Assuming the /input volume will contain the validation/test data for the leadership board and beyond, is there a specific file name and format we should be looking for when loading that? For example, should we be retrieving input/inputdata.h5ad or so? The docker template looks for /input/data.csv but I'm thinking that's not applicable here. We found documentation for the output file name and format, so no questions there at this point.
Thank you
Created by Phillipe Loher PhillipeLoher Ok thanks for the clarification now i understood.
No, we expect you to give us the trained model. In other words, the docker should be able to directly provide the predictions from the validation data and the test data. Sage will change the data input accordingly to test your model's ability to predict the validation dataset and the test dataset. However, we will not provide the training datasets for you to train on the Sage server during validation and testing if that makes sense. Hope this helps. So in /input volume we can look for 3 files right ? including given 2 training dataset .
Thank you. Hi Phillipe,
Thanks for catching it. The input file name will be input/data.h5ad. We will edit the description of the input data on the website accordingly. Thanks.