Dear all,
In the resource folder, you can find a docker tutorial, in case it is still not clear.
Please have a look at some scripts from our dry run: https://github.com/Sage-Bionetworks/NCI-CPTAC-Challenge-Examples
Best,
Mi
Created by Mi YANG MI_YANG Dear Artem,
Yes that is correct. You should not be doing any training in your docker container.
Best,
Tom Dear Mi,
Just to double check: am I correct to understand that our Docker container code can simply load pre-trained models from an .RData file (as you've done in your GitHub example), without having to re-run training on the original data files from scratch?
-Artem Dear Irene,
1) In the docker folder, you will include your script which impute the missing value in test data. There is no mentioning of the training data.
2) The assessment will run the program on the test data.
PS: the main difference between sc1 and sc2/3, is that here you only need your imputation codes. whereas in sc2/3, you need to upload your saved models from training.
Best,
Mi Hi, Mi,
Could I please clarify the process for submission.
1) Is this correct that,
For sub1, in the docker file, we will need to include the final prediction results ( predicted value and predicted interval) from the training data which are the 10 obs files, and the program for the algorithm.
2) The assessment will then also include re-run the program on the test data. After this, will the assessment matrices being calculated from the results from test data, or the training data?
thank you,
Irene
Drop files to upload
Docker tutorial and examples of submission scripts page is loading…