TheBloke's LLM work is generously supported by a grant from andreessen horowitz (a16z)
SellarX 4B V0.2 - GPTQ
- Model creator: Dampish
- Original model: SellarX 4B V0.2
Description
This repo contains GPTQ model files for Dampish's SellarX 4B V0.2.
Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
Repositories available
- GPTQ models for GPU inference, with multiple quantisation parameter options.
- Dampish's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions
Prompt template: Unknown
{prompt}
Provided files and GPTQ parameters
Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
Each separate quant is in a different branch. See below for instructions on fetching from different branches.
All recent GPTQ files are made with AutoGPTQ, and all files in non-main branches are made with AutoGPTQ. Files in the main branch which were uploaded before August 2023 were made with GPTQ-for-LLaMa.
Explanation of GPTQ parameters
- Bits: The bit size of the quantised model.
- GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
- Act Order: True or False. Also known as
desc_act. True results in better quantisation accuracy. Some GPTQ clients have had issues with models that use Act Order plus Group Size, but this is generally resolved now. - Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
- GPTQ dataset: The dataset used for quantisation. Using a dataset more appropriate to the model's training can improve quantisation accuracy. Note that the GPTQ dataset is not the same as the dataset used to train the model - please refer to the original model repo for details of the training dataset(s).
- Sequence Length: The length of the dataset sequences used for quantisation. Ideally this is the same as the model sequence length. For some very long sequence models (16+K), a lower sequence length may have to be used. Note that a lower sequence length does not limit the sequence length of the quantised model. It only impacts the quantisation accuracy on longer inference sequences.
- ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama models in 4-bit.
| Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
|---|---|---|---|---|---|---|---|---|---|
| main | 4 | 128 | Yes | 0.1 | wikitext | 2048 | 1.83 GB | No | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. |
| gptq-4bit-32g-actorder_True | 4 | 32 | Yes | 0.1 | wikitext | 2048 | 1.98 GB | No | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. |
| gptq-8bit-128g-actorder_True | 8 | 128 | Yes | 0.1 | wikitext | 2048 | 3.10 GB | No | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. |
| gptq-8bit-32g-actorder_True | 8 | 32 | Yes | 0.1 | wikitext | 2048 | 3.27 GB | No | 8-bit, with group size 32g and Act Order for maximum inference quality. |
How to download from branches
- In text-generation-webui, you can add
:branchto the end of the download name, egTheBloke/StellarX-4B-V0.2-GPTQ:main - With Git, you can clone a branch with:
git clone --single-branch --branch main https://huggingface.co/TheBloke/StellarX-4B-V0.2-GPTQ
- In Python Transformers code, the branch is the
revisionparameter; see below.
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/StellarX-4B-V0.2-GPTQ.
- To download from a specific branch, enter for example
TheBloke/StellarX-4B-V0.2-GPTQ:main - see Provided Files above for the list of branches for each option.
- 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:
StellarX-4B-V0.2-GPTQ - The model will automatically 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.
- Note that you do not need to and should not set manual GPTQ parameters any more. These are set automatically from the file
quantize_config.json.
- Once you're ready, click the Text Generation tab and enter a prompt to get started!
How to use this GPTQ model from Python code
Install the necessary packages
Requires: Transformers 4.32.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
pip3 install transformers>=4.32.0 optimum>=1.12.0
pip3 install auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/ # Use cu117 if on CUDA 11.7
If you have problems installing AutoGPTQ using the pre-built wheels, install it from source instead:
pip3 uninstall -y auto-gptq
git clone https://github.com/PanQiWei/AutoGPTQ
cd AutoGPTQ
pip3 install .
For CodeLlama models only: you must use Transformers 4.33.0 or later.
If 4.33.0 is not yet released when you read this, you will need to install Transformers from source:
pip3 uninstall -y transformers
pip3 install git+https://github.com/huggingface/transformers.git
You can then use the following code
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
model_name_or_path = "TheBloke/StellarX-4B-V0.2-GPTQ"
# To use a different branch, change revision
# For example: revision="main"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
device_map="auto",
trust_remote_code=False,
revision="main")
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
prompt = "Tell me about AI"
prompt_template=f'''{prompt}
'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' 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 AutoGPTQ, both via Transformers and using AutoGPTQ directly. They should also work with Occ4m's GPTQ-for-LLaMa fork.
ExLlama is compatible with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
Huggingface Text Generation Inference (TGI) is compatible with all GPTQ models.
Discord
For further support, and discussions on these models and AI in general, join us at:
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: Dampish's SellarX 4B V0.2
StellarX: A Base Model by Dampish and Arkane
StellarX is a powerful autoregressive language model designed for various natural language processing tasks. It has been trained on a massive dataset containing 810 billion tokens(trained on 300B tokens), trained on "redpajama," and is built upon the popular GPT-NeoX architecture. With approximately 4 billion parameters, StellarX offers exceptional performance and versatility.
Model Details
- Training Data: StellarX is trained on a large-scale dataset provided by "redpajama" maintained by the group "togethercumputer." This dataset has been instrumental in shaping StellarX's language capabilities and general-purpose understanding.
- Model Architecture: StellarX is built upon the GPT-NeoX architecture, which may, be, inspired by GPT-3 and shares similarities with GPT-J-6B. The architecture incorporates key advancements in transformer-based language models, ensuring high-quality predictions and contextual understanding.
- Model Size: StellarX consists of approximately 4 billion parameters, making it a highly capable language model for a wide range of natural language processing tasks.
