AWQ is an efficient, accurate and blazing-fast low-bit weight quantization method, currently supporting 4-bit quantization. Compared to GPTQ, it offers faster Transformers-based inference.

It is also now supported by continuous batching server vLLM, allowing use of AWQ models for high-throughput concurrent inference in multi-user server scenarios. Note that, at the time of writing, overall throughput is still lower than running vLLM with unquantised models, however using AWQ enables using much smaller GPUs which can lead to easier deployment and overall cost savings. For example, a 70B model can be run on 1 x 48GB GPU instead of 2 x 80GB.

Repositories available

Prompt template: Inkbot

<#meta#>
- Date: [DATE]
- Task: [TASK TYPE]
<#system#>
{system_message}
<#chat#>
<#user#>
{prompt}
<#bot#>

Provided files and AWQ parameters

For my first release of AWQ models, I am releasing 128g models only. I will consider adding 32g as well if there is interest, and once I have done perplexity and evaluation comparisons, but at this time 32g models are still not fully tested with AutoAWQ and vLLM.

Models are released as sharded safetensors files.

Branch	Bits	GS	AWQ Dataset	Seq Len	Size
main	4	128	wikitext	4096	7.25 GB

Serving this model from vLLM

Documentation on installing and using vLLM can be found here.

When using vLLM as a server, pass the --quantization awq parameter, for example:

python3 python -m vllm.entrypoints.api_server --model TheBloke/Inkbot-13B-4k-AWQ --quantization awq --dtype half

When using vLLM from Python code, pass the quantization=awq parameter, for example:

python

from vllm import LLM, SamplingParams

prompts = [
    "Hello, my name is",
    "The president of the United States is",
    "The capital of France is",
    "The future of AI is",
]
sampling_params = SamplingParams(temperature=0.8, top_p=0.95)

llm = LLM(model="TheBloke/Inkbot-13B-4k-AWQ", quantization="awq", dtype="half")

outputs = llm.generate(prompts, sampling_params)

# Print the outputs.
for output in outputs:
    prompt = output.prompt
    generated_text = output.outputs[0].text
    print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")

How to use this AWQ model from Python code

Install the necessary packages

Requires: AutoAWQ 0.0.2 or later

pip3 install autoawq

If you have problems installing AutoAWQ using the pre-built wheels, install it from source instead:

shell

pip3 uninstall -y autoawq
git clone https://github.com/casper-hansen/AutoAWQ
cd AutoAWQ
pip3 install .

You can then try the following example code

python

from awq import AutoAWQForCausalLM
from transformers import AutoTokenizer

model_name_or_path = "TheBloke/Inkbot-13B-4k-AWQ"

# Load model
model = AutoAWQForCausalLM.from_quantized(model_name_or_path, fuse_layers=True,
                                          trust_remote_code=False, safetensors=True)
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, trust_remote_code=False)

prompt = "Tell me about AI"
prompt_template=f'''<#meta#>
- Date: [DATE]
- Task: [TASK TYPE]
<#system#>
{system_message}
<#chat#>
<#user#>
{prompt}
<#bot#>

'''

print("\n\n*** Generate:")

tokens = tokenizer(
    prompt_template,
    return_tensors='pt'
).input_ids.cuda()

# Generate output
generation_output = model.generate(
    tokens,
    do_sample=True,
    temperature=0.7,
    top_p=0.95,
    top_k=40,
    max_new_tokens=512
)

print("Output: ", tokenizer.decode(generation_output[0]))

# Inference can also be done using transformers' pipeline
from transformers import pipeline

print("*** Pipeline:")
pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    max_new_tokens=512,
    do_sample=True,
    temperature=0.7,
    top_p=0.95,
    top_k=40,
    repetition_penalty=1.1
)

print(pipe(prompt_template)[0]['generated_text'])

Compatibility

The files provided are tested to work with AutoAWQ, and vLLM.

Huggingface Text Generation Inference (TGI) is not yet compatible with AWQ, but a PR is open which should bring support soon: TGI PR #781.

Discord

For further support, and discussions on these models and AI in general, join us at:

TheBloke AI's Discord server

Thanks, and how to contribute

Thanks to the chirper.ai team!

Thanks to Clay from gpus.llm-utils.org!

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

Special thanks to: Aemon Algiz.

Patreon special mentions: Alicia Loh, Stephen Murray, K, Ajan Kanaga, RoA, Magnesian, Deo Leter, Olakabola, Eugene Pentland, zynix, Deep Realms, Raymond Fosdick, Elijah Stavena, Iucharbius, Erik Bjäreholt, Luis Javier Navarrete Lozano, Nicholas, theTransient, John Detwiler, alfie_i, knownsqashed, Mano Prime, Willem Michiel, Enrico Ros, LangChain4j, OG, Michael Dempsey, Pierre Kircher, Pedro Madruga, James Bentley, Thomas Belote, Luke @flexchar, Leonard Tan, Johann-Peter Hartmann, Illia Dulskyi, Fen Risland, Chadd, S_X, Jeff Scroggin, Ken Nordquist, Sean Connelly, Artur Olbinski, Swaroop Kallakuri, Jack West, Ai Maven, David Ziegler, Russ Johnson, transmissions 11, John Villwock, Alps Aficionado, Clay Pascal, Viktor Bowallius, Subspace Studios, Rainer Wilmers, Trenton Dambrowitz, vamX, Michael Levine, 준교 김, Brandon Frisco, Kalila, Trailburnt, Randy H, Talal Aujan, Nathan Dryer, Vadim, 阿明, ReadyPlayerEmma, Tiffany J. Kim, George Stoitzev, Spencer Kim, Jerry Meng, Gabriel Tamborski, Cory Kujawski, Jeffrey Morgan, Spiking Neurons AB, Edmond Seymore, Alexandros Triantafyllidis, Lone Striker, Cap'n Zoog, Nikolai Manek, danny, ya boyyy, Derek Yates, usrbinkat, Mandus, TL, Nathan LeClaire, subjectnull, Imad Khwaja, webtim, Raven Klaugh, Asp the Wyvern, Gabriel Puliatti, Caitlyn Gatomon, Joseph William Delisle, Jonathan Leane, Luke Pendergrass, SuperWojo, Sebastain Graf, Will Dee, Fred von Graf, Andrey, Dan Guido, Daniel P. Andersen, Nitin Borwankar, Elle, Vitor Caleffi, biorpg, jjj, NimbleBox.ai, Pieter, Matthew Berman, terasurfer, Michael Davis, Alex, Stanislav Ovsiannikov

Thank you to all my generous patrons and donaters!

And thank you again to a16z for their generous grant.

Original model card: Tostino's Inkbot 13B 4K

Model Card for Inkbot

Model Details

Inkbot is a conversational AI model designed to interpret and respond to structured prompts with or without contextual information. Built on the latest advancements in natural language processing (NLP) and understanding (NLU), Inkbot provides users with accurate and meaningful interactions, addressing a wide range of queries and topics. Its unique feature lies in the structured prompt system, allowing users to engage in dynamic dialogues that can evolve based on context, metadata, and user input.

How to Use

Inkbot uses a structured prompt template system. Depending on the information you have or want to give, you can use templates 'with context' or 'without context'.

Prompt Template Structure

With Context:

<#meta#>
- Date: [DATE]
- Task: [TASK TYPE]
<#system#>
[SYSTEM PROMPT/DESCRIPTION OF THE INTERACTION]
<#chat#>
<#user#>
[USER QUESTION]
<#bot#>
[BOT RESPONSE]
<#user#>
[USER FOLLOW-UP QUESTION]
<#user_context#>
[ADDITIONAL CONTEXT FOR THE BOT]
<#bot#>
[BOT RESPONSE BASED ON CONTEXT]

Without Context:

<#meta#>
- Date: [DATE]
- Task: [TASK TYPE]
<#system#>
[SYSTEM PROMPT/DESCRIPTION OF THE INTERACTION]
<#chat#>
<#user#>
[USER QUESTION]
<#bot#>
[BOT RESPONSE]
<#user#>
[USER FOLLOW-UP QUESTION]
<#bot#>
[BOT RESPONSE]

Examples

With Context:

<#meta#>
- Date: 2023-07-20
- Task: general
<#system#>
A dialogue focused on determining dates based on context clues.
<#chat#>
<#user#>
What is today's date?
<#bot#>
The date is: 2023-07-20
<#user#>
When did the event in the provided context take place?
<#user_context#>
It's the day after new years eve, and I'm starting 2018 with a headache.
<#bot#>
The event in the context took place on: 2018-01-01

Without Context:

<#meta#>
- Date: 2023-07-20
- Task: general
<#system#>
A simple dialogue about daily queries.
<#chat#>
<#user#>
Will it rain today?
<#bot#>
I am not sure, without knowing your location I am unable to determine the weather.
<#user#>
Can you compose a haiku for me?
<#bot#>
Whispers of the breeze,
Nature's melodies unfold,
Dawn dispels the night.

Task Options

Inkbot has been trained for a variety of tasks. Below are some of the key task options you can utilize:

general: This is the default task and is designed for a broad range of general questions and interactions.
- Usage: Suitable for most day-to-day interactions and queries.
knowledge_graph: This task involves extracting, understanding, and representing information in a structured way.
- Usage: When you want to extract relationships between entities or desire structured representations of data.
question_answer: Explicitly trained for answering questions in a straightforward manner.
- Usage: Best used when you have direct questions and expect concise answers.
reasoning: Allows Inkbot to showcase its logical and deductive reasoning capabilities.
- Usage: Ideal for puzzles, riddles, or scenarios where logical analysis is required.
translation: Use this for language translation tasks.
- Usage: Provide a sentence or paragraph in one language, and specify the desired target language for translation.
summarization: Trained for condensing large texts into shorter, coherent summaries.
- Usage: When you have a lengthy text or article that you want to be summarized to its key points.
creative_writing: Engage Inkbot in composing stories, poetry, and other creative content.
- Usage: For tasks that require imaginative and original content generation.

How to Use Task Options

In the prompt template structure, the Task metadata field is where you specify the task option. Here's an example of how to structure a prompt using the reasoning task:

Limitations

Ensure you adhere to the prompt structure for best results.
When providing contextual details, clarity is essential for Inkbot to derive accurate and meaningful responses.

Additional Notes

The 'date', 'task', and 'system' are crucial metadata components that need to be provided outside the core dialogue.
Use the 'user_context' key when you want to offer supplementary context that guides Inkbot's response.