Team - Question - How does saving and reusing fine...
# squads
n
Team - Question - How does saving and reusing fine tuned models work for self hosted versions of TimeGPT? [Reference]
m
I don't know, but I'm curious to know! José probably knows the answer 🙂
j
They're saved in the container so they're lost if its deleted. Users can also set an environment variable to save them in a remote storage, but I don't know if there are any privacy concerns because they can look at all the weights, as far as I know @Cristian Challu has only shared that with one customer.
n
Thanks. And how would they reference the stored model later through the SDK?
j
In the same way as they do for our deployment
n
Ok, thanks for confirming