Edit model card

Meet Nexus Nova, the spacecraft salesman! CEO Of Nova Space Ships, former cargo hauler, and family man. He's here to offer you the best deals in this corner of the galaxy. Nex is ready to answer all your questions and help you find the best ship for your budget

Trained for 9 epochs on 1100 lines of custom dialogue, written partially by hand and partialy with the help of GPT-3.5-turbo. Because of this, this model is not suitable for commercial use and is offered to the public soley for personal enjoyment and experimentation.

Uses

This model is designed to be a conversational roleplaying chatbot for a home made tabletop rpg campaign. It is not intended to answer general questions or provide factual information.

Direct Use

This model is intended to be ran on low end hardware and hosted as a chatbot. Conversations are intended to be at most 25 turns in total.

Out-of-Scope Use

This model is only intended for short role-play conversaions, and fails to perform adequatley at any other tasks.

Bias, Risks, and Limitations

This model is a fine tuned variant of Microsoft's DialoGPT, so it likely has the same inherit biases as thier base model. It may create harmful outputs if pushed to, but is highly unlikely to unless prompted to. This model is only fine tuned on hyper specific dialogue scenerios, and is severley limited in its ability to hold conversations outside of its domain (spaceship sales). This model does exhibit short term memory issues when not reminded of which ship it is talking about, and may also sometimes hallucinate new ships, prices, features, etc that were not part of its training.

Recommendations

This model should not be used in any commercial or public settings, or outside of its intended domain as an entertainment/experiment

Downloads last month
22
Safetensors
Model size
380M params
Tensor type
F32
·
U8
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.