Can I run Qwen3 in an optimized way on the Hat?

General Discussion
1

I was using qwen3:0.6B and qwen3:1.7B for a project I am working on and I bought the board to decrease latency but I was surprised to not see 3 in the model zoo. Is there a reason the model zoo is behind? Is there a way I can optimize it with my own resources? Can I just use the hailo-ollama with it?

Thanks

1 Like
2

Hi @user449 ,

Thanks for reaching out. We are planning to increase the range of supported models, however for time being we cannot share specific details about timelines and/or models.

”Is there a way I can optimize it with my own resources”: We are planning to release support for LoRA fine tuning for our supported models.

“Can I just use the hailo-ollama with it”: hailo-ollama works with only our supported models.

Thanks,