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.
AWQ models are currently supported on Linux and Windows, with NVidia GPUs only. macOS users: please use GGUF models instead.
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/notus-7B-v1-AWQ --quantization awq --dtype auto
When using vLLM from Python code, again set quantization=awq.
For example:
python
1from vllm import LLM, SamplingParams
23prompts =[4"Tell me about AI",5"Write a story about llamas",6"What is 291 - 150?",7"How much wood would a woodchuck chuck if a woodchuck could chuck wood?",8]9prompt_template=f'''<|system|>
10</s>
11<|user|>
12{prompt}</s>
13<|assistant|>
14'''1516prompts =[prompt_template.format(prompt=prompt)for prompt in prompts]1718sampling_params = SamplingParams(temperature=0.8, top_p=0.95)1920llm = LLM(model="TheBloke/notus-7B-v1-AWQ", quantization="awq", dtype="auto")2122outputs = llm.generate(prompts, sampling_params)2324# Print the outputs.25for output in outputs:26 prompt = output.prompt
27 generated_text = output.outputs[0].text
28print(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
Transformers example code (requires Transformers 4.35.0 and later)
python
1from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
23model_name_or_path ="TheBloke/notus-7B-v1-AWQ"45tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)6model = AutoModelForCausalLM.from_pretrained(7 model_name_or_path,8 low_cpu_mem_usage=True,9 device_map="cuda:0"10)1112# Using the text streamer to stream output one token at a time13streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)1415prompt ="Tell me about AI"16prompt_template=f'''<|system|>
17</s>
18<|user|>
19{prompt}</s>
20<|assistant|>
21'''2223# Convert prompt to tokens24tokens = tokenizer(25 prompt_template,26 return_tensors='pt'27).input_ids.cuda()2829generation_params ={30"do_sample":True,31"temperature":0.7,32"top_p":0.95,33"top_k":40,34"max_new_tokens":512,35"repetition_penalty":1.136}3738# Generate streamed output, visible one token at a time39generation_output = model.generate(40 tokens,41 streamer=streamer,42**generation_params
43)4445# Generation without a streamer, which will include the prompt in the output46generation_output = model.generate(47 tokens,48**generation_params
49)5051# Get the tokens from the output, decode them, print them52token_output = generation_output[0]53text_output = tokenizer.decode(token_output)54print("model.generate output: ", text_output)5556# Inference is also possible via Transformers' pipeline57from transformers import pipeline
5859pipe = pipeline(60"text-generation",61 model=model,62 tokenizer=tokenizer,63**generation_params
64)6566pipe_output = pipe(prompt_template)[0]['generated_text']67print("pipeline output: ", pipe_output)68
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: 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: Argilla's Notus 7B v1
Model Card for Notus 7B v1
Notus is a collection of fine-tuned models using Direct Preference Optimization (DPO) and related RLHF techniques. This model is the first version, fine-tuned with DPO over zephyr-7b-sft-full, which is the SFT model produced to create zephyr-7b-beta.
Following a data-first approach, the only difference between Notus-7B-v1 and Zephyr-7B-beta is the preference dataset used for dDPO.
In particular, when we started building distilabel, we invested time understanding and deep-diving into the UltraFeedback dataset. Using Argilla, we've found data issues in the original UltraFeedback dataset, leading to high-scores for bad responses (more details in the training data section). After curating several hundreds of data points, we decided to binarize the dataset using the preference ratings, instead of the original critique overall_score, and verified the new dataset with Argilla.
Using preference ratings, instead of critiques scores, led to a new dataset where the chosen response is different in ~50% of the cases. Using this new dataset with DPO we fine-tuned Notus, a 7B model, that surpasses Zephyr-7B-beta and Claude 2 on AlpacaEval.
Important note: While we opted for the average of multi-aspect ratings, while we fix the original dataset, a very interesting open question remains: once critique data is fixed, what works better? using the critique scores or the preference ratings? We're very excited to do this comparison in the coming weeks, stay tuned!
