Hugging Face's logo

TheBloke
/
Dawn-v2-70B-AWQ

Text Generation
Transformers
Safetensors
llama
Not-For-All-Audiences
nsfw
text-generation-inference
4-bit precision
awq
Model card Files Community
1
Edit model card
TheBlokeAI

TheBloke's LLM work is generously supported by a grant from andreessen horowitz (a16z)

Dawn V2 70B - AWQ

Description

This repo contains AWQ model files for Undi's Dawn V2 70B.

These files were quantised using hardware kindly provided by Massed Compute.

About AWQ

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 with equivalent or better quality compared to the most commonly used GPTQ settings.

It is supported by:

Repositories available

Prompt template: Alpaca 

Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{prompt}

### Response:

Licensing

The creator of the source model has listed its license as cc-by-nc-4.0, and this quantization has therefore used that same license.

As this model is based on Llama 2, it is also subject to the Meta Llama 2 license terms, and the license files for that are additionally included. It should therefore be considered as being claimed to be licensed under both licenses. I contacted Hugging Face for clarification on dual licensing but they do not yet have an official position. Should this change, or should Meta provide any feedback on this situation, I will update this section accordingly.

In the meantime, any questions regarding licensing, and in particular how these two licenses might interact, should be directed to the original model repository: Undi's Dawn V2 70B.

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 36.61 GB

How to easily download and use this model in text-generation-webui

Please make sure you're using the latest version of text-generation-webui.

It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.

  1. Click the Model tab.
  2. Under Download custom model or LoRA, enter TheBloke/Dawn-v2-70B-AWQ.
  3. Click Download.
  4. The model will start downloading. Once it's finished it will say "Done".
  5. In the top left, click the refresh icon next to Model.
  6. In the Model dropdown, choose the model you just downloaded: Dawn-v2-70B-AWQ
  7. Select Loader: AutoAWQ.
  8. Click Load, and the model will load and is now ready for use.
  9. If you want any custom settings, set them and then click Save settings for this model followed by Reload the Model in the top right.
  10. Once you're ready, click the Text Generation tab and enter a prompt to get started!

Multi-user inference server: vLLM

Documentation on installing and using vLLM can be found here.

  • Please ensure you are using vLLM version 0.2 or later.
  • When using vLLM as a server, pass the --quantization awq parameter.

For example:

python3 -m vllm.entrypoints.api_server --model TheBloke/Dawn-v2-70B-AWQ --quantization awq
  • When using vLLM from Python code, again set quantization=awq.

For example:

from vllm import LLM, SamplingParams

prompts = [
    "Tell me about AI",
    "Write a story about llamas",
    "What is 291 - 150?",
    "How much wood would a woodchuck chuck if a woodchuck could chuck wood?",
]
prompt_template=f'''Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{prompt}

### Response:
'''

prompts = [prompt_template.format(prompt=prompt) for prompt in prompts]

sampling_params = SamplingParams(temperature=0.8, top_p=0.95)

llm = LLM(model="TheBloke/Dawn-v2-70B-AWQ", quantization="awq", dtype="auto")

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}")

Multi-user inference server: Hugging Face Text Generation Inference (TGI)

Use TGI version 1.1.0 or later. The official Docker container is: ghcr.io/huggingface/text-generation-inference:1.1.0

Example Docker parameters:

--model-id TheBloke/Dawn-v2-70B-AWQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096

Example Python code for interfacing with TGI (requires huggingface-hub 0.17.0 or later):

pip3 install huggingface-hub

from huggingface_hub import InferenceClient

endpoint_url = "https://your-endpoint-url-here"

prompt = "Tell me about AI"
prompt_template=f'''Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{prompt}

### Response:
'''

client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
                                  max_new_tokens=128,
                                  do_sample=True,
                                  temperature=0.7,
                                  top_p=0.95,
                                  top_k=40,
                                  repetition_penalty=1.1)

print(f"Model output: ", response)

Inference from Python code using AutoAWQ

Install the AutoAWQ package

Requires: AutoAWQ 0.1.1 or later.

pip3 install autoawq

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

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

AutoAWQ example code 

from awq import AutoAWQForCausalLM
from transformers import AutoTokenizer

model_name_or_path = "TheBloke/Dawn-v2-70B-AWQ"

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

prompt = "Tell me about AI"
prompt_template=f'''Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{prompt}

### Response:
'''

print("*** Running model.generate:")

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

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

# Get the tokens from the output, decode them, print them
token_output = generation_output[0]
text_output = tokenizer.decode(token_output)
print("LLM output: ", text_output)

"""
# Inference should be possible with transformers pipeline as well in future
# But currently this is not yet supported by AutoAWQ (correct as of September 25th 2023)
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:

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.

Special thanks to: Aemon Algiz.

Patreon special mentions: Brandon Frisco, LangChain4j, Spiking Neurons AB, transmissions 11, Joseph William Delisle, Nitin Borwankar, Willem Michiel, Michael Dempsey, vamX, Jeffrey Morgan, zynix, jjj, Omer Bin Jawed, Sean Connelly, jinyuan sun, Jeromy Smith, Shadi, Pawan Osman, Chadd, Elijah Stavena, Illia Dulskyi, Sebastain Graf, Stephen Murray, terasurfer, Edmond Seymore, Celu Ramasamy, Mandus, Alex, biorpg, Ajan Kanaga, Clay Pascal, Raven Klaugh, 阿明, K, ya boyyy, usrbinkat, Alicia Loh, John Villwock, ReadyPlayerEmma, Chris Smitley, Cap'n Zoog, fincy, GodLy, S_X, sidney chen, Cory Kujawski, OG, Mano Prime, AzureBlack, Pieter, Kalila, Spencer Kim, Tom X Nguyen, Stanislav Ovsiannikov, Michael Levine, Andrey, Trailburnt, Vadim, Enrico Ros, Talal Aujan, Brandon Phillips, Jack West, Eugene Pentland, Michael Davis, Will Dee, webtim, Jonathan Leane, Alps Aficionado, Rooh Singh, Tiffany J. Kim, theTransient, Luke @flexchar, Elle, Caitlyn Gatomon, Ari Malik, subjectnull, Johann-Peter Hartmann, Trenton Dambrowitz, Imad Khwaja, Asp the Wyvern, Emad Mostaque, Rainer Wilmers, Alexandros Triantafyllidis, Nicholas, Pedro Madruga, SuperWojo, Harry Royden McLaughlin, James Bentley, Olakabola, David Ziegler, Ai Maven, Jeff Scroggin, Nikolai Manek, Deo Leter, Matthew Berman, Fen Risland, Ken Nordquist, Manuel Alberto Morcote, Luke Pendergrass, TL, Fred von Graf, Randy H, Dan Guido, NimbleBox.ai, Vitor Caleffi, Gabriel Tamborski, knownsqashed, Lone Striker, Erik Bjäreholt, John Detwiler, Leonard Tan, Iucharbius

Thank you to all my generous patrons and donaters!

And thank you again to a16z for their generous grant.

Original model card: Undi's Dawn V2 70B

[fp16 - gguf - exl2 : 2.40bpw - 2.55bpw]

Description

This repo contains fp16 files of Dawn-70B, a merge I have done with the new layer shuffle method from mergekit.

UtopiaXL was a huge success for me, I really liked it, so I took the same path to do this 70B: A good base, some psychologic data, some medical data, a little bit of this, of that, and LimaRP at the end as always.

NOTE: This repo contain the file measurement.json needed to do your own exl2 quant (I use wikitext).

Models and loras used

The sauce 

!mergekit-layershuffle ./Dawn-v2-70B \
  --model Sao10K/Euryale-1.3-L2-70B --weight 0.3 \
  --model Xwin-LM/Xwin-LM-70B-V0.1 --weight 0.2 \
  --model ehartford/Samantha-1.11-70b --weight 0.1 \
  --model NousResearch/Nous-Hermes-Llama2-70b --weight 0.05 \
  --model augtoma/qCammel-70-x --weight 0.05 \
  --model jondurbin/airoboros-l2-c70b-3.1.2 --weight 0.2 \
  --model fangloveskari/ORCA_LLaMA_70B_QLoRA --weight 0.1 \
  --write-yaml Dawn-v2-70B.yaml

