aya-23-8B/README.md

inference, library_name, language, license, extra_gated_prompt, extra_gated_fields

inference	library_name	language	license	extra_gated_prompt	extra_gated_fields
false	transformers	en fr de es it pt ja ko zh ar el fa pl id cs he hi nl ro ru tr uk vi	cc-by-nc-4.0	By submitting this form, you agree to the [License Agreement](https://cohere.com/c4ai-cc-by-nc-license) and acknowledge that the information you provide will be collected, used, and shared in accordance with Cohere’s [Privacy Policy]( https://cohere.com/privacy). You’ll receive email updates about Cohere Labs and Cohere research, events, products and services. You can unsubscribe at any time.	Name Affiliation Country I agree to use this model for non-commercial use ONLY text text type options select Aruba Afghanistan Angola Anguilla Åland Islands Albania Andorra United Arab Emirates Argentina Armenia American Samoa Antarctica French Southern Territories Antigua and Barbuda Australia Austria Azerbaijan Burundi Belgium Benin Bonaire Sint Eustatius and Saba Burkina Faso Bangladesh Bulgaria Bahrain Bahamas Bosnia and Herzegovina Saint Barthélemy Belarus Belize Bermuda Plurinational State of Bolivia Brazil Barbados Brunei-Darussalam Bhutan Bouvet-Island Botswana Central African Republic Canada Cocos (Keeling) Islands Switzerland Chile China Côte-dIvoire Cameroon Democratic Republic of the Congo Cook Islands Colombia Comoros Cabo Verde Costa Rica Cuba Curaçao Christmas Island Cayman Islands Cyprus Czechia Germany Djibouti Dominica Denmark Dominican Republic Algeria Ecuador Egypt Eritrea Western Sahara Spain Estonia Ethiopia Finland Fiji Falkland Islands (Malvinas) France Faroe Islands Federated States of Micronesia Gabon United Kingdom Georgia Guernsey Ghana Gibraltar Guinea Guadeloupe Gambia Guinea Bissau Equatorial Guinea Greece Grenada Greenland Guatemala French Guiana Guam Guyana Hong Kong Heard Island and McDonald Islands Honduras Croatia Haiti Hungary Indonesia Isle of Man India British Indian Ocean Territory Ireland Islamic Republic of Iran Iraq Iceland Israel Italy Jamaica Jersey Jordan Japan Kazakhstan Kenya Kyrgyzstan Cambodia Kiribati Saint-Kitts-and-Nevis South Korea Kuwait Lao-Peoples-Democratic-Republic Lebanon Liberia Libya Saint-Lucia Liechtenstein Sri Lanka Lesotho Lithuania Luxembourg Latvia Macao Saint Martin (French-part) Morocco Monaco Republic of Moldova Madagascar Maldives Mexico Marshall Islands North Macedonia Mali Malta Myanmar Montenegro Mongolia Northern Mariana Islands Mozambique Mauritania Montserrat Martinique Mauritius Malawi Malaysia Mayotte Namibia New Caledonia Niger Norfolk Island Nigeria Nicaragua Niue Netherlands Norway Nepal Nauru New Zealand Oman Pakistan Panama Pitcairn Peru Philippines Palau Papua New Guinea Poland Puerto Rico North Korea Portugal Paraguay State of Palestine French Polynesia Qatar Réunion Romania Russia Rwanda Saudi Arabia Sudan Senegal Singapore South Georgia and the South Sandwich Islands Saint Helena Ascension and Tristan da Cunha Svalbard and Jan Mayen Solomon Islands Sierra Leone El Salvador San Marino Somalia Saint Pierre and Miquelon Serbia South Sudan Sao Tome and Principe Suriname Slovakia Slovenia Sweden Eswatini Sint Maarten (Dutch-part) Seychelles Syrian Arab Republic Turks and Caicos Islands Chad Togo Thailand Tajikistan Tokelau Turkmenistan Timor Leste Tonga Trinidad and Tobago Tunisia Turkey Tuvalu Taiwan United Republic of Tanzania Uganda Ukraine United States Minor Outlying Islands Uruguay United-States Uzbