README.md

---
datasets:
- Norquinal/claude_multiround_chat_1k
- jondurbin/airoboros-gpt4-1.4
- Squish42/bluemoon-fandom-1-1-rp-cleaned
- totally-not-an-llm/EverythingLM-data-V2-sharegpt
- OpenLeecher/Teatime
- PygmalionAI/PIPPA
tags:
- not-for-all-audiences
- nsfw
license: cc-by-nc-4.0
---
## What is PetrolLM?
PetrolLM is [Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) model fine-tune using QLoRA (4-bit precision) for the purposes of creative writing and roleplay. 

The dataset consists of 5800 samples, with the composition as follows:
* AICG Logs (~17%)
* PygmalionAI/PIPPA (~17%)
* Squish42/bluemoon-fandom-1-1-rp-cleaned (~13%)
* OpenLeecher/Teatime (~2%)
* Norquinal/claude_multiround_chat_1k (~17%)
* jundurbin/airoboros-gpt4-1.4 (~17%)
* totally-not-an-llm/EverythingLM-data-V2-sharegpt (~17%)

These samples were then back-filled using gpt-4/gpt-3.5-turbo-16k or otherwise converted to fit the prompt format.

## Prompt Format
The model was finetuned with a prompt format similar to the original SuperHOT prototype:
```
---
style: roleplay
characters:
  [char]: [description]
summary: [scenario]
---
<chat_history>
Format:
[char]: [message]
Human: [message]
```

## Use in Text Generation Web UI
Install the bleeding-edge version of `transformers` from source:
```
pip install git+https://github.com/huggingface/transformers
```
Or, alternatively, change `model_type` in `config.json` from `mistral` to `llama`.

## Use in SillyTavern UI
![](https://files.catbox.moe/2dkr28.png)
As an addendum, you can include one of the following as the `Last Output Sequence`:
```
Human: In your next reply, write at least two paragraphs. Be descriptive and immersive, providing vivid details about {{char}}'s actions, emotions, and the environment.
{{char}}:
```
```
{{char}} (2 paragraphs, engaging, natural, authentic, descriptive, creative):
```
```
[System note: Write at least two paragraphs. Be descriptive and immersive, providing vivid details about {{char}}'s actions, emotions, and the environment.]
{{char}}:
```
The third one seems to work the best. I would recommend experimenting with creating your own to best suit your needs.

## Finetuing Parameters
- LoRA Rank: 64
- LoRA Alpha: 16
- LoRA Dropout: 0.1
- BF16 Training
- Cutoff Length: 2048
- Training Epoch(s): 2