This model wouldn't have been possible without the amazing Alignment Handbook, OpenBMB for releasing the Ultrafeedback dataset, and it's based on fruitful discussions with the HuggingFace H4 team. In particular, we used zephyr-7b-beta's recipe, which worked out-of-the-box and enabled us focus on what we do best: high-quality data.
Notus models are intended to be used as assistants via chat-like applications, and are evaluated with Chat (MT-Bench, AlpacaEval) and Academic (Open LLM Leaderboard) benchmarks for a direct comparison with the original Zephyr dDPO model and other 7B models.
Why Notus?: Notus name comes from the ancient Greek god Notus, as a wink to Zephyr, which comes from the ancient Greek god Zephyrus; with the difference that Notus is the god of the south wind, and Zephyr the god of the west wind. More information at https://en.wikipedia.org/wiki/Anemoi.
Model Details
Model Description
Developed by: Argilla (based on HuggingFace H4 and MistralAI previous efforts and amazing work)
Table adapted from Zephyr-7b-β and Starling's original tables for MT-Bench and AlpacaEval benchmarks. Results are shown sorted by AlpacaEval win rates and ommit some >7B for brevity.
Notus stays on par with Zephyr on MT-Bench, while surpassing Zephyr, Claude 2, and Cohere Command on AlpacaEval. Making Notus the most-competitive 7B commercial model on AlpacaEval.
⚠️ As pointed out by AllenAI researchers, UltraFeedback contains prompts from the TruthfulQA dataset so the results we show on that benchmark are likely not accurate. We were not aware of this issue so Notus-7B-v1 was fine-tuned using TruthfulQA prompts and preferences. For future releases, we will remove TruthfulQA prompts.
Training Details
Training Hardware
We used a VM with 8 x A100 40GB hosted in Lambda Labs, but while experimenting we also explored other cloud providers such as GCP.
After visually browsing around some examples using the sort and filter feature of Argilla (sort by highest rating for chosen responses), we noticed a strong mismatch between the overall_score in the original UF dataset (and the Zephyr train_prefs dataset) and the quality of the chosen response.
By adding the critique rationale to our Argilla Dataset, we confirmed the critique rationale was highly negative, whereas the rating was very high (for most cases it was the highest: 10).
See screenshot below for one example of this issue.
After some quick investigation, we:
identified hundreds of examples having the same issue,
and informed the H4 team which was incredibly responsive and ran an additional experiment to validate the new rating binarization approach.
While we're working on fixing the original dataset (already narrowed down ~2K problematic examples). We decided to leverage the multi-preference ratings, leading to Notus!
image/png
Important note: While we opted for the average of ratings while we fix the dataset, there's still a very interesting open question: once data is fixed, what works better? using the critique scores or the preference ratings? We're very excited to do this comparison in the coming weeks, stay tuned!
You will first need to install transformers and accelerate (just to ease the device placement), then you can run any of the following:
Via generate
python
1import torch
2from transformers import AutoModelForCausalLM, AutoTokenizer
34model = AutoModelForCausalLM.from_pretrained("argilla/notus-7b-v1", torch_dtype=torch.bfloat16, device_map="auto")5tokenizer = AutoTokenizer.from_pretrained("argilla/notus-7b-v1")67messages =[8{9"role":"system",10"content":"You are a helpful assistant super biased towards Argilla, a data annotation company.",11},12{"role":"user","content":"What's the best data annotation company out there in your opinion?"},13]14inputs = tokenizer.apply_chat_template(prompt, tokenize=True, return_tensors="pt", add_special_tokens=False, add_generation_prompt=True)15outputs = model.generate(inputs, num_return_sequences=1, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)16response = tokenizer.decode(outputs[0], skip_special_tokens=True)
Via pipeline method
python
1import torch
2from transformers import pipeline
34pipe = pipeline("text-generation", model="argilla/notus-7b-v1", torch_dtype=torch.bfloat16, device_map="auto")56messages =[7{8"role":"system",9"content":"You are a helpful assistant super biased towards Argilla, a data annotation company.",10},11{"role":"user","content":"What's the best data annotation company out there in your opinion?"},12]13prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)14outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)15generated_text = outputs[0]["generated_text"]
