1. This version is the latest supported version running on LazyCat Server CPU. If you have purchased an ai pod, we recommend using the version provided by the pod. If you encounter any issues during usage, please contact our official support team for troubleshooting. Thank you. 2. A default environment variable HF_ENDPOINT=https://hf-mirror.com has been added. If you are unable to download models in your current environment, please contact us promptly. 3. Open-WebUI uses its own account system. During first login, you will need to configure your account and password (this is unrelated to the LazyCat Server account system). 4. LzcOllama does not include any models by default. You must manually download models via the interface, such as qwen3:0.6b. 5. During the initial query, the model needs to be loaded and compiled, which takes some time. Please wait a moment. Once the model is loaded, subsequent queries will proceed normally. 6. Ollama's concurrency is disabled by default, as is the auto-completion feature in Open-WebUI's input box. 7. If the model download times out or progresses slowly, you can click the X on the right to cancel the download and then restart the download. 8. Other applications using the Ollama interface can access it via https://lzcollama.$boxname.heiyu.space/lzcollama
