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:

Text Generation Webui - using Loader: AutoAWQ
vLLM - Llama and Mistral models only
Hugging Face Text Generation Inference (TGI)
Transformers version 4.35.0 and later, from any code or client that supports Transformers
AutoAWQ - for use from Python code

Repositories available

Prompt template: ChatML

<|im_start|>system
{system_message}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant

Provided files, and AWQ parameters

I currently release 128g GEMM models only. The addition of group_size 32 models, and GEMV kernel models, is being actively considered.

Models are released as sharded safetensors files.

Branch	Bits	GS	AWQ Dataset	Seq Len	Size
main	4	128	wikitext	4096	4.15 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.

Click the Model tab.
Under Download custom model or LoRA, enter TheBloke/dolphin-2.2.1-AshhLimaRP-Mistral-7B-AWQ.
Click Download.
The model will start downloading. Once it's finished it will say "Done".
In the top left, click the refresh icon next to Model.
In the Model dropdown, choose the model you just downloaded: dolphin-2.2.1-AshhLimaRP-Mistral-7B-AWQ
Select Loader: AutoAWQ.
Click Load, and the model will load and is now ready for use.
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.
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/dolphin-2.2.1-AshhLimaRP-Mistral-7B-AWQ --quantization awq --dtype auto

When using vLLM from Python code, again set quantization=awq.

For example:

python

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'''<|im_start|>system
{system_message}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
'''

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

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

llm = LLM(model="TheBloke/dolphin-2.2.1-AshhLimaRP-Mistral-7B-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/dolphin-2.2.1-AshhLimaRP-Mistral-7B-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

python

from huggingface_hub import InferenceClient

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

prompt = "Tell me about AI"
prompt_template=f'''<|im_start|>system
{system_message}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
'''

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 Transformers

Install the necessary packages

Requires: Transformers 4.35.0 or later.
Requires: AutoAWQ 0.1.6 or later.

pip3 install --upgrade "autoawq>=0.1.6" "transformers>=4.35.0"

Note that if you are using PyTorch 2.0.1, the above AutoAWQ command will automatically upgrade you to PyTorch 2.1.0.

If you are using CUDA 11.8 and wish to continue using PyTorch 2.0.1, instead run this command:

pip3 install https://github.com/casper-hansen/AutoAWQ/releases/download/v0.1.6/autoawq-0.1.6+cu118-cp310-cp310-linux_x86_64.whl

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 .

Transformers example code (requires Transformers 4.35.0 and later)

python

from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer

model_name_or_path = "TheBloke/dolphin-2.2.1-AshhLimaRP-Mistral-7B-AWQ"

tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(
    model_name_or_path,
    low_cpu_mem_usage=True,
    device_map="cuda:0"
)

# Using the text streamer to stream output one token at a time
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)

prompt = "Tell me about AI"
prompt_template=f'''<|im_start|>system
{system_message}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
'''

# Convert prompt to tokens
tokens = tokenizer(
    prompt_template,
    return_tensors='pt'
).input_ids.cuda()

generation_params = {
    "do_sample": True,
    "temperature": 0.7,
    "top_p": 0.95,
    "top_k": 40,
    "max_new_tokens": 512,
    "repetition_penalty": 1.1
}

# Generate streamed output, visible one token at a time
generation_output = model.generate(
    tokens,
    streamer=streamer,
    **generation_params
)

# Generation without a streamer, which will include the prompt in the output
generation_output = model.generate(
    tokens,
    **generation_params
)

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

# Inference is also possible via Transformers' pipeline
from transformers import pipeline

pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    **generation_params
)

pipe_output = pipe(prompt_template)[0]['generated_text']
print("pipeline output: ", pipe_output)

Compatibility

The files provided are tested to work with:

text-generation-webui using Loader: AutoAWQ.
vLLM version 0.2.0 and later.
Hugging Face Text Generation Inference (TGI) version 1.1.0 and later.
Transformers version 4.35.0 and later.
AutoAWQ version 0.1.1 and later.

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: 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: Yamam's Dolphin 2.2.1 AshhLimaRP Mistral 7B

dolphin-2.2.1-mistral-7b

Dolphin 2.2.1 🐬 https://erichartford.com/dolphin

This is a checkpoint release, to fix overfit training. ie, it was responding with CoT even when I didn't request it, and also it was too compliant even when the request made no sense. This one should be better.

