Working Notes: a commonplace notebook for recording & exploring ideas.
Home. Site Map. Subscribe. More at expLog.
Misc links for transformers:
Generally seems to be pretty interesting, follows the patterns I expected to see more of
using any compute and smaller models to do anything
followed by using full device access to work around limitations for integrations)
a desire to not be tied to large models
Charlotte talking about the complexity in serving LLaMa3.
— Kunal