Thank you for your work. However, could you provide more detailed information about the environment configuration?

[TianYaDY]

In fact, I spent about 3 hours configuring the environment, and these environment issues always appeared only after the LLM tokens had been consumed. First, RMBG-2.0 reported an error; I installed the corresponding dependencies as prompted and then reran it. Then another error occurred. I tried downgrading numpy, downgrading transformers, downgrading various things, and finally was informed that the older CUDA version cannot be used with the RTX5090; I could only continue by disabling the 5090 and relying on the CPU for computation.
If you could provide a detailed environment setup tutorial, that would be great!

[ResearAI]

Stay tuned for updates.

[TuchuanLin]

In fact, I spent about 3 hours configuring the environment, and these environment issues always appeared only after the LLM tokens had been consumed. First, RMBG-2.0 reported an error; I installed the corresponding dependencies as prompted and then reran it. Then another error occurred. I tried downgrading numpy, downgrading transformers, downgrading various things, and finally was informed that the older CUDA version cannot be used with the RTX5090; I could only continue by disabling the 5090 and relying on the CPU for computation. If you could provide a detailed environment setup tutorial, that would be great!

Thanks for the detailed feedback, and sorry for the setup friction.
We’ve now improved this by adding a clearer Docker deployment guide (README.md / README_ZH.md), a more complete .env.example (including RMBG-2.0 + HF_TOKEN requirements), and troubleshooting notes for dependency/network issues. Hope this makes the setup smoother and improves your experience going forward^_^