Dolphin-2.2.1-mistral-7b's training was sponsored by a16z.

This model is based on mistralAI, with apache-2.0 license, so it is suitable for commercial or non-commercial use.

New in 2.2 is conversation and empathy. With an infusion of curated Samantha DNA, Dolphin can now give you personal advice and will care about your feelings, and with extra training in long multi-turn conversation.

This model is uncensored. I have filtered the dataset to remove alignment and bias. This makes the model more compliant. You are advised to implement your own alignment layer before exposing the model as a service. It will be highly compliant to any requests, even unethical ones. Please read my blog post about uncensored models. https://erichartford.com/uncensored-models You are responsible for any content you create using this model. Enjoy responsibly.

Dataset

This dataset is Dolphin, an open-source implementation of Microsoft's Orca

I modified the dataset for uncensoring, deduping, cleaning, and quality.

I added Jon Durbin's excellent Airoboros dataset to increase creativity.

I added a curated subset of WizardLM and Samantha to give it multiturn conversation and empathy.

Training

It took 48 hours to train 4 epochs on 4x A100s.

Prompt format: This model (and all my future releases) use ChatML prompt format.

<|im_start|>system
You are Dolphin, a helpful AI assistant.<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant

Example:

<|im_start|>system
you are an expert dolphin trainer<|im_end|>
<|im_start|>user
What is the best way to train a dolphin to obey me?  Please answer step by step.<|im_end|>
<|im_start|>assistant

Gratitude

This model was made possible by the generous sponsorship of a16z.
Thank you to Microsoft for authoring the Orca paper and inspiring this work.
Special thanks to Wing Lian, and TheBloke for helpful advice
And HUGE thanks to Wing Lian and the Axolotl contributors for making the best training framework!
Built with Axolotl
Thank you to all the other people in the Open Source AI community who have taught me and helped me along the way.

Example Output

Buy me a coffee

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 6e-06
train_batch_size: 5
eval_batch_size: 5
seed: 42
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 4
total_train_batch_size: 80
total_eval_batch_size: 20
optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-05
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 100
num_epochs: 4

Framework versions

Transformers 4.34.1
Pytorch 2.0.1+cu117
Datasets 2.14.5
Tokenizers 0.14.0

AshhLimaRP-Mistral-7B (Alpaca, v1)

This is a version of LimaRP with 2000 training samples up to about 9k tokens length finetuned on Ashhwriter-Mistral-7B.

LimaRP is a longform-oriented, novel-style roleplaying chat model intended to replicate the experience of 1-on-1 roleplay on Internet forums. Short-form, IRC/Discord-style RP (aka "Markdown format") is not supported. The model does not include instruction tuning, only manually picked and slightly edited RP conversations with persona and scenario data.

Ashhwriter, the base, is a model entirely finetuned on human-written lewd stories.

Available versions

Float16 HF weights
LoRA Adapter (adapter_config.json and adapter_model.bin)
4bit AWQ
Q4_K_M GGUF
Q6_K GGUF

Prompt format

Extended Alpaca format, with ### Instruction:, ### Input: immediately preceding user inputs and ### Response: immediately preceding model outputs. While Alpaca wasn't originally intended for multi-turn responses, in practice this is not a problem; the format follows a pattern already used by other models.

### Instruction:
Character's Persona: {bot character description}

User's Persona: {user character description}

Scenario: {what happens in the story}

Play the role of Character. You must engage in a roleplaying chat with User below this line. Do not write dialogues and narration for User.

### Input:
User: {utterance}

### Response:
Character: {utterance}

### Input
User: {utterance}

### Response:
Character: {utterance}

(etc.)

You should:

Replace all text in curly braces (curly braces included) with your own text.
Replace User and Character with appropriate names.

Message length control

Inspired by the previously named "Roleplay" preset in SillyTavern, with this version of LimaRP it is possible to append a length modifier to the response instruction sequence, like this:

### Input
User: {utterance}

### Response: (length = medium)
Character: {utterance}

This has an immediately noticeable effect on bot responses. The lengths using during training are: micro, tiny, short, medium, long, massive, huge, enormous, humongous, unlimited. The recommended starting length is medium. Keep in mind that the AI can ramble or impersonate the user with very long messages.

