Text only testing out Meta's CM3leon Multi-modal model

We were provided access to Meta's latest iteration of the CM3leon pronounced "Chameleon". The 1st model that uses a single tokenizer for images as well as text.
GitHub: github.com/fac...