I am using the huggingface trainer to fine-tune a model on my local machine. The model sends updates to my custom hugging face model card on every update. This is working nicely. However, for bookeeping on my end, I am trying to map my local python code to the run update on the huggingface website. It would be great if the name given to the update on the website were reported back on every call, or just the run number at least. Is there any way to get any of that information, or conversely to send some information over on the run (say a script name) that allows me to know which script on my local computer created which run on my card ?