These files were quantised using hardware kindly provided by Massed Compute.
About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
llama.cpp. The source project for GGUF. Offers a CLI and a server option.
text-generation-webui, the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
KoboldCpp, a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
GPT4All, a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel.
LM Studio, an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023.
LoLLMS Web UI, a great web UI with many interesting and unique features, including a full model library for easy model selection.
Faraday.dev, an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
llama-cpp-python, a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
candle, a Rust ML framework with a focus on performance, including GPU support, and ease of use.
ctransformers, a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models.
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit d0cee0d
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
Explanation of quantisation methods
Click to see details
The new methods available are:
GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
very large, extremely low quality loss - not recommended
Note: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
How to download GGUF files
Note for manual downloaders: You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
LM Studio
LoLLMS Web UI
Faraday.dev
In text-generation-webui
Under Download Model, you can enter the model repo: TheBloke/openchat-3.5-1210-GGUF and below it, a specific filename to download, such as: openchat-3.5-1210.Q4_K_M.gguf.
Then click Download.
On the command line, including multiple files at once
I recommend using the huggingface-hub Python library:
pip3 install huggingface-hub
Then you can download any individual model file to the current directory, at high speed, with a command like this:
Change -ngl 32 to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change -c 8192 to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically. Note that longer sequence lengths require much more resources, so you may need to reduce this value.
If you want to have a chat-style conversation, replace the -p <PROMPT> argument with -i -ins
You can use GGUF models from Python using the llama-cpp-python or ctransformers libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.
How to load this model in Python code, using llama-cpp-python
Run one of the following commands, according to your system:
shell
1# Base ctransformers with no GPU acceleration2pip install llama-cpp-python
3# With NVidia CUDA acceleration4CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
5# Or with OpenBLAS acceleration6CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
7# Or with CLBLast acceleration8CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
9# Or with AMD ROCm GPU acceleration (Linux only)10CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
11# Or with Metal GPU acceleration for macOS systems only12CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python
1314# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:15$env:CMAKE_ARGS ="-DLLAMA_OPENBLAS=on"16pip install llama-cpp-python
Simple llama-cpp-python example code
python
1from llama_cpp import Llama
23# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.4llm = Llama(5 model_path="./openchat-3.5-1210.Q4_K_M.gguf",# Download the model file first6 n_ctx=8192,# The max sequence length to use - note that longer sequence lengths require much more resources7 n_threads=8,# The number of CPU threads to use, tailor to your system and the resulting performance8 n_gpu_layers=35# The number of layers to offload to GPU, if you have GPU acceleration available9)1011# Simple inference example12output = llm(13"GPT4 Correct User: {prompt}<|end_of_turn|>GPT4 Correct Assistant:",# Prompt14 max_tokens=512,# Generate up to 512 tokens15 stop=["</s>"],# Example stop token - not necessarily correct for this specific model! Please check before using.16 echo=True# Whether to echo the prompt17)1819# Chat Completion API2021llm = Llama(model_path="./openchat-3.5-1210.Q4_K_M.gguf", chat_format="llama-2")# Set chat_format according to the model you are using22llm.create_chat_completion(23 messages =[24{"role":"system","content":"You are a story writing assistant."},25{26"role":"user",27"content":"Write a story about llamas."28}29]30)
How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
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 special mentions: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
Original model card: OpenChat's Openchat 3.5 1210
Advancing Open-source Language Models with Mixed-Quality Data
To use this model, we highly recommend installing the OpenChat package by following the installation guide in our repository and using the OpenChat OpenAI-compatible API server by running the serving command from the table below. The server is optimized for high-throughput deployment using vLLM and can run on a consumer GPU with 24GB RAM. To enable tensor parallelism, append --tensor-parallel-size N to the serving command.
Once started, the server listens at localhost:18888 for requests and is compatible with the OpenAI ChatCompletion API specifications. Please refer to the example request below for reference. Additionally, you can use the OpenChat Web UI for a user-friendly experience.
If you want to deploy the server as an online service, you can use --api-keys sk-KEY1 sk-KEY2 ... to specify allowed API keys and --disable-log-requests --disable-log-stats --log-file openchat.log for logging only to a file. For security purposes, we recommend using an HTTPS gateway in front of the server.
💡 Default Mode (GPT4 Correct): Best for coding, chat and general tasks
bash
1curl http://localhost:18888/v1/chat/completions \2 -H "Content-Type: application/json"\3 -d '{
4 "model": "openchat_3.5",
5 "messages": [{"role": "user", "content": "You are a large language model named OpenChat. Write a poem to describe yourself"}]
6 }'
🧮 Mathematical Reasoning Mode: Tailored for solving math problems
💡 Default Mode (GPT4 Correct): Best for coding, chat and general tasks
GPT4 Correct User: Hello<|end_of_turn|>GPT4 Correct Assistant: Hi<|end_of_turn|>GPT4 Correct User: How are you today?<|end_of_turn|>GPT4 Correct Assistant:
🧮 Mathematical Reasoning Mode: Tailored for solving math problems
Math Correct User: 10.3 − 7988.8133=<|end_of_turn|>Math Correct Assistant:
⚠️ Notice: Remember to set <|end_of_turn|> as end of generation token.
The default (GPT4 Correct) template is also available as the integrated tokenizer.chat_template,
which can be used instead of manually specifying the template:
python
1messages =[2{"role":"user","content":"Hello"},3{"role":"assistant","content":"Hi"},4{"role":"user","content":"How are you today?"}5]6tokens = tokenizer.apply_chat_template(messages, add_generation_prompt=True)7assert tokens ==[1,420,6316,28781,3198,3123,1247,28747,22557,32000,420,6316,28781,3198,3123,21631,28747,15359,32000,420,6316,28781,3198,3123,1247,28747,1602,460,368,3154,28804,32000,420,6316,28781,3198,3123,21631,28747]
(Experimental) Evaluator / Feedback Capabilities
We've included evaluator capabilities in this release to advance open-source models as evaluators. You can use `Default Mode (GPT4 Correct)` with the following prompt (same as [Prometheus](https://huggingface.co/datasets/kaist-ai/Feedback-Collection)) to evaluate a response.
###Task Description:
An instruction (might include an Input inside it), a response to evaluate, a reference answer that gets a score of 5, and a score rubric representing a evaluation criteria are given.
1. Write a detailed feedback that assess the quality of the response strictly based on the given score rubric, not evaluating in general.
2. After writing a feedback, write a score that is an integer between 1 and 5. You should refer to the score rubric.
3. The output format should look as follows: "Feedback: (write a feedback for criteria) [RESULT] (an integer number between 1 and 5)"
4. Please do not generate any other opening, closing, and explanations.
###The instruction to evaluate:
{orig_instruction}
###Response to evaluate:
{orig_response}
###Reference Answer (Score 5):
{orig_reference_answer}
###Score Rubrics:
[{orig_criteria}]
Score 1: {orig_score1_description}
Score 2: {orig_score2_description}
Score 3: {orig_score3_description}
Score 4: {orig_score4_description}
Score 5: {orig_score5_description}
###Feedback:
Benchmarks
Model
# Params
Average
MT-Bench
HumanEval
BBH MC
AGIEval
TruthfulQA
MMLU
GSM8K
BBH CoT
OpenChat-3.5-1210
7B
63.8
7.76
68.9
49.5
48.0
61.8
65.3
77.3
61.8
OpenChat-3.5
7B
61.6
7.81
55.5
47.6
47.4
59.1
64.3
77.3
63.5
ChatGPT (March)*
?
