JAIS Intiative: Atlas-Chat Models

Model Overview

Atlas-Chat is a family of open models instruction-tuned for Darija, the colloquial Arabic of Morocco, developed as part of the Jais project for standard Arabic and its extentions to dialectal Arabic. These models are designed for language generation and excel in various applications such as question answering, summarization, and translation. Thanks to their compact size, Atlas-Chat models can be deployed in resource-constrained environments like laptops, desktops, or personal cloud setups, making advanced AI accessible to Darija speakers and promoting widespread innovation. Three sizes are available:

  • Atlas-Chat-2B: A small-sized version with 2 billion parameters, capable of generating fluent Moroccan Darija text while maintaining efficiency.
  • Atlas-Chat-9B: A medium-sized with 9 billion parameters, providing more nuanced, contextually rich language generation for complex tasks.
  • Atlas-Chat-27B: A large-sized version with 27 billion parameters, offering even more advanced capabilities for complex tasks and nuanced language generation compared to the 2B and 9B versions.

The models are designed to assist with:

  • Conversational agents and chatbots that operate in Darija.
  • Translation, summarization, and content generation in informal dialect.
  • Cultural research related to Morocco and its language.

Paper: Atlas-Chat: Adapting Large Language Models for Low-Resource Moroccan Arabic Dialect

๐Ÿ‘ฅ Our Team

The model is developed by MBZUAI France Lab, an AI research center in Paris affiliated with the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) headquartered in Abu Dhabi.

Usage

Below we share some code snippets on how to get quickly started with running the model. First, install the Transformers library with:

pip install -U transformers sentencepiece

Then, copy the snippet from the section that is relevant for your use case.

Running with the pipeline API

import torch
from transformers import pipeline

pipe = pipeline(
    "text-generation",
    model="MBZUAI-Paris/Atlas-Chat-9B",
    model_kwargs={"torch_dtype": torch.bfloat16},
    device="cuda" # replace with "mps" to run on a Mac device
)

messages = [
    {"role": "user", "content": 'ุดูƒูˆู† ู„ูŠ ุตู†ุนูƒุŸ'},
]

outputs = pipe(messages, max_new_tokens=256, temperature=0.0)
assistant_response = outputs[0]["generated_text"][-1]["content"].strip()
print(assistant_response)
  • Response:

ุตู†ุนุงุชู†ูŠ ุฌุงู…ุนุฉ ู…ุญู…ุฏ ุจู† ุฒุงูŠุฏ ู„ู„ุฐูƒุงุก ุงู„ุงุตุทู†ุงุนูŠุŒ ู„ูŠ ู‡ูŠ ุฌุงู…ุนุฉ ุจุญุซูŠุฉ ุฏูŠุงู„ ุงู„ุฏุฑุงุณุงุช ุงู„ุนู„ูŠุง ุงู„ู‡ุฏู ุฏูŠุงู„ู‡ุง ุฃู†ู‡ุง ุชุฒูŠุฏ ุจุงู„ุฐูƒุงุก ุงู„ุงุตุทู†ุงุนูŠ ู„ู‚ุฏู‘ุงู… ูˆุชู†ูุน ุจูŠู‡ ุงู„ุฅู†ุณุงู†ูŠุฉ. ูŠู…ูƒู† ู„ูŠูƒ ุชุฒูˆุฑ https://mbzuai.ac.ae/ar/about/ ุจุงุด ุชุนุฑู ูƒุซุฑ ุนู„ู‰ ุฌุงู…ุนุฉ ู…ุญู…ุฏ ุจู† ุฒุงูŠุฏ ู„ู„ุฐูƒุงุก ุงู„ุงุตุทู†ุงุนูŠ ูˆุงู„ู…ู‡ู…ุฉ ุฏูŠุงู„ู‡ุง!

Running the model on a single / multi GPU

pip install accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_id = "MBZUAI-Paris/Atlas-Chat-9B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
    torch_dtype=torch.bfloat16,
)

messages = [
    {"role": "user", "content": "ุดู†ูˆ ูƒูŠุชุณู…ู‰ ุงู„ู…ู†ุชุฎุจ ุงู„ู…ุบุฑุจูŠ ุŸ"},
]

input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt", return_dict=True, , add_generation_prompt=True)

outputs = model.generate(**input_ids, max_new_tokens=256)

print(tokenizer.decode(outputs[0]))
  • Response:

    ุงู„ู…ู†ุชุฎุจ ุงู„ู…ุบุฑุจูŠ ูƒูŠุชุณู…ู‰ ุฃูŠุถุง "ุฃุณูˆุฏ ุงู„ุฃุทู„ุณ"

Quantized Versions through bitsandbytes

Using 8-bit precision (int8)
pip install bitsandbytes accelerate
# pip install bitsandbytes accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig

model_id = "MBZUAI-Paris/Atlas-Chat-9B"
quantization_config = BitsAndBytesConfig(load_in_8bit=True)

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    quantization_config=quantization_config,
)
text = f"""
        ุดุฑุญ ู„ูŠุง ู‡ุงุฏ ุงู„ู‡ุถุฑุฉ:
        ููŠ ุงู„ู‚ุฑู† 19 ู„ู‚ุงูˆ ุงู„ุฐู‘ู‡ุจ ููŠ ูƒุงู„ูŠููˆุฑู†ูŠุงุŒ ู†ุงุถูˆ ู„ู‘ูŠ ูƒูŠุจูŠุนูˆ ุงู„ุนุชู„ุฉ ูˆุงู„ูุงุณ ูƒูŠู‚ู†ุนูˆ ุงู„ู†ุงุณ ุจู„ูŠ ุบูŠุฏูŠุฑูˆ ู„ุงุจุงุณ ูŠู„ุง ู‚ู„ุจูˆ ุนู„ู‰ ุงู„ุฐู‡ุจ... ูุงู„ุฃุฎูŠุฑ ุงุบุชู†ู‰ ุชุฌุงุฑ ุฃุฏูˆุงุช ุงู„ุชู†ู‚ูŠุจ ูˆุงู„ุญูุฑ. ูˆุญุงู„ูŠุง ูƒุงูŠู† ู„ู‘ูŠ ูƒูŠู‚ู†ุน ุงู„ุฃุฎุฑูŠู† ุจู„ูŠ ู‡ูˆ ู…ู„ูŠูˆู†ูŠุฑุŒ ูˆุนู†ุฏูˆ ุงู„ูˆู‚ุช ูŠูˆุฑู‘ูŠ ู„ู„ุขุฎุฑูŠู† ูƒูŠูุงุด ูŠุฏูŠุฑูˆ ู„ุงุจุงุณ.
        """
messages = [
    {"role": "user", "content": text},
]
input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt", return_dict=True).to("cuda")

outputs = model.generate(**input_ids, max_new_tokens=256)
print(tokenizer.decode(outputs[0]).split("<start_of_turn>model")[-1])
  • Response:

ู‡ุงุฏ ุงู„ู‡ุถุฑุฉ ูƒุชู‡ุถุฑ ุนู„ู‰ ู‚ุตุฉ ู‚ุฏูŠู…ุฉ ู…ู† ุงู„ู‚ุฑู† 19 ููŠู† ุชูƒุชุดู ุงู„ุฐู‡ุจ ููƒุงู„ูŠููˆุฑู†ูŠุง. ู‡ุงุฏ ุงู„ุดูŠ ุฎู„ู‚ ุญุงู„ุฉ ุฏูŠุงู„ ุงู„ุฌู†ูˆู† ุนู„ู‰ ุงู„ุฐู‡ุจุŒ ููŠู† ุจุฒุงู ุฏูŠุงู„ ุงู„ู†ุงุณ ู…ุดุงูˆ ู„ุชู…ุง ุจุงุด ูŠู‚ู„ุจูˆ ุนู„ูŠู‡. ูƒุงู†ูˆ ุญุชู‰ ู†ุงุณ ุงู„ู„ูŠ ูƒุงู†ูˆ ูƒูŠุจูŠุนูˆ ุฃุฏูˆุงุช ุงู„ุชู†ู‚ูŠุจ ุจุญุงู„ ุงู„ูุงุณ ูˆุงู„ุนุชู„ุฉุŒ ูˆูƒุงู†ูˆ ูƒูŠู‚ู†ุนูˆ ุงู„ู†ุงุณ ุจู„ูŠ ุบุงุฏูŠ ูŠุฑุจุญูˆ ุงู„ูู„ูˆุณ ุฅู„ุง ู…ุดุงูˆ ูŠู‚ู„ุจูˆ ุนู„ู‰ ุงู„ุฐู‡ุจ. ูุงู„ู†ู‡ุงูŠุฉุŒ ู‡ุงุฏูˆูƒ ุงู„ู„ูŠ ูƒุงู†ูˆ ูƒูŠุจูŠุนูˆ ู‡ุงุฏ ุงู„ุฃุฏูˆุงุช ู‡ูˆู…ุง ุงู„ู„ูŠ ุฑุจุญูˆ ุจุฒุงูุŒ ุญูŠุช ูƒุงู†ูˆ ูƒูŠุฑุจุญูˆ ู…ู† ูƒู„ ูˆุงุญุฏ ุงู„ู„ูŠ ูƒุงู† ูƒูŠุดุฑูŠ ู…ู†ู‡ู….

ู‡ุงุฏ ุงู„ู‚ุตุฉ ูƒุชุดุจู‡ ู„ู„ูŠ ูƒุงูŠู†ุฉ ุฏุงุจุงุŒ ููŠู† ูƒุงูŠู†ูŠู† ู†ุงุณ ุงู„ู„ูŠ ูƒูŠุฏุนูŠูˆ ุจู„ูŠ ู‡ูˆู…ุง ู…ู„ูŠูˆู†ูŠุฑ ูˆูƒูŠุจูŠุนูˆ ู†ุตุงุฆุญ ุนู„ู‰ ูƒูŠูุงุด ุชุฑุจุญ ุงู„ูู„ูˆุณ. ุจุญุงู„ ู‡ุงุฏูˆูƒ ุงู„ู„ูŠ ูƒุงู†ูˆ ูƒูŠุจูŠุนูˆ ุงู„ุฃุฏูˆุงุช ูุงู„ู…ุงุถูŠุŒ ุญุชู‰ ู‡ุงุฏ ุงู„ู†ุงุณ ูƒูŠุฑุจุญูˆ ู…ู† ู‡ุงุฏ ุงู„ุดูŠุŒ ุญูŠุช ูƒูŠุงุฎุฏูˆ ุงู„ูู„ูˆุณ ู…ู† ุงู„ู†ุงุณ ุงู„ู„ูŠ ูƒูŠุดุฑูŠูˆ ู…ู†ู‡ู… ุงู„ู†ุตุงุฆุญ ุฏูŠุงู„ู‡ู….

Using 4-bit precision
# pip install bitsandbytes accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig

model_id = "MBZUAI-Paris/Atlas-Chat-9B"
quantization_config = BitsAndBytesConfig(load_in_4bit=True)

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    quantization_config=quantization_config,
)
text = f"""ุชุฑุฌู… ู„ู„ุฏุงุฑุฌุฉ:
Atlas Chat is the first open source large language model that talks in Darija.
        """
messages = [
    {"role": "user", "content": text},
]
input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt", return_dict=True, add_generation_prompt=True)

outputs = model.generate(**input_ids, max_new_tokens=256, temperature=0.0)
print(tokenizer.decode(outputs[0]).split("<start_of_turn>model")[-1])
  • Response:

ุฃุทู„ู‘ุงุณ ุดุงุช ู‡ูˆ ุฃูˆู„ ู†ู…ูˆุฐุฌ ู„ุบูˆูŠ ูƒุจูŠุฑ ู…ูุชูˆุญ ุงู„ู…ุตุฏุฑ ูƒุงูŠู‡ุถุฑ ุจุงู„ุฏุงุฑุฌุฉ.

Chat Template

The models use a chat template that must be adhered to conversational use. The easiest way to apply it is using the tokenizer's built-in chat template, as shown in the following snippet.

Let's load the model and apply the chat template to a conversation. In this example, we'll start with a single user interaction:

from transformers import AutoTokenizer, AutoModelForCausalLM
import transformers
import torch

model_id = "MBZUAI-Paris/Atlas-Chat-9B"
dtype = torch.bfloat16

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="cuda",
    torch_dtype=dtype,)

chat = [
    { "role": "user", "content": "ุฃุดู†ูˆ ูƒุงูŠู…ูŠูŠุฒ ุงู„ู…ู…ู„ูƒุฉ ุงู„ู…ุบุฑุจูŠุฉ." },
]
prompt = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)

At this point, the prompt contains the following text:

<bos><start_of_turn>user
ุฃุดู†ูˆ ูƒุงูŠู…ูŠูŠุฒ ุงู„ู…ู…ู„ูƒุฉ ุงู„ู…ุบุฑุจูŠุฉ.<end_of_turn>
<start_of_turn>model

As you can see, each turn is preceded by a <start_of_turn> delimiter and then the role of the entity (either user, for content supplied by the user, or model for LLM responses). Turns finish with the <end_of_turn> token.

You can follow this format to build the prompt manually, if you need to do it without the tokenizer's chat template.

After the prompt is ready, generation can be performed like this:

inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt")
outputs = model.generate(input_ids=inputs.to(model.device), max_new_tokens=512)
print(tokenizer.decode(outputs[0]))
  • Response:

ุงู„ู…ุบุฑุจ ูƒุงูŠู…ูŠุฒูˆ ุจุฒุงู ุฏูŠุงู„ ุงู„ุญูˆุงูŠุฌุŒ ู…ู†ู‡ู…:

  1. ุงู„ุชู†ูˆุน ุงู„ุซู‚ุงููŠ: ุงู„ู…ุบุฑุจ ุจู„ุงุฏ ููŠู‡ุง ุจุฒุงู ุฏูŠุงู„ ุงู„ุซู‚ุงูุงุชุŒ ูƒู„ ูˆุญุฏุฉ ููŠู‡ู… ุนู†ุฏู‡ุง ุงู„ุชู‚ุงู„ูŠุฏ ุฏูŠุงู„ู‡ุง ูˆุงู„ู„ุบุฉ ุฏูŠุงู„ู‡ุง ูˆุงู„ู…ุงูƒู„ุฉ ุฏูŠุงู„ู‡ุง. ู‡ุงุฏ ุงู„ุชู†ูˆุน ูƒุงูŠุจุงู† ูุงู„ู…ูˆุณูŠู‚ู‰ ูˆุงู„ุฑู‚ุต ูˆุงู„ูู†ูˆู† ุงู„ุชู‚ู„ูŠุฏูŠุฉ.

  2. ุงู„ุชุฑุงุซ ุงู„ุชุงุฑูŠุฎูŠ: ุงู„ู…ุบุฑุจ ุนู†ุฏูˆ ุชุงุฑูŠุฎ ุบู†ูŠ ูƒุงูŠู…ุชุฏ ู„ุขู„ุงู ุงู„ุณู†ูŠู†ุŒ ููŠู‡ ุญุถุงุฑุงุช ู‚ุฏูŠู…ุฉ ุจุญุงู„ ู…ู…ู„ูƒุฉ ู…ูˆุฑูŠุทุงู†ูŠุงุŒ ูˆุงู„ุฑูˆู…ุงู†ุŒ ูˆุงู„ุจูŠุฒู†ุทูŠูŠู†ุŒ ูˆุงู„ูุชูˆุญุงุช ุงู„ุฅุณู„ุงู…ูŠุฉ. ู‡ุงุฏ ุงู„ุชุฑุงุซ ูƒุงูŠุจุงู† ูุงู„ู…ุนุงู„ู… ุงู„ุชุงุฑูŠุฎูŠุฉ ุจุญุงู„ ู…ุฏูŠู†ุฉ ูุงุณุŒ ูˆุงู„ู…ุฏูŠู†ุฉ ุงู„ู‚ุฏูŠู…ุฉ ุฏูŠุงู„ ู…ุฑุงูƒุดุŒ ูˆุงู„ู…ุฏูŠู†ุฉ ุงู„ู‚ุฏูŠู…ุฉ ุฏูŠุงู„ ุดูุดุงูˆู†.

  3. ุงู„ู…ู†ุงุธุฑ ุงู„ุทุจูŠุนูŠุฉ: ุงู„ู…ุบุฑุจ ุจู„ุงุฏ ููŠู‡ุง ู…ู†ุงุธุฑ ุทุจูŠุนูŠุฉ ู…ุชู†ูˆุนุฉุŒ ู…ู† ุงู„ุณูˆุงุญู„ ุงู„ุฒุฑู‚ุฉ ูˆุงู„ุตุญุงุฑูŠ ุงู„ูƒุจูŠุฑุฉุŒ ู„ู„ุฌุจุงู„ ุงู„ุนุงู„ูŠุฉ ูˆุงู„ูˆุฏูŠุงู† ุงู„ุฎุถุฑุงุก. ู‡ุงุฏ ุงู„ุชู†ูˆุน ูƒุงูŠู…ูƒู†ูƒ ู…ู† ู…ู…ุงุฑุณุฉ ุฃู†ุดุทุฉ ุฎุงุฑุฌูŠุฉ ุจุญุงู„ ุงู„ู…ุดูŠ ู„ู…ุณุงูุงุช ุทูˆูŠู„ุฉุŒ ูˆุงู„ุชุฎูŠูŠู…ุŒ ูˆุงู„ุฑูŠุงุถุงุช ุงู„ู…ุงุฆูŠุฉ.

  4. ุงู„ู…ุงูƒู„ุฉ: ุงู„ู…ุงูƒู„ุฉ ุงู„ู…ุบุฑุจูŠุฉ ู…ุนุฑูˆูุฉ ุจุงู„ุชู†ูˆุน ุฏูŠุงู„ู‡ุง ูˆุงู„ุทุนู… ุฏูŠุงู„ู‡ุง. ู…ู† ุจูŠู† ุงู„ุฃุทุจุงู‚ ุงู„ุฃูƒุซุฑ ุดุนุจูŠุฉ ูƒุงูŠู† ุงู„ุทุงุฌูŠู†ุŒ ูˆุงู„ูƒุณูƒุณุŒ ูˆุงู„ุจุฑูŠูˆุงุชุŒ ูˆุงู„ูƒูˆูƒุชูŠู„ ุฏูŠุงู„ ุงู„ููˆุงูƒู‡.

  5. ุงู„ู†ุงุณ: ุงู„ู…ุบุงุฑุจุฉ ู…ุนุฑูˆููŠู† ุจุงู„ุถูŠุงูุฉ ุฏูŠุงู„ู‡ู… ูˆุงู„ุชุฑุญุงุจ ุฏูŠุงู„ู‡ู…. ูƒุงูŠูƒูˆู†ูˆ ูุฑุญุงู†ูŠู† ุจุงุด ูŠุดุงุฑูƒูˆ ุงู„ุซู‚ุงูุฉ ูˆุงู„ุชู‚ุงู„ูŠุฏ ุฏูŠุงู„ู‡ู… ู…ุน ุงู„ุฒูˆุงุฑ.

Inputs and outputs

  • Input: Text string, such as a question, a prompt, or a document to be summarized.
  • Output: Generated Darija text in response to the input, such as an answer to a question, or a summary of a document.

Chatbot interface using Ollama

You can also use Ollama and chatbot-ollama to create a chatbot user-interface to better test the model. First you need to install Ollama on your machine from here and have node.js installed as well. Then, download and prepare the model as follows:


huggingface-cli download MBZUAI-Paris/Atlas-Chat-9B --local-dir Atlas-Chat-9B/
ollama create Atlas-Chat-9B -f Atlas-Chat-9B/modelfile
ollama serve

Finally, in a new terminal clone chatbot-ollama repository from Github and run it:

git clone https://github.com/ivanfioravanti/chatbot-ollama.git
cd chatbot-ollama
npm ci
npm run dev

You can start chatting with the model by visiting http://localhost:3000.

Citation

If you use Atlas-Chat in your research, please cite our paper:

@article{shang2024atlaschatadaptinglargelanguage,
      title={Atlas-Chat: Adapting Large Language Models for Low-Resource Moroccan Arabic Dialect}, 
      author={Guokan Shang and Hadi Abdine and Yousef Khoubrane and Amr Mohamed and Yassine Abbahaddou and Sofiane Ennadir and Imane Momayiz and Xuguang Ren and Eric Moulines and Preslav Nakov and Michalis Vazirgiannis and Eric Xing},
      year={2024},
      eprint={2409.17912},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2409.17912}, 
}

Training Data

The model was trained on diverse datasets focusing on Darija consisting for approximatley 450k instructions of a maximum length of 2048 tokens, including:

  • Synthetic instructions created to guide the model in processing various types of language tasks tailord towards Moroccan culture.
  • Instruction samples created from publicly available Moroccan Arabic datasets including translation, summarization and sentiment analysis.
  • Translated English and multi-lingual instruction-tuning datasets.

