This topic lists common problems and solutions for deploying model services using Function AI. If your problem is not listed here, contact us and provide the necessary information, such as the traceID and instanceName.
Why does my deployment fail with the error "Resources are being replenished, please wait"?
This error indicates that the GPU resources in the current region are insufficient. To resolve this issue, switch to a different GPU type or region and then retry.
How do I resolve the "Logon failed, AccessToken error" message?
This error indicates that the AccessToken is incorrect. This token is required when you deploy a model from a custom source, such as ModelScope, OSS, or a custom image. For example, when you use a ModelScope source, you must provide the model ID, version number, and the token. To resolve this error, obtain the correct AccessToken from the User Center and enter it, or refresh the token and then retry.
How do I resolve the "Message: code: 403, Current user is in debt" error?
This error message indicates that your Alibaba Cloud account has an overdue payment. To resolve this issue, go to the Expenses and Costs page to view the overdue payment, add funds to your account, and then retry.
How do I resolve the "Incorrect ModelScopeModelID format" error?
This error message indicates that the entered ModelScope model ID is in an incorrect format. The ID must use the "group/model" format, such as Qwen/Qwen3-32B. For more information, see the Model Library.
How do I resolve a ParseModelFileError or ModelTaskError?
These error messages indicate that the model failed to download. This can happen if the model is in a private repository and a valid ModelScope token was not provided, or if the config.json file is missing.