Hi, how are we supposed to train our models, given that we cannot download the data?
Are computational resources provided?
P
Created by Pablo Meyer jeriscience Hi @yuanfang.guan,
I see your request and we are in the process of creating a space for your team. I'll let you know if we need any additional information from you.
Thank you,
@trberg @trberg
"The data is hosted within an instance of Palantir's Foundry. When you get access, you'll have resources and tools available to design and train your models within that environment."
When do we expect to get access to this instance?
"Are you onboarded in the N3C enclave? That's step one. "
We onboarded and applied for access (twice) quite a while ago. Never got a reply. What should we expect after we send data access request
thank you Hi @trberg
DPM360 is not a simple Python package. So it is not available on conda/pytorch.
It requires a cluster of machines to be available in a cloud against which the installer could be run to deploy the various components of the DPM360 (OHDSI tooling, service builder etc).
Thanks for trying this, let me loop the team and get instructions on how to install. Hi @jeriscience,
Is the DPM360 package available through conda? The github has a [conda yaml file](https://github.com/IBM/DPM360/blob/main/lightsaber/conda.yaml) but when I tried to use conda to download it, it doesn't seem to be on any of the channels (pytorch, conda-forge, default).
Should this currently be available on conda?
Hi @jeriscience,
I've started the process of getting that package added into the enclave. I'll keep you updated as progress is made.
Thanks!
@trberg This is what we want to install in order to cohort and train/run our models: https://github.com/IBM/DPM360
Thanks Are you onboarded in the N3C enclave? That's step one. What software would you like to install? Thanks, I know we can install software on the environment, what are the steps to do so?
thanks
Pablo Hi @jeriscience,
The data is hosted within an instance of Palantir's Foundry. When you get access, you'll have resources and tools available to design and train your models within that environment.
Thank you,
@trberg