You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This model uses T5 architecture so we firstly need to support T5, see this issue: #5763
Beside that, this model doesn't have a tokenizer (since time-series signal is converted directly into tokens), so convert script may fail. We can simply add a dummy tokenizer before converting to solve that problem. An example of the HF model (without tokenizer) can be found here: https://huggingface.co/amazon/chronos-t5-small/tree/main
Chronos appears to be a great open source time series forecasting model.
Is it possible to integrate this model into llama.cpp?
https://github.com/amazon-science/chronos-forecasting
The text was updated successfully, but these errors were encountered: