This notebook will need high ram and a premium GPU. It blew past the 16gb GPU that comes with Colab Pro
Your session crashed after using all available RAM. If you are interested in access to high-RAM runtimes, you may want to check out Colab PRO
You can try 8bit
Using 8-bit still runs out of RAM for both the 3B and 7B models. It's unclear if it's because it still uses more than the available RAM, or if it's just quietly not using 8-bit since it's not implemented.