61.5
7.94
48.1
47.6
47.1
57.7
67.3
74.9
70.1
OpenHermes 2.5
7B
59.3
7.54
48.2
49.4
46.5
57.5
63.8
73.5
59.9
OpenOrca Mistral
7B
52.7
6.86
38.4
49.4
42.9
45.9
59.3
59.1
58.1
Zephyr-β^
7B
34.6
7.34
22.0
40.6
39.0
40.8
39.8
5.1
16.0
Mistral
7B
-
6.84
30.5
39.0
38.0
-
60.1
52.2
-
Evaluation Details(click to expand)
*: ChatGPT (March) results are from [GPT-4 Technical Report](https://arxiv.org/abs/2303.08774), [Chain-of-Thought Hub](https://github.com/FranxYao/chain-of-thought-hub), and our evaluation. Please note that ChatGPT is not a fixed baseline and evolves rapidly over time.
^: Zephyr-β often fails to follow few-shot CoT instructions, likely because it was aligned with only chat data but not trained on few-shot data.
**: Mistral and Open-source SOTA results are taken from reported results in instruction-tuned model papers and official repositories.
All models are evaluated in chat mode (e.g. with the respective conversation template applied). All zero-shot benchmarks follow the same setting as in the AGIEval paper and Orca paper. CoT tasks use the same configuration as Chain-of-Thought Hub, HumanEval is evaluated with EvalPlus, and MT-bench is run using FastChat. To reproduce our results, follow the instructions in our repository.
⚠️ Note that this model was not explicitly trained in Chinese (only < 0.1% of the data is in Chinese). 请注意本模型没有针对性训练中文（中文数据占比小于0.1%）。
Multi-Level Multi-Discipline Chinese Evaluation Suite (CEVAL)
Model
Avg
STEM
Social Science
Humanities
Others
ChatGPT
54.4
52.9
61.8
50.9
53.6
OpenChat
47.29
45.22
52.49
48.52
45.08
Massive Multitask Language Understanding in Chinese (CMMLU, 5-shot)
Models
STEM
Humanities
SocialSciences
Other
ChinaSpecific
Avg
ChatGPT
47.81
55.68
56.5
62.66
50.69
55.51
OpenChat
38.7
45.99
48.32
50.23
43.27
45.85
Limitations
Foundation Model Limitations
Despite its advanced capabilities, OpenChat is still bound by the limitations inherent in its foundation models. These limitations may impact the model's performance in areas such as:
Complex reasoning
Mathematical and arithmetic tasks
Programming and coding challenges
Hallucination of Non-existent Information
OpenChat may sometimes generate information that does not exist or is not accurate, also known as "hallucination". Users should be aware of this possibility and verify any critical information obtained from the model.
Safety
OpenChat may sometimes generate harmful, hate speech, biased responses, or answer unsafe questions. It's crucial to apply additional AI safety measures in use cases that require safe and moderated responses.
License
Our OpenChat 3.5 code and models are distributed under the Apache License 2.0.
Dataset Details
OpenChat 3.5 was trained with C-RLFT on a collection of publicly available high-quality instruction data, with a custom processing pipeline. We detail some notable subsets included here:
@article{wang2023openchat,
title={OpenChat: Advancing Open-source Language Models with Mixed-Quality Data},
author={Wang, Guan and Cheng, Sijie and Zhan, Xianyuan and Li, Xiangang and Song, Sen and Liu, Yang},
journal={arXiv preprint arXiv:2309.11235},
year={2023}
}
Acknowledgments
We extend our heartfelt gratitude to AutoMeta and caesus from Alignment Lab AI, LDJ and Teknium from Nous Research, alpin and TearGosling from Pygmalion AI for their substantial contributions to data collection and model training.
Special thanks go to Changling Liu from GPT Desk Pte. Ltd., Qiying Yu at Tsinghua University, Baochang Ma, and Hao Wan from 01.AI company for their generous provision of resources. We are also deeply grateful to Jianxiong Li and Peng Li at Tsinghua University for their insightful discussions.
