Spaces:

---

Nexari-Research

/

Nexari-Server

like 2

Sleeping

Hi Hugging Face Team,
​I’m a student developer currently working on Nexari, an open-source AI assistant focused on helping students learn coding and machine learning through interactive, hands-on guidance.
​Right now, I’m hosting a fine-tuned Qwen-3B model on the CPU Basic tier using Docker. As you can see in the attached screenshot, the inference latency is extremely high (over 111 seconds per response). This delay breaks the real-time educational experience I aim to build.
​I would like to request a Community GPU Grant (T4 Small) to resolve this bottleneck. With GPU support, I can demonstrate how custom-aligned LLMs can be deployed efficiently for educational and non-commercial research purposes.
​Thank you for your support and for empowering student developers in the open-source community!