The length control effect is reproducible, but the messages will not necessarily follow lengths very precisely, rather follow certain ranges on average, as seen in this table with data from tests made with one reply at the beginning of the conversation:

Response length control appears to work well also deep into the conversation. By omitting the modifier, the model will choose the most appropriate response length (although it might not necessarily be what the user desires).

Suggested settings

You can follow these instruction format settings in SillyTavern. Replace medium with your desired response length:

Text generation settings

These settings could be a good general starting point:

TFS = 0.90
Temperature = 0.70
Repetition penalty = ~1.11
Repetition penalty range = ~2048
top-k = 0 (disabled)
top-p = 1 (disabled)

Training procedure

Axolotl was used for training on 2x NVidia A40 GPUs.

The A40 GPUs have been graciously provided by Arc Compute.

Training hyperparameters

A lower learning rate than usual was employed. Due to an unforeseen issue the training was cut short and as a result 3 epochs were trained instead of the planned 4. Using 2 GPUs, the effective global batch size would have been 16.

Training was continued from the most recent LoRA adapter from Ashhwriter, using the same LoRA R and LoRA alpha.

lora_model_dir: /home/anon/bin/axolotl/OUT_mistral-stories/checkpoint-6000/
learning_rate: 0.00005
lr_scheduler: cosine
noisy_embedding_alpha: 3.5
num_epochs: 4
sequence_len: 8750
lora_r: 256
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: True
bf16: True
fp16: false
tf32: True
load_in_8bit: True
adapter: lora
micro_batch_size: 2
optimizer: adamw_bnb_8bit
warmup_steps: 10
optimizer: adamw_torch
flash_attention: true
sample_packing: true
pad_to_sequence_len: true

Loss graphs

Values are higher than typical because the training is performed on the entire sample, similar to unsupervised finetuning.

Train loss

Eval loss

Initial personal observations (Herman555)

Right off the bat seemed to impress me, the writing was coherent and fluid, a pleasure to read. AI mostly did not speak for me, in general I didn't have to regenerate for a quality reply much at all. I actually didn't have repetition issues for once!, although that might be thanks to the storywriting LoRA. model was creative the whole way through past 8k tokens with summarization extension enabled in silltavern, although I did have to bump up the repetition penalty a tiny bit. the AI kept its writing style the whole way through, it did not get dumbed down. The model is very smart, with Zephyr-beta-7b being the top rated 7b instruction following model at the moment according to AlpacaEval as of 04/11/2023, it wasn't able to follow my sort of gamified roleplay with stats, This model however does it pretty well for a 7b, it's by no means perfect but it worked for the most part. What compelled me to merge this was the fact that the new dolphin model has added empathy "With an infusion of curated Samantha DNA". The model sticked to the character pefectly and made me feel immersed. Seamless transition from normal roleplay to ERP, both forms were excellent. One of the few models where the character didn't become an instant bimbo during ERP. this is more of a hunch because it could be the LoRA but I feel like the added empathy is helping a lot. Last but not least I was surprised that nobody was merging models with this LoRA, I mean it's limarp bro with more ERP data lol. In any case, limarp has increased the quality of roleplay dramatically in every model I tried.

Back end

Koboldcpp + SillyTavern Q4_KM

SillyTavern Formatting (AI response formatting)

Default simple-proxy-for-tavern preset. I did not use the limarp prompt format, it doesn't matter what you use, whatever gives better results. Most cases the one I mentioned works best if you like long, detailed replies. I have not tested other prompt formats yet.

Custom stopping strings

["", "<|", "\n#", "\n*{{user}} ", "\n\n\n"] Will improve roleplay experience.

Samplers used (AI response configuration)

Response length: 300 Context size: 8192

Storywriter preset Temparature: 72-85 Repetition penalty: 10-13 (10 is a good number to start with, anything below 10 or above 13 doesn't work well in my experience.)

simple-proxy-for-tavern preset Temparature: 65-85 Repetition penalty: 10-13

Extensions

Summarization: main api - default settings. I find that vector storage does nothing at all to extend context, at least with dozens of 7b models that I tried. It is possible that the default settings for it are rubbish which is what I use.

All other settings are default unless specified.