2.2.2

Minor library version updates, model list updates, etc.

2.2.1

Adds preliminary support for VertexAI chat models.

2.2.0

Basic awareness (temporal, corporeal, sensorial), remove USER/ASSISTANT from rp/gtkm outputs to just use character name directly, misc other updates.

2.1.6

Misc fixes and improvements to LMoE inference, including bettertransformer/flash

2.1.4

More LMoE fixes

2.1.3

Misc LMoE fixes.

2.1.2

LMoE - HF inference (better quality, slower), vllm inference (much faster, much lower quality for some adapters)

2.1.1

LMoE - Dynamic LoRA switching mixture of experts (sort of), built on top of vllm

2.0.24

More chat updates.

2.0.23

Misc fixes.