Failed when saving task, and error neither pass nor fail when evaluate

[Cedsbstn]

The following list consists of impacted links:

• https://www.kaggle.com/code/cedsbstn/three-body-problem
• https://www.kaggle.com/benchmarks/tasks/cedsbstn/calculate-sun-temperature

Error message when evaluate the task: "The task code doesn't contain any statement to use the kbench.llm model, this usually means that you hard-coded a model in the code. Please revise the task code to use kbench.llm instead. More examples: https://github.com/Kaggle/kaggle-benchmarks/blob/ci/cookbook.md"

[yibinlin-google]

Hello @Cedsbstn thanks for reporting the bug. I took a look at https://www.kaggle.com/benchmarks/tasks/cedsbstn/calculate-sun-temperature and the error didn't show up in the latest version of the task. Did you manage to resolve it? Thanks.

[Cedsbstn]

Hi @yibinlin-google, thanks for your response. I did, but the problem happened when I chose to write a task instead of generating a task. I resolved it by creating a new task, choosing the generate a task option, and it was able to save the task as well. After that, I downloaded the notebook and imported it into the task, and then I adjusted the prompt, code, and configuration, so the problem was resolved. Thanks.