stabilityai / ar-stablelm-2-chat

huggingface.co
Total runs: 462
24-hour runs: 100
7-day runs: 118
30-day runs: 271
Model's Last Updated: 12月 06 2024

Introduction of ar-stablelm-2-chat

Model Details of ar-stablelm-2-chat

Arabic StableLM 2 Chat 1.6B

Model Description

Arabic Stable LM 2 Chat 1.6B is a 1.6 billion parameter instruction tuned language model from the ar-stablelm-base. The model is trained on a mix of publicly available datasets and synthetic datasets.

Usage

StableLM 2 1.6B Chat uses the following ChatML format:

from transformers import AutoModelForCausalLM, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained('stabilityai/ar-stablelm-2-chat')
model = AutoModelForCausalLM.from_pretrained(
    'stabilityai/ar-stablelm-2-chat',
    device_map="auto",
)

prompt = [{'role': 'user', 'content': 'ماهو إعراب الجملة: ذهبت إلى السوق'}]
inputs = tokenizer.apply_chat_template(
    prompt,
    add_generation_prompt=True,
    return_tensors='pt'
)

tokens = model.generate(
    inputs.to(model.device),
    max_new_tokens=100,
    temperature=0.7,
    do_sample=True
)
output = tokenizer.decode(tokens[:, inputs.shape[-1]:][0], skip_special_tokens=False)

print(output)
Model Details
Training Dataset

The dataset is comprised of a mixture of open datasets large-scale datasets available on the HuggingFace Hub :

Evaluation

We use the following datasets for evaluation :

Below we list the a comparison of our models against 28 models

Model Params CIDAR-MCQ ArabicMMLU ACVA AlGhafa Average
AraGPT2-base 135M 39.0 31.7 51.7 37.9 40.1
AraT5v2-base-1024 220M 38.0 28.3 37.3 36.8 35.1
AraGPT2-medium 370M 39.0 32.2 44.5 38.3 38.5
jais-family-590m 590M 33.0 35.7 51.2 38.8 39.7
jais-family-590m-chat 590M 28.0 31.4 51.8 36.8 37.0
AraGPT2-large 792M 37.0 32.6 46.2 37.9 38.4
AraGPT2-mega 1.46B 40.0 33.3 51.0 38.4 40.7
jais-family-1p3b 1.3B 41.0 37.5 51.3 39.6 42.3
jais-family-1p3b-chat 1.3B 30.0 34.6 56.0 44.7 41.3
Qwen2-1.5B 1.5B 30.0 30.5 61.1 38.7 40.1
Qwen2-1.5B-Instruct 1.5B 28.0 30.8 63.9 38.9 40.4
bloom-1b7 1.72B 37.0 33.0 52.2 39.1 40.3
bloomz-1b7 1.72B 25.0 33.5 45.3 42.1 36.5
jais-family-2p7b 2.7B 43.0 39.0 48.5 40.9 42.8
jais-family-2p7b-chat 2.7B 36.0 36.1 49.5 43.7 41.3
jais-family-6p7b 6.7B 44.0 40.7 55.8 41.3 45.4
jais-family-6p7b-chat 6.7B 38.0 38.7 64.1 39.9 45.2
AceGPT-7B 7B 46.0 40.9 63.1 40.7 47.7
AceGPT-7B-chat 7B 37.0 38.2 65.9 45.1 46.6
SILMA-9B-Instruct-v1.0 9B 27.0 30.8 56.2 42.0 39.0
AceGPT-13B 13B 41.0 40.7 65.3 38.8 46.4
AceGPT-13B-chat 13B 42.0 41.1 70.2 41.6 48.7
AceGPT-v1.5-13B 13B 40.0 40.3 66.1 39.5 46.5
AceGPT-v1.5-13B-chat 13B 36.0 40.9 69.0 39.8 46.4
jais-13b 13B 47.0 40.9 59.6 41.2 47.2
jais-13b-chat 13B 41.0 40.7 61.2 41.4 46.1
jais-family-13b 13B 45.0 41.9 58.3 40.5 46.4
jais-family-13b-chat 13B 37.0 39.9 63.2 41.4 45.4
ar-stablelm-2-base 1.64B 43.0 41.1 52.0 43.9 45.0
ar-stablelm-2-chat 1.64B 46.0 45.5 57.0 50.1 49.6
Use and Limitations
Intended Use

The model is intended to be used in chat-like applications for research only. Users should evaluate the model for safety performance in their specific use case and apply the necessary safeguards and fine-tune the model to facilitate safe performance in downstream applications. Read more about safety and limitations below.

Out-of-scope Use

Out-of-scope uses include use in any manner that violates applicable laws or regulations, Stability AI’s Acceptable Use Policy or license agreement, or use in languages outside of those explicitly supported by this model.

Limitations and Bias

As a base model, this model may exhibit unreliable or other undesirable behaviors that should be corrected through evaluation and fine-tuning prior to deployment. Given that each use case is unique, running a suite of tests may help facilitate proper performance of this model. Using this model will require guardrails around the user’s inputs and outputs to ensure that any outputs returned are not harmful. Pairing this model with an input and output classifier may help prevent harmful responses. Users should exercise caution when using these models in production systems and should not use the models if they are unsuitable for the user’s application.

How to Cite
@misc{alyafeai2024arabicstablelmadapting,
      title={Arabic Stable LM: Adapting Stable LM 2 1.6B to Arabic}, 
      author={Zaid Alyafeai and Michael Pieler and Hannah Teufel and Jonathan Tow and Marco Bellagente and Duy Phung and Nikhil Pinnaparaju and Reshinth Adithyan and Paulo Rocha and Maksym Zhuravinskyi and Carlos Riquelme},
      year={2024},
      eprint={2412.04277},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2412.04277}, 
}

Runs of stabilityai ar-stablelm-2-chat on huggingface.co

462
Total runs
100
24-hour runs
106
3-day runs
118
7-day runs
271
30-day runs

More Information About ar-stablelm-2-chat huggingface.co Model

More ar-stablelm-2-chat license Visit here:

https://choosealicense.com/licenses/other

ar-stablelm-2-chat huggingface.co

ar-stablelm-2-chat huggingface.co is an AI model on huggingface.co that provides ar-stablelm-2-chat's model effect (), which can be used instantly with this stabilityai ar-stablelm-2-chat model. huggingface.co supports a free trial of the ar-stablelm-2-chat model, and also provides paid use of the ar-stablelm-2-chat. Support call ar-stablelm-2-chat model through api, including Node.js, Python, http.

ar-stablelm-2-chat huggingface.co Url

https://huggingface.co/stabilityai/ar-stablelm-2-chat

stabilityai ar-stablelm-2-chat online free

ar-stablelm-2-chat huggingface.co is an online trial and call api platform, which integrates ar-stablelm-2-chat's modeling effects, including api services, and provides a free online trial of ar-stablelm-2-chat, you can try ar-stablelm-2-chat online for free by clicking the link below.

stabilityai ar-stablelm-2-chat online free url in huggingface.co:

https://huggingface.co/stabilityai/ar-stablelm-2-chat

ar-stablelm-2-chat install

ar-stablelm-2-chat is an open source model from GitHub that offers a free installation service, and any user can find ar-stablelm-2-chat on GitHub to install. At the same time, huggingface.co provides the effect of ar-stablelm-2-chat install, users can directly use ar-stablelm-2-chat installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

ar-stablelm-2-chat install url in huggingface.co:

https://huggingface.co/stabilityai/ar-stablelm-2-chat

Url of ar-stablelm-2-chat

ar-stablelm-2-chat huggingface.co Url

Provider of ar-stablelm-2-chat huggingface.co

stabilityai
ORGANIZATIONS

Other API from stabilityai

huggingface.co

Total runs: 161.5K
Run Growth: 26.8K
Growth Rate: 17.30%
Updated: 8月 04 2023
huggingface.co

Total runs: 129.7K
Run Growth: 2.8K
Growth Rate: 2.13%
Updated: 7月 10 2024
huggingface.co

Total runs: 33.7K
Run Growth: 3.4K
Growth Rate: 9.99%
Updated: 8月 09 2024
huggingface.co

Total runs: 363
Run Growth: -97.9K
Growth Rate: -26316.13%
Updated: 8月 03 2024
huggingface.co

Total runs: 0
Run Growth: 0
Growth Rate: 0.00%
Updated: 7月 10 2024
huggingface.co

Total runs: 0
Run Growth: 0
Growth Rate: 0.00%
Updated: 4月 13 2024