# ai-msgbot GPT2-L

_NOTE: model card is WIP_

GPT2-L (774M parameters) trained on [the Wizard of Wikipedia dataset](https://parl.ai/projects/wizard_of_wikipedia/) for 40k steps with 34/36 layers frozen using `aitextgen`. 


Designed for use with [ai-msgbot](https://github.com/pszemraj/ai-msgbot) to create an open-ended chatbot (of course, if other use cases arise have at it).


## conversation data

The dataset was tokenized and fed to the model as a conversation between two speakers, whose names are below. this is relevant for writing prompts and filtering/extracting text from responses.

`script_speaker_name` = `person alpha`

`script_responder_name` = `person beta`


## examples

- the default inference API examples should work _okay_
- an ideal test would be explicitly adding `person beta` to the **end** of the prompt text. The model is forced to respond to the entered chat prompt instead of adding to the entered prompt and then responding to that (which may cut off the response text due to the Inference API limits).