Some question related to environment configuration. #33
Description
- I would like to know how the official environment is configured for the purpose of performance comparison.
- Based on statements such as:
“To allow fair comparison, our model is trained on a subset of the MetaUrban-train dataset, and is tested on 1,000 scenes in the MetaUrban-test dataset and 100 scenes in the MetaUrban-unseen dataset,”
I would like to ask whether there is a separate mechanism for controlling this via scenario IDs, and if so, how the environment should be configured accordingly.
I look forward to your response. Thank you.