=========================

merge_method: passthrough
slices:
- sources:
  - layer_range:
    - 0
    - 1
    model: fangloveskari/ORCA_LLaMA_70B_QLoRA
- sources:
  - layer_range:
    - 1
    - 2
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 2
    - 3
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 3
    - 4
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 4
    - 5
    model: fangloveskari/ORCA_LLaMA_70B_QLoRA
- sources:
  - layer_range:
    - 5
    - 6
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 6
    - 8
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 8
    - 9
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 9
    - 10
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 10
    - 11
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 11
    - 12
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 12
    - 13
    model: fangloveskari/ORCA_LLaMA_70B_QLoRA
- sources:
  - layer_range:
    - 13
    - 14
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 14
    - 15
    model: fangloveskari/ORCA_LLaMA_70B_QLoRA
- sources:
  - layer_range:
    - 15
    - 16
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 16
    - 17
    model: fangloveskari/ORCA_LLaMA_70B_QLoRA
- sources:
  - layer_range:
    - 17
    - 18
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 18
    - 19
    model: NousResearch/Nous-Hermes-Llama2-70b
- sources:
  - layer_range:
    - 19
    - 20
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 20
    - 21
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 21
    - 22
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 22
    - 23
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 23
    - 24
    model: augtoma/qCammel-70-x
- sources:
  - layer_range:
    - 24
    - 25
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 25
    - 27
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 27
    - 28
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 28
    - 29
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 29
    - 30
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 30
    - 32
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 32
    - 33
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 33
    - 34
    model: augtoma/qCammel-70-x
- sources:
  - layer_range:
    - 34
    - 35
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 35
    - 37
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 37
    - 38
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 38
    - 39
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 39
    - 40
    model: augtoma/qCammel-70-x
- sources:
  - layer_range:
    - 40
    - 41
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 41
    - 42
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 42
    - 43
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 43
    - 44
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 44
    - 45
    model: NousResearch/Nous-Hermes-Llama2-70b
- sources:
  - layer_range:
    - 45
    - 46
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 46
    - 48
    model: ehartford/Samantha-1.11-70b
- sources:
  - layer_range:
    - 48
    - 49
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 49
    - 50
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 50
    - 51
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 51
    - 54
    model: fangloveskari/ORCA_LLaMA_70B_QLoRA
- sources:
  - layer_range:
    - 54
    - 55
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 55
    - 56
    model: fangloveskari/ORCA_LLaMA_70B_QLoRA
- sources:
  - layer_range:
    - 56
    - 58
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 58
    - 59
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 59
    - 60
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 60
    - 62
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 62
    - 63
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 63
    - 64
    model: fangloveskari/ORCA_LLaMA_70B_QLoRA
- sources:
  - layer_range:
    - 64
    - 65
    model: NousResearch/Nous-Hermes-Llama2-70b
- sources:
  - layer_range:
    - 65
    - 66
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 66
    - 67
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 67
    - 68
    model: augtoma/qCammel-70-x
- sources:
  - layer_range:
    - 68
    - 70
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 70
    - 71
    model: augtoma/qCammel-70-x
- sources:
  - layer_range:
    - 71
    - 72
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 72
    - 73
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 73
    - 75
    model: jondurbin/airoboros-l2-c70b-3.1.2
- sources:
  - layer_range:
    - 75
    - 76
    model: Sao10K/Euryale-1.3-L2-70B
- sources:
  - layer_range:
    - 76
    - 77
    model: augtoma/qCammel-70-x
- sources:
  - layer_range:
    - 77
    - 78
    model: Xwin-LM/Xwin-LM-70B-V0.1
- sources:
  - layer_range:
    - 78
    - 79
    model: NousResearch/Nous-Hermes-Llama2-70b
- sources:
  - layer_range:
    - 79
    - 80
    model: Xwin-LM/Xwin-LM-70B-V0.1


=========================

=> Applying Doctor-Shotgun/limarpv3-llama2-70b-qlora x 0.35

Prompt template: Alpaca 

Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{prompt}

### Response:

A big thanks to Charles for adding the layer shuffle method to his tool mergekit and Henky/KoboldAI for the machine he let me use.

If you want to support me, you can here.

Downloads last month
6
Safetensors
Model size
9.68B params
Tensor type
I32
·
FP16
·
Inference Examples
Text Generation
Inference API (serverless) has been turned off for this model.

Model tree for TheBloke/Dawn-v2-70B-AWQ

Base model

Undi95/Dawn-v2-70B
Quantized
(5)
this model