v1.1.1

This is a patch release that fixes a crash in the speculative_orig example. The cause of the crash was that the KV cache keep functions were not sending a transaction start command, leading to an unknown operation error.

v1.1.0

Add outputting results to CSV files for easier analysis. Also restores the original speculative inference example as the speculative_orig example for easier comparisons. Finally, a script was added to aid in plotting results.

v1.0.0

Version 1 of PipeInfer. This version corresponds to the CPU-only implementation used in our paper. GPU support will be added in a future release.

Full Changelog: https://github.com/AutonomicPerfectionist/PipeInfer/commits/v1.0.0