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: User-Assistant-Hashes

### User: {prompt}
### Assistant:

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	3.89 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/Tulpar-7B-v0-AWQ --quantization awq

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/Tulpar-7B-v0-AWQ", quantization="awq")

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/Tulpar-7B-v0-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'''### User: {prompt}
### Assistant:

'''

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: HyperbeeAI's Tulpar 7B v0

Model Description

Tulpar-7b is a LLama2-7b-based model trained by HyperbeeAI. Training is done on a filtered and preprocessed instruction finetuning dataset that includes GPT-4 generated and generally curated datasets like Airoboros and Platypus.

Example Usage

Loading the model:

python

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("HyperbeeAI/Tulpar-7b-v0")
model = AutoModelForCausalLM.from_pretrained("HyperbeeAI/Tulpar-7b-v0", device_map="auto")

You can run inference with both of the following prompts:

python

input_text="What is deep learning?"
prompt = f"### User: {input_text}\n\n### Assistant:\n"
inputs = tokenizer(prompt, return_tensors="pt")
output = model.generate(**inputs, do_sample=True, top_p=0.95, top_k=0, max_new_tokens=512)
print(tokenizer.decode(output[0]))

python

input_text="What is deep learning?"
prompt = f"Question: {input_text}\n\nAnswer:"
inputs = tokenizer(prompt, return_tensors="pt")
output = model.generate(**inputs, do_sample=True, top_p=0.95, top_k=0, max_new_tokens=512)
print(tokenizer.decode(output[0]))

Evaluation

Our offline HF Leaderboard evaluation results:


Task	Metric	Value
arc_challenge	acc_norm	0.5614
hellaswag	acc_norm	0.7901
mmlu	acc_norm	0.5242
truthfulqa_mc	mc2	0.5160
Average	-	0.5979

Other GPT4All evaluation results:


Task	Metric	Value
boolq	acc	0.8306
piqa	acc	0.7905
	acc_norm	0.7884
winogrande	acc	0.7159
openbookqa	acc	0.356
	acc_norm	0.448
Average (including HF leaderboard datasets)		0.6468

BigBenchHard results:


Task	Metric	Value
bigbench_causal_judgement	multiple_choice_grade	0.6105
bigbench_date_understanding	multiple_choice_grade	0.6423
bigbench_disambiguation_qa	multiple_choice_grade	0.3643
bigbench_dyck_languages	multiple_choice_grade	0.2000
bigbench_formal_fallacies_syllogisms_negation	multiple_choice_grade	0.5002
bigbench_geometric_shapes	multiple_choice_grade	0.0000
	exact_str_match	0.0000
bigbench_hyperbaton	multiple_choice_grade	0.6754
bigbench_logical_deduction_five_objects	multiple_choice_grade	0.2700
bigbench_logical_deduction_seven_objects	multiple_choice_grade	0.1929
bigbench_logical_deduction_three_objects	multiple_choice_grade	0.4133
bigbench_movie_recommendation	multiple_choice_grade	0.3000
bigbench_navigate	multiple_choice_grade	0.5000
bigbench_reasoning_about_colored_objects	multiple_choice_grade	0.5750
bigbench_ruin_names	multiple_choice_grade	0.3281
bigbench_salient_translation_error_detection	multiple_choice_grade	0.2976
bigbench_snarks	multiple_choice_grade	0.6022
bigbench_sports_understanding	multiple_choice_grade	0.5122
bigbench_temporal_sequences	multiple_choice_grade	0.1450
bigbench_tracking_shuffled_objects_five_objects	multiple_choice_grade	0.1976
bigbench_tracking_shuffled_objects_seven_objects	multiple_choice_grade	0.1440
bigbench_tracking_shuffled_objects_three_objects	multiple_choice_grade	0.4133
Average		0.3754

Ethical Considerations and Limitations

Tulpar is a technology with potential risks and limitations. This model is finetuned only in English and all language-related scenarios are not covered. As HyperbeeAI, we neither guarantee ethical, accurate, unbiased, objective responses nor endorse its outputs. Before deploying this model, you are advised to make safety tests for your use case.