search

tomaarsen / attention_sinks

Extend existing LLMs way beyond the original training length with constant memory usage, without retraining
https://huggingface.co/blog/tomaarsen/attention-sinks
Apache License 2.0
650 stars 41 forks source link

Add support for StableLM 3b 4e1t model #20

Closed kmn1024 closed 9 months ago

kmn1024 commented 9 months ago

To resolve #19 (add support for StableLM 3b 4e1t model).

attention_sink_stablelm

tomaarsen commented 9 months ago

This is awesome! Thanks a bunch. Exactly what I'm looking for in these PRs. I've added the model info to the README & added a "Model Contributors" section in the README with some credits.