JonathanBhimani-Burrows

Thanks! Turns out that the architecture of the two was slightly differently implemented, so it wasn't possible to load the weights (not just different state dict names, but entirely different...

Re-opening this Trained a model with Pytorch tabular (FT Trans) and saved _file_path = os.path.join(out_path, 'best_model') print(f"Saving Best Model at {file_path}") tabular_model.save_model(file_path)_ Attempting to reload the model (checking if weights...

Did you re-initialize the model in between these tests? Otherwise, the best model would be already loaded and you wouldn't see this issue I had to save, re-initialize and reload,...

Step 1 and 2 were for replication information Regarding your workflow, that would make sense, thank you for clarifying! I had in my mind the regular pytorch approach( not lightning)...

commenting here to support these suggestion - I run out of memory with 1 mil rows/10 features on a 40 GB GPU is there a reason that there isn't a...

Thanks for the reply, but this didn't really answer the question The decision of making both first name and primary address with pii = False, was a design decision. I...

Any update on this? Is this a mistake on our side, or is this expected?

We have several models fpr our clients that use the older version For us to do a full migration, we would need to re-run all models for all clients to...