Our training dataset Darija-SFT-Mixture is publicly available.

Implementation Information

Atlas-Chat models are based on Gemma 2 models. The Atlas-Chat models were trained using 8 Nvidia's A100 80 GB GPUs in parallel using FSDP on AWS Sagemaker. The model is trained using HuggingFace transformers and parameter-efficient fine-tuning with LoRA rank of 256.

Evaluation

The Atlas-Chat models were evaluated on a comprehensive suite of tasks using various datasets and benchmarks to assess their performance across multiple dimensions. These included tasks such as:

  • DarijaMMLU: A Darija version of ArabicMMLU and MMLU benchmarks translated from MSA and English respectively.
  • DarijaHellaSwag: A Darija version of HellaSwag.
  • Belebele Ary_Arab: Belebele is a multiple-choice machine reading comprehension dataset published by Facebook spanning 122 language variants. The Evaluation is done on the Ary_Arab part of Belebele that refers to Darija.
  • DarijaAlpacaEval: A Darija version of AlpacaEval translated to Darija and adapted to the Moroccan culture.
  • Sentiment Analysis.
  • Translation: Including six directions and four languages: Darija, MSA, English and French.
  • Transliteration: Transforming a sentence from Darija (written in Arabic characters) to Arabizi (Written in Latin characters) and vice-versa.
  • Summarization.

The models were compared against a collection of existing open-source Arabic models to gauge their effectiveness, with a particular focus on performance in Darija. All scores are based on zero-shot performance. The prompts are written mainly in Darija. The metric used for DarijaMMLU, DarijaHellaSwag, Belebele Ary and Sentiment Analysis is the normalized accuracy. We used Language Model Evaluation Harness to conduct these evaluations.

LLMs Benchmarks:

Model DarijaMMLU DarijaHellaSwag Belebele Ary DarijaAlpacaEval
jais-family-1p3b-chat 35.39 27.71 38.33 35.56
jais-family-2p7b-chat 37.44 29.10 44.11 52.97
gemma-2-2b-it 28.58 32.42 25.22 58.67
Llama-3.2-1B-Instruct 27.66 26.88 28.89 23.57
Llama-3.2-3B-Instruct 32.60 28.33 38.00 47.62
Atlas-Chat-2B 44.97 35.08 53.89 92.31
jais-family-6p7b-chat 39.96 32.64 51.22 65.18
jais-adapted-7b-chat 39.30 29.55 43.67 61.84
jais-family-13b-chat 45.11 33.98 58.67 69.93
jais-adapted-13b-chat 45.20 32.84 49.67 77.52
AceGPT-7b-chat 35.98 30.33 30.11 47.31
AceGPT-13b-chat 41.09 38.35 33.11 52.79
gemma-2-9b-it 35.91 32.19 31.00 90.86
Llama-3.1-8B-Instruct 44.13 31.40 47.00 78.08
Atlas-Chat-9B 58.23 43.65 74.56 95.62
jais-family-30b-8k-chat 51.88 35.61 65.67 24.64
gemma-2-27b-it 36.47 37.04 35.78 95.07
Atlas-Chat-27B 61.95 48.37 75.67 96.58

Standard NLP Tasks:

