---
library_name: transformers
tags:
- axolotl
- qlora
- not-for-all-audiences
license: apache-2.0
language:
- en
---
Atlantis-v0.1-12B
Props to invisietech for the model. This is the 4BPW version of it. [Find the original model here](https://huggingface.co/invisietch/Atlantis-v0.1-12B)
[Find the 6bpw version quanted by Hydrafig here](https://huggingface.co/Hydrafig/Atlantis-v0.1-12B-6BPW-EXL2)
[Find the 8bpw version quanted by myself here](https://huggingface.co/Statuo/Atlantis-v0.1-EXL2-8bpw)
# Model Details
Atlantis 12B is a finetune of [Mistral-Nemo-2407-Instruct](https://huggingface.co/mistralai/Mistral-Nemo-Instruct-2407) in an attempt to make it a more RP/storytelling
friendly model.
The model should support up to 128K context.
This is a very early-stage and brand new finetune, so I expect it to have issues. Please read through the 'Usage' section before reporting any issues as Nemo has some
specific requirements regarding backends, prompting formats & sampler settings.
# Feedback
I appreciate all feedback on any of my model merges, you can use:
* [The Community tab](https://huggingface.co/invisietch/Atlantis-v0.1-12B/discussions) - requires HF login.
* [SillyTavern Discord thread](https://discord.com/channels/1100685673633153084/1266471239397019780) - Must be on SillyTavern Discord.
Your feedback is how I improve these models for future versions.
# Formats
* [FP16 Safetensors](https://huggingface.co/invisietch/Atlantis-v0.1-12B)
* [Static GGUF](https://huggingface.co/invisietch/Atlantis-v0.1-12B-GGUF)
* [6BPW EXL2](https://huggingface.co/Hydrafig/Atlantis-v0.1-12B-6BPW-EXL2) (thanks to [Hydrafig](https://huggingface.co/Hydrafig))
# Disclaimer
This model is built on top of Mistral's Nemo-2407-Instruct model and released under the Apache 2.0 license.
This model is fairly unaligned and can generate explicit, disturbing or offensive responses.
Use responsibly. I am not responsible for your use of this model.
# Usage
## Inferencing
If you are running a GGUF quantization, you need to make sure that you are running:
- llama.cpp b3436 or newer
- koboldcpp 1.71 or newer
- oobabooga 1.11 or newer
If the GGUF refuses to load, it is probably because your software is out of date.
## Prompt Format
I would suggest using the Mistral prompt format:
```
[INST]Please write me a song.[/INST]
```
Other prompt formats may or may not work, but I've had most success with this.
## Recommended Sampler Settings
**Note the very low temperature, this is the same as Mistral's recommendation for the base model.**
I'd recommend starting with:
- **Temp**: 0.3-0.5
- **Top-P**: 1
- **Min-P**: 0
- **Rep Pen**: 1.13-1.16
- **Response Length**: 350+
The model does prefer a longer response length.
# Example Storywriting
These prompts are used on SillyTavern with a fairly basic narrator card. I have trimmed the start and finish where the narrator decided to add
chapter headings, commentary and the like. All samples are made with the Q8_0 GGUF loaded with koboldcpp, with response length capped at 4096 tokens.
Continue was pressed a maximum of once if the model cut off halfway through a chapter.