Extend context to 262144 with YaRN rope scaling

#16

by logan-vegna-shopify - opened Apr 15

base: refs/heads/main

←

from: refs/pr/16

Discussion Files changed

+10

-7

Extend context to 262144 with YaRN rope scalingd92fdebc

logan-vegna-shopify

Apr 15

Extend max_position_embeddings from 128K to 262144 (256K) using YaRN rope scaling (factor=2.05). This enables vLLM to serve the model with longer context windows.

Patch tokenizer_class to PreTrainedTokenizerFast for vLLM/cserve compatibilitybfdfafb2

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

Ready to merge

This branch is ready to get merged automatically.

· Sign up or log in to comment