llama2.c calculation in FP16 #539

Open

Open

llama2.c calculation in FP16#539

Is it possible to replace all calculation in float32 with float16?
How would this impact accuracy? Is this attempted before?

to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Assignees

No one assigned

Labels

No labels

No labels

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests