Hello,
I recently submitted my model, which successfully ran through the synthetic data, and got an error that it did not pass through the UW data. The error message I got in my email was that "Submission 'person_id' does not have scores for all goldstandard patients." Did anyone else get this error? I can't figure out why it would work fine on the synthetic data but not the real data (I know that the goldstandard.csv file does not exist in the challenge's data and I did not use that file). I also checked the error logs and couldn't figure out where the issue was.
Created by Emma Becker becker.743 Thank you so much for all your help @trberg , I made those changes and it was successfully submitted!! Another possibility is that your code `read.csv("data/measurement.csv")` is trying to access a data folder in the same working directory, when mapping `-v /synthetic_data:/data` actually mounts a volume to the root directory. So you should try `read.csv("/data/measurement.csv")` and see if that works.
Thanks,
@trberg Thank you so much, @trberg ! This is how I've been running the command, from the example:
```
docker run \
-v $(pwd)/synthetic_data:/data:ro \
-v $(pwd)/output:/output:rw \
-v $(pwd)/scratch:/scratch:rw \
awesome-covid19-q1-model:v1 bash /app/infer.sh
```
And just now I tried:
```
docker run -v /path/to/the/synthetic_data:/data awesome-covid19-q1-model:v1 bash infer.sh
```
I tried copying an empty /data directory to my Dockerfile, but both run commands still give the error that there is no such file or directory when trying to load data/measurement.csv onto the model.
I am further looking into how to create volumes in Docker, in case there is something I am missing.
Thank you so much again,
@becker.743 Hi @becker.743,
No problem, happy to help! How are you running your docker container locally? Does it look something like this?
```
docker run -v /path/to/synthetic_data:/data name_of_container:tag bash train.sh
```
The data should not be in your container when you run it, it should be mounted as a volume from outside the container to the `/data` directory.
Let me know if that is not the issue,
Thanks,
@trberg
I'm so sorry to keep asking @trberg , but I cannot get my model to run and generate predictions without adding the synthetic data to the container, even when I use the "tar xvf synthetic_data.tar.gz" command before running. I also cannot get it to run after changing the data loading commands from read.csv("synthetic_data/measurement.csv") to read.csv("data/measurement.csv") like it is in the example model, even after changing the name of the synthetic_data.tar.gz file. Am I just using the wrong data altogether, and I should be adding something different to the container? I've tried to follow the tutorial and example exactly. Thank you so much @trberg for your help, I can't tell you how much I appreciate it!! Hi @becker.743,
We took a deeper dive into your model and it looks like the container has the synthetic data inside and is making predictions on the synthetic person table rather than the real data that is mapped into the `/data` folder. You should just be using the data in the `\data` folder. We'd also recommend removing the synthetic data from the container. You can see the [tutorial for more information](https://www.synapse.org/#!Synapse:syn21849255/wiki/602419).
Thank you and I hope this helps!
@trberg Hi @becker.743 ,
That error is thrown when the prediction file your model output is either missing person_ids for which we are expecting predictions or when the file has too many or non-existent person_ids. I looked into the output prediction file and there were more person_ids than we were expecting. Is your model making predictions for all patients in the person.csv file?
Thank you,
Tim