---
license: agpl-3.0
language:
- en
model_creator: PygmalionAI
quantized_by: Crataco
tags:
- ggml
- text generation
- conversational
inference: false
---

*(Not to be confused with [Pygmalion 13B](https://huggingface.co/TehVenom/Pygmalion-13b-GGML) and [Pygmalion 2 13B](https://huggingface.co/TheBloke/Pygmalion-2-13B-GGUF).)*

# Pygmalion 1.3B GGML
### This repository contains quantized conversions of the Pygmalion 1.3B checkpoint.
*For use with frontends that support GGML quantized GPT-NeoX models, such as KoboldCpp and Oobabooga (with the CTransformers loader).*

*Last updated on 2023-09-23.*

Model | Startup RAM usage (KoboldCpp) | Startup RAM usage (Oobabooga)
:--:|:--:|:--:
pygmalion-1.3b.q4_0.bin | 1.0 GiB | 1.3 GiB
pygmalion-1.3b.q4_1.bin | 1.1 GiB | 1.4 GiB
pygmalion-1.3b.q5_0.bin | 1.2 GiB | 1.5 GiB
pygmalion-1.3b.q5_1.bin | 1.3 GiB | 1.6 GiB
pygmalion-1.3b.q8_0.bin | 1.7 GiB | 2.0 GiB
pygmalion-1.3b.f16.bin | 2.9 GiB | 3.2 GiB

**Recommended settings:**

Pygmalion 1.3B is a limited model, left in the dust by the Pygmalion project's advancements since then. Which is a shame, as it remains one of the few conversational models available for systems with less than 2GB RAM, at least before we get [TinyLLaMA](https://github.com/jzhang38/TinyLlama) and quantized [Phi-1.5](https://huggingface.co/microsoft/phi-1_5).

Here are some tips to get the best results you can out of this model:
- Stick to a low temperature, preferably between 0.2 and 0.7.
- Keep your repetition penalty between 1.0 and 1.02. These tiny values are required for models based on Pythia Deduped.
- If using SillyTavern, follow these settings:
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6251b9851842c08ef3111c4f/Yqvgv428hA9V67jC9VZTp.png)
- You also have to keep character descriptions to a few sentences, possibly following CharacterAI's 500-character descriptions.

**Notes:**
- KoboldCpp [[bfc696f]](https://github.com/LostRuins/koboldcpp/tree/bfc696fcc452975dbe8967c39301ba856d04a030) was tested without OpenBLAS.
- Oobabooga [[895ec9d]](https://github.com/oobabooga/text-generation-webui/tree/895ec9dadb96120e8202a83052bf9032ca3245ae) was tested with with the `--model <model> --loader ctransformers --model_type gptneox` launch arguments.
- ggerganov/ggml [[8ca2c19]](https://github.com/ggerganov/ggml/tree/8ca2c19a3bb8622954d858fbf6383522684eaf34) was used for conversion and quantization.
- The original model is available at [PygmalionAI/pygmalion-1.3b](https://huggingface.co/PygmalionAI/pygmalion-1.3b).
- Earlier ggmlv2 quantizations are available [here](https://huggingface.co/Crataco/Pygmalion-1.3B-GGML/tree/15d3aa5e07372e4200c598443d211a2976db47f9).

Below is the original model card for Pygmalion 1.3B.

* * *

# Pygmalion 1.3B

## Model description

Pymalion 1.3B is a proof-of-concept dialogue model based on EleutherAI's [pythia-1.3b-deduped](https://huggingface.co/EleutherAI/pythia-1.3b-deduped).

**Warning:** This model is **NOT** suitable for use by minors. It **will** output X-rated content under certain circumstances.

## Training data

The fine-tuning dataset consisted of 56MB of dialogue data gathered from multiple sources, which includes both real _and_ partially machine-generated conversations.

## Training procedure

Fine-tuning was done using [ColossalAI](https://github.com/hpcaitech/ColossalAI) (specifically, with a slightly modified version of their [OPT fine-tune example](https://github.com/hpcaitech/ColossalAI/blob/78509124d32b63b7fc36f6508e0576a326d51422/examples/language/opt/run_clm.py)) for around 11.4 million tokens over 5440 steps on a single 24GB GPU. The run took just under 21 hours.

## Intended use

### The easy way

We provide a notebook with a Gradio UI for playing around with the model without having to manually format inputs. This notebook can be found [here](https://github.com/PygmalionAI/gradio-ui/blob/master/notebooks/GPU.ipynb).

### The manual way

The model can be used as a regular text generation model, but it'll perform best if the input prompt adheres to the following format:

```
[CHARACTER]'s Persona: [A few sentences about the character you want the model to play]

[DIALOGUE HISTORY]
You: [Your input message here]
[CHARACTER]:
```

Where `[CHARACTER] `is, as you can probably guess, the name of the character you want the model to portray, and `[DIALOGUE HISTORY]` is chat history so the model can have some conversational context to draw from. Ideally it'll be pairs of messages like:

```
[CHARACTER]: [some dialogue here]
You: [your response to the dialogue above]
```

Apart from chat history, you can also just add example conversations in `[DIALOGUE HISTORY]` to show how the character should speak - ideally at the beginning, so it doesn't get confused as to what's conversation history vs. character definition.

## Known issues

- The model can get stuck repeating certain phrases, or sometimes even entire sentences.
  - We believe this is due to that behavior being present in the training data itself, and plan to investigate and adjust accordingly for future versions.