README.md

---
datasets:
- anon8231489123/ShareGPT_Vicuna_unfiltered
- ehartford/wizard_vicuna_70k_unfiltered
- ehartford/WizardLM_alpaca_evol_instruct_70k_unfiltered
- QingyiSi/Alpaca-CoT
- teknium/GPT4-LLM-Cleaned
- teknium/GPTeacher-General-Instruct
- metaeval/ScienceQA_text_only
- hellaswag
- tasksource/mmlu
- openai/summarize_from_feedback
language:
- en
library_name: transformers
pipeline_tag: text-generation
inference: false
---

# Manticore 13B GGML

This is GGML format quantised 4bit and 5bit models of [OpenAccess AI Collective's Manticore 13B](https://huggingface.co/openaccess-ai-collective/manticore-13b).

This repo is the result of quantising to 4-bit, 5-bit and 8-bit GGML for CPU (+CUDA) inference using [llama.cpp](https://github.com/ggerganov/llama.cpp).

## Repositories available

* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Manticore-13B-GPTQ).
* [4-bit, 5-bit 8-bit GGML models for llama.cpp CPU (+CUDA) inference](https://huggingface.co/TheBloke/Manticore-13B-GGML).
* [OpenAccess AI Collective's original float16 HF format repo for GPU inference and further conversions](https://huggingface.co/openaccess-ai-collective/manticore-13b).

## THE FILES IN MAIN BRANCH REQUIRES LATEST LLAMA.CPP (May 12th 2023 - commit b9fd7ee)!

llama.cpp recently made a breaking change to its quantisation methods.

I have quantised the GGML files in this repo with the latest version. Therefore you will require llama.cpp compiled on May 12th or later (commit `b9fd7ee` or later) to use them.

## Provided files
| Name | Quant method | Bits | Size | RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
`manticore-13B.ggmlv2.q4_0.bin` | q4_0 | 4bit | 8.14GB | 10.5GB | 4-bit. |
`manticore-13B.ggmlv2.q4_1.bin` | q5_0 | 5bit | 8.95GB | 11.0GB | 5-bit. Higher accuracy, higher resource usage and slower inference. |
`manticore-13B.ggmlv2.q5_0.bin` | q5_1 | 5bit | 9.76GB | 12.25GB | 5-bit. Even higher accuracy, and higher resource usage and slower inference. |
`manticore-13B.ggmlv2.q8_0.bin` | q8_0 | 8bit | 14.6GB | 17GB | 8-bit. Almost indistinguishable from float16. Huge resource use and slow. Not recommended for normal use. |

## How to run in `llama.cpp`

I use the following command line; adjust for your tastes and needs:

```
./main -t 8 -m manticore-13B-.ggmlv2.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### Instruction: write a story about llamas ### Response:"
```

Change `-t 8` to the number of physical CPU cores you have.

## How to run in `text-generation-webui`

GGML models can be loaded into text-generation-webui by installing the llama.cpp module, then placing the ggml model file in a model folder as usual.

Further instructions here: [text-generation-webui/docs/llama.cpp-models.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp-models.md).

# Original Model Card: Manticore 13B - Preview Release (previously Wizard Mega)

Manticore 13B is a Llama 13B model fine-tuned on the following datasets:
- [ShareGPT](https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered) - based on a cleaned and de-suped subset
- [WizardLM](https://huggingface.co/datasets/ehartford/WizardLM_alpaca_evol_instruct_70k_unfiltered)
- [Wizard-Vicuna](https://huggingface.co/datasets/ehartford/wizard_vicuna_70k_unfiltered)
- [subset of QingyiSi/Alpaca-CoT for roleplay and CoT](https://huggingface.co/QingyiSi/Alpaca-CoT)
- [GPT4-LLM-Cleaned](https://huggingface.co/datasets/teknium/GPT4-LLM-Cleaned)
- [GPTeacher-General-Instruct](https://huggingface.co/datasets/teknium/GPTeacher-General-Instruct)
- ARC-Easy & ARC-Challenge - instruct augmented  for detailed responses
- mmlu: instruct augmented for detailed responses subset including
  - abstract_algebra
  - conceptual_physics
  - formal_logic
  - high_school_physics
  - logical_fallacies
- [hellaswag](https://huggingface.co/datasets/hellaswag) - 5K row subset of instruct augmented for concise responses
- [metaeval/ScienceQA_text_only](https://huggingface.co/datasets/metaeval/ScienceQA_text_only) - instruct for concise responses
- [openai/summarize_from_feedback](https://huggingface.co/datasets/openai/summarize_from_feedback) - instruct augmented tl;dr summarization


# Demo

Try out the model in HF Spaces. The demo uses a quantized GGML version of the model to quickly return predictions on smaller GPUs (and even CPUs). Quantized GGML may have some minimal loss of model quality.
- https://huggingface.co/spaces/openaccess-ai-collective/manticore-ggml

## Release Notes

- https://wandb.ai/wing-lian/manticore-13b/runs/nq3u3uoh/workspace

## Build

Manticore was built with [Axolotl](https://github.com/OpenAccess-AI-Collective/axolotl) on 8xA100 80GB 
 - Preview Release: 1 epoch taking 8 hours. 
 - The configuration to duplicate this build is provided in this repo's [/config folder](https://huggingface.co/openaccess-ai-collective/manticore-13b/tree/main/configs).

## Bias, Risks, and Limitations
Manticore has not been aligned to human preferences with techniques like RLHF or deployed with in-the-loop filtering of responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so).
Manticore was fine-tuned from the base model LlaMa 13B, please refer to its model card's Limitations Section for relevant information.

## Examples

````
### Instruction: write Python code that returns the first n numbers of the Fibonacci sequence using memoization.

### Assistant: 
````

```
### Instruction: Finish the joke, a mechanic and a car salesman walk into a bar...  

### Assistant:
```