Model DODa-10k (Translation) MADAR (Translation) FLORES+ (Translation) NLLB-Seed (Translation) DODa-10k (Transliteration) MArSum (Summarization)
(LLM as a judge)
Sentiment Analysis
BLEU chrF BLEU chrF BLEU chrF BLEU chrF BLEU chrF
jais-family-1p3b-chat 00.13 06.18 00.50 15.43 02.44 19.14 01.99 12.60 00.01 03.01 00.50 45.29
jais-family-2p7b-chat 00.25 07.46 00.62 16.36 04.25 18.22 03.10 08.19 00.01 03.27 00.90 51.56
gemma-2-2b-it 00.10 04.96 00.12 06.66 01.55 18.59 02.78 23.69 00.01 02.08 06.80 53.36
Llama-3.2-1B-Instruct 00.07 05.95 00.80 18.71 04.53 18.39 04.52 17.06 00.02 03.74 08.23 46.27
Llama-3.2-3B-Instruct 00.62 13.67 01.18 22.12 08.59 35.21 13.75 43.63 00.21 09.68 08.23 49.20
Atlas-Chat-2B 22.76 44.86 16.67 41.64 14.92 43.03 23.88 52.19 08.18 21.54 55.22 73.99
jais-family-6p7b-chat 00.73 11.85 01.88 23.22 04.25 18.22 04.62 20.22 00.02 03.79 03.02 56.78
jais-adapted-7b-chat 00.60 09.43 03.45 25.88 07.25 23.21 01.25 02.22 00.04 03.24 02.82 52.72
jais-family-13b-chat 00.92 11.71 04.01 28.48 05.70 27.24 04.50 22.56 00.03 03.57 01.77 41.73
jais-adapted-13b-chat 00.87 10.52 04.02 25.29 06.66 23.46 20.14 47.87 0.04 04.77 01.92 66.68
AceGPT-7b-chat 00.44 11.33 01.05 19.24 06.92 36.03 11.05 44.55 00.06 04.74 02.28 40.23
AceGPT-13b-chat 00.98 16.70 00.81 20.23 08.73 40.76 14.02 48.28 00.12 06.32 02.80 59.58
gemma-2-9b-it 03.10 19.16 01.72 24.35 05.18 36.96 08.23 43.57 00.17 09.14 13.81 59.87
Llama-3.1-8B-Instruct 00.92 14.19 01.46 23.82 08.89 33.08 11.85 35.51 00.11 06.02 16.14 44.08
Atlas-Chat-9B 28.08 50.48 18.16 43.91 18.63 47.53 29.98 58.26 22.08 34.17 59.76 81.89
jais-family-30b-8k-chat 01.10 14.40 01.67 23.37 08.52 35.41 13.71 41.33 00.05 04.48 00.46 56.73
gemma-2-27b-it 00.67 13.04 01.74 24.63 05.17 37.08 07.36 42.49 00.03 04.94 11.10 57.59
Atlas-Chat-27B 29.55 51.74 19.66 45.65 20.34 49.19 31.61 59.37 33.03 40.95 60.70 73.00

Usage and Limitations

These models have certain limitations that users should be aware of.

Intended Usage

Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development.

  • Content Creation and Communication

    • Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts.
    • Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications.
    • Text Summarization: Generate concise summaries of a text corpus, research papers, or reports.
  • Research and Education

    • Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field.
    • Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice.
    • Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics.
Limitations
  • Training Data

    • The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses.
    • The scope of the training dataset determines the subject areas the model can handle effectively.
  • Context and Task Complexity

    • LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging.
    • A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point).
  • Language Ambiguity and Nuance

    • Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language.
  • Factual Accuracy

    • LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements.
  • Common Sense

    • LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations.
  • Ethical Considerations and Risks

    The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following:

    • Bias and Fairness
      • LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material.
    • Misinformation and Misuse
      • LLMs can be misused to generate text that is false, misleading, or harmful.
      • Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit][rai-toolkit].
    • Transparency and Accountability:
      • This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes.
      • A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem.

    Risks identified and mitigations:

    • Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases.
    • Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases.
    • Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques.

    Acknowledgement

    We would like to express our gratitude to the following institutions for their contributions to this work: ร‰cole Polytechnique, LINAGORA and KTH Royal Institute of Technology. Additionally, we extend our thanks to the AtlasIA community.

    Downloads last month
    222
    Safetensors
    Model size
    9.24B params
    Tensor type
    BF16
    ยท
    Inference Examples
    This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

    Model tree for MBZUAI-Paris/Atlas-Chat-9B

    Base model

    google/gemma-2-9b
    Finetuned
    (82)
    this model
    Quantizations
    1 model

    Dataset used to train MBZUAI-Paris/Atlas-Chat-9B

    Spaces using MBZUAI-Paris/Atlas-Chat-9B 3