- Carbon-Friendly and Resource-Efficient: StellarX has been optimized for carbon efficiency and can be comfortably run on local devices. When loaded in 8 bits, the model requires only about 5GB of storage, making it more accessible and convenient for various applications.
- V0.2 Meaning what version it is on, currently version 0.2, Assume version 0.2 has only been trained on 300B tokens and the goal is 810B tokens. The next version aims to have a way higher accuracy.
How to Use
To load StellarX using the Hugging Face Transformers library, you can use the following code snippet:
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("Dampish/StellarX-4B-V0")
model = AutoModelForCausalLM.from_pretrained("Dampish/StellarX-4B-V0")
This model is particularly beneficial for those seeking a language model that is powerful, compact, and can be run on local devices without a hefty carbon footprint. Remember, when considering Darius1, it's not just about the impressive numbers—it's about what these numbers represent: powerful performance, optimized resources, and responsible computing.
For any queries related to this model, feel free to reach out to "Dampish#3607" on discord.
Licensing and Usage
StellarX, developed by the Dampish, is made available under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC-BY-NC-SA-4.0). This license ensures that you can utilize the model for research purposes and personal use without any restrictions, while also promoting the sharing and adaptation of the model under certain conditions.
Research and Personal Use
StellarX can be freely used for research purposes, allowing you to explore its capabilities, conduct experiments, and develop novel applications. Whether you're a student, researcher, or hobbyist, the model's availability under the CC-BY-NC-SA-4.0 license empowers you to unlock the potential of StellarX for your own non-commercial projects.
Commercial Usage
For commercial usage of StellarX, an additional licensing arrangement must be established. If you intend to leverage the model for any commercial purpose, such as integrating it into a product or service, you are required to reach an agreement with the Dampish. This agreement will specify the terms, including the agreed-upon percentage or licensing fee to be paid for the commercial use of StellarX.
To initiate discussions regarding commercial usage, please contact Dampish through the designated channels mentioned earlier. They will be able to provide you with further information and guide you through the process of establishing a licensing arrangement tailored to your specific requirements.
Importance of Licensing Compliance
It is crucial to respect the licensing terms to ensure the fair usage and continued development of StellarX. The revenue generated from commercial licensing supports the efforts of the Dampish in advancing the model and making it more widely accessible.
Note on CC-BY-NC-SA-4.0
Under the CC-BY-NC-SA-4.0 license, you are allowed to modify and adapt StellarX, incorporating it into your own projects. However, any derivative work or modifications should also be shared under the same license terms, ensuring the continued openness and collaborative spirit of the project.
Please review the complete text of the CC-BY-NC-SA-4.0 license to familiarize yourself with its provisions and requirements. It is essential to comply with the terms of the license to respect the intellectual property rights and contributions of the Dampish and the wider community involved in developing StellarX.
GPT-NeoX and Model Selection
GPT-NeoX-20B, a sibling model to StellarX, is a 20 billion parameter autoregressive language model trained on the Pile using the GPT-NeoX library. StellarX draws inspiration from the architectural advancements and performance of GPT-NeoX models. While the specifics of StellarX's architecture and parameters may differ, it benefits from the proven capabilities of GPT-NeoX and its suitability for diverse natural language processing tasks.
Training and Evaluation
StellarX's training dataset comprises a comprehensive collection of English-language texts, covering various domains, thanks to the efforts of "redpajama" dataset by the group "togethercumputer" group.
Evaluation of GPT-NeoX 20B performance has demonstrated its competence across different natural language tasks. Although since this description provides a brief summary, we refer to the GPT-NeoX Paper https://arxiv.org/abs/2204.06745, comparing GPT-NeoX 20B to other models on tasks such as OpenAI's LAMBADA, SciQ, PIQA, TriviaQA, and ARC Challenge.
Limitations and Considerations
StellarX, like its sibling models, is intended primarily for research purposes. It provides a powerful foundation for extracting useful features and insights from the English language. While StellarX can be further fine-tuned and adapted for deployment, users should conduct their own risk and bias assessments before using it as a basis for downstream tasks.
It's important to note that StellarX is not intended for direct deployment without supervision. It is not designed for human-facing interactions, unlike models like ChatGPT, which have been fine-tuned using reinforcement learning from human feedback to better understand human instructions and dialogue.
Furthermore, StellarX is not limited to the English language if trained properly and can sometimes be used for translation aswell as text generation in other languages.
Lastly, users should be aware of potential biases and limitations inherent in
Special thanks to the group that created the training dataset. The Redpajama dataset, used to train StellarX, thank you togethercumputer.
Community and Support
To inquire about StellarX and receive support, you can join the Dampish's server and engage in discussions in the #questions channel. It is recommended to explore the existing documentation and resources available for GPT-NeoX-20B to familiarize yourself with the model before seeking assistance on. For better information about GPT-NeoX, you can reach out to eleutherAI.
Summary
StellarX, a base language model developed by the Dampish, offers impressive language capabilities and flexibility. Trained on an extensive dataset and built upon the GPT-NeoX architecture, StellarX excels in various natural language processing tasks. Its carbon-friendly and resource-efficient design makes it accessible for local device deployment. Researchers and enthusiasts can freely explore StellarX for research purposes and personal use, while commercial users should adhere to the licensing terms.
Again i am really grateful for the data made by togethercumputers and their willingness to opensource, they inspired this project and sparked the idea in Stellar-models, i am truly really really grateful to them. -dampish
Discord: https://discord.gg/vasyNnUa
OR Reach out to me personally on Discord via the username: Dampish#3607
Thank you for your time.
- Downloads last month
- 4