poisson Modeling.
Our work introduces a novel framework for crowd counting that achieves state-of-the-art results by modeling the sparse, discrete nature of crowd data using a Zero-Inflated Poisson (ZIP) distribution. This approach significantly improves accuracy over traditional methods, particularly on challenging benchmarks like ShanghaiTech A & B, UCF-QNRF, and NWPU-Crowd.
The goal of our Space is to provide the community with a live, interactive demonstration of this cutting-edge model. Users can upload their own images and see the model's predictions in real-time, which we believe is a powerful way to showcase the effectiveness and practical utility of our research.
Currently, the demo runs on a CPU, which results in slow inference times, especially for the larger, more accurate variants of our model (like ZIP-B). This latency significantly hinders the user experience and does not accurately reflect the model's potential for real-world, near-real-time applications. A GPU is essential to enable real-time inference: Provide immediate feedback to users, making the demo more engaging and impressive.
We believe that providing a fast, accessible demo will greatly benefit the computer vision community and highlight the power of the Hugging Face platform for sharing cutting-edge research.
Thank you for considering our application. We are excited about the possibility of leveraging the Zero GPU program to enhance our project's impact.
Best regards,
Yiming
