This is the original AI Dungeon 2 model back from when AI Dungeon 2 was an open source project. The model used for this conversion has been converted prior to the Huggingface Pytorch 16-bit format and was then further quantized to a 4.0 quantization (The original model was a 32-bit Tensorflow model).
The purpose of this model is to allow faster access to the classic AI Dungeon 2 experience, for the sake of nostalgia or for use on low end hardware. It is best used with Koboldcpp and was tested on the Windows 7 version of Koboldcpp 1.25.
Inference Providers
NEW
This model is not currently available via any of the supported third-party Inference Providers, and
HF Inference API was unable to determine this model's library.