# Koala: A Dialogue Model for Academic Research
This repo contains the weights of the Koala 7B model produced at Berkeley. It is the result of combining the diffs from https://huggingface.co/young-geng/koala with the original Llama 7B model.
This version has then been quantized to 4-bit using [GPTQ-for-LLaMa](https://github.com/qwopqwop200/GPTQ-for-LLaMa).
## My Koala repos
I have the following Koala model repositories available:
**13B models:**
* [Unquantized 13B model in HF format](https://huggingface.co/TheBloke/koala-13B-HF)
* [GPTQ quantized 4bit 13B model in `pt` and `safetensors` formats](https://huggingface.co/TheBloke/koala-13B-GPTQ-4bit-128g)
* [4-bit, 5-bit and 8-bit GGML models for `llama.cpp`](https://huggingface.co/TheBloke/koala-13B-GGML)
**7B models:**
* [Unquantized 7B model in HF format](https://huggingface.co/TheBloke/koala-7B-HF)
* [Unquantized 7B model in GGML format for llama.cpp](https://huggingface.co/TheBloke/koala-7b-ggml-unquantized)
* [GPTQ quantized 4bit 7B model in `pt` and `safetensors` formats](https://huggingface.co/TheBloke/koala-7B-GPTQ-4bit-128g)
* [4-bit, 5-bit and 8-bit GGML models for `llama.cpp`](https://huggingface.co/TheBloke/koala-7B-GGML)
## GETTING GIBBERISH OUTPUT?
Please read the sections below carefully. Gibberish output is expected if you are using the `safetensors` file without the latest GPTQ-for-LLaMa code.
Your options are either to update GPTQ-for-LLaMa under `text-generation-webui/repositories` to a more recent version, or use the other file provided, `koala-7B-4bit-128g.no-act-order.ooba.pt` which will work immediately.
Unfortunately right now it is a bit more complext to update GPTQ-for-LLaMa because the most recent code has breaking changes which are not supported by `text-generation-webui`.
Therefore it's currently recommended to use `koala-7B-4bit-128g.no-act-order.ooba.pt`.
## Provided files
Three model files are provided. You don't need all three - choose the one that suits your needs best!
Details of the files provided:
* `koala-7B-4bit-128g.safetensors`
* newer `safetensors` format, with improved file security, created with the latest [GPTQ-for-LLaMa](https://github.com/qwopqwop200/GPTQ-for-LLaMa) code.
* Command to create:
* `python3 llama.py koala-7B-HF c4 --wbits 4 --true-sequential --act-order --groupsize 128 --save_safetensors koala-7B-4bit-128g.safetensors`
* `koala-7B-4bit-128g.no-act-order.ooba.pt`
* `pt` format file, created with [oobabooga's older CUDA fork of GPTQ-for-LLaMa](https://github.com/oobabooga/GPTQ-for-LLaMa).
* This file is included primarily for Windows users, as it can be used without needing to compile the latest GPTQ-for-LLaMa code.
* It should hopefully therefore work with one-click-installers on Windows, which include the older GPTQ-for-LLaMa code.
* The older GPTQ code does not support all the latest features, so the quality may be fractionally lower.
* Command to create:
* `python3 llama.py koala-7B-HF c4 --wbits 4 --true-sequential --groupsize 128 --save koala-7B-4bit-128g.no-act-order.ooba.pt`
## How to run in `text-generation-webui`
File `koala-7B-4bit-128g.no-act-order.ooba.pt` can be loaded the same as any other GPTQ file, without requiring any updates to [oobaboogas text-generation-webui](https://github.com/oobabooga/text-generation-webui).
The other two model files were created with the latest GPTQ code, and require that the latest GPTQ-for-LLaMa is used inside the UI.
Here are the commands I used to clone the Triton branch of GPTQ-for-LLaMa, clone text-generation-webui, and install GPTQ into the UI:
```
git clone https://github.com/qwopqwop200/GPTQ-for-LLaMa
git clone https://github.com/oobabooga/text-generation-webui
mkdir -p text-generation-webui/repositories
ln -s GPTQ-for-LLaMa text-generation-webui/repositories/GPTQ-for-LLaMa
```
Then install this model into `text-generation-webui/models` and launch the UI as follows:
```
cd text-generation-webui
python server.py --model koala-7B-GPTQ-4bit-128g --wbits 4 --groupsize 128 --model_type Llama # add any other command line args you want
```
The above commands assume you have installed all dependencies for GPTQ-for-LLaMa and text-generation-webui. Please see their respective repositories for further information.
If you are on Windows, or cannot use the Triton branch of GPTQ for any other reason, you can instead use the CUDA branch:
```
git clone https://github.com/qwopqwop200/GPTQ-for-LLaMa -b cuda
cd GPTQ-for-LLaMa
python setup_cuda.py install
```
Then link that into `text-generation-webui/repositories` as described above.
Or just use `koala-7B-4bit-128g.no-act-order.ooba.pt` as mentioned above.
## How the Koala delta weights were merged
The Koala delta weights were originally merged using the following commands, producing [koala-7B-HF](https://huggingface.co/TheBloke/koala-7B-HF):
```
git clone https://github.com/young-geng/EasyLM
git clone https://huggingface.co/nyanko7/LLaMA-7B
mkdir koala_diffs && cd koala_diffs && wget https://huggingface.co/young-geng/koala/resolve/main/koala_7b_diff_v2
cd EasyLM
PYTHON_PATH="${PWD}:$PYTHONPATH" python \
-m EasyLM.models.llama.convert_torch_to_easylm \
--checkpoint_dir=/content/LLaMA-7B \
--output_file=/content/llama-7B-LM \
--streaming=True
PYTHON_PATH="${PWD}:$PYTHONPATH" python \
-m EasyLM.scripts.diff_checkpoint --recover_diff=True \
--load_base_checkpoint='params::/content/llama-7B-LM' \
--load_target_checkpoint='params::/content/koala_diffs/koala_7b_diff_v2' \
--output_file=/content/koala_7b.diff.weights \
--streaming=True
PYTHON_PATH="${PWD}:$PYTHONPATH" python \
-m EasyLM.models.llama.convert_easylm_to_hf --model_size=7b \
--output_dir=/content/koala-7B-HF \
--load_checkpoint='params::/content/koala_7b.diff.weights' \
--tokenizer_path=/content/LLaMA-7B/tokenizer.model
```
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/Jq4vkcDakD)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Patreon special mentions**: Aemon Algiz, Dmitriy Samsonov, Nathan LeClaire, Trenton Dambrowitz, Mano Prime, David Flickinger, vamX, Nikolai Manek, senxiiz, Khalefa Al-Ahmad, Illia Dulskyi, Jonathan Leane, Talal Aujan, V. Lukas, Joseph William Delisle, Pyrater, Oscar Rangel, Lone Striker, Luke Pendergrass, Eugene Pentland, Sebastain Graf, Johann-Peter Hartman.
Thank you to all my generous patrons and donaters!
## Further info
Check out the following links to learn more about the Berkeley Koala model.
* [Blog post](https://bair.berkeley.edu/blog/2023/04/03/koala/)
* [Online demo](https://koala.lmsys.org/)
* [EasyLM: training and serving framework on GitHub](https://github.com/young-geng/EasyLM)
* [Documentation for running Koala locally](https://github.com/young-geng/EasyLM/blob/main/docs/koala.md)
## License
The model weights are intended for academic research only, subject to the
[model License of LLaMA](https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md),
[Terms of Use of the data generated by OpenAI](https://openai.com/policies/terms-of-use),
and [Privacy Practices of ShareGPT](https://chrome.google.com/webstore/detail/sharegpt-share-your-chatg/daiacboceoaocpibfodeljbdfacokfjb).
Any other usage of the model weights, including but not limited to commercial usage, is strictly prohibited.
