Store trained models without wandb as artifacts

Hello wandb community!

I have a large collection of trained models (with pytorch-lightning), but unfortunately without using wandb back then. So, I am wondering if there is a way to store them as artifacts in a model registry.

Thanks in advance! :slight_smile:

Hi Dimitris, you can definitely create a new model collection, log models and then link them to the collection via Model Registry. Please refer to a detailed guide on how to do so here.

Note: A companion notebook is also provided to walk you through all these steps in detail. Please let us know in case of any further questions.

Hi @wdika , I wanted to follow up on this request. Please let us know if we can be of further assistance or if your issue has been resolved.

Hi @anmolmann, thanks for your kind reply! I have seen your comment, but unfortunately I didn’t have the time yet to go through it. Eventually, I haven’t read the section from the documentation you pointed me to. I will catch up as soon as I find some time to work on this :slight_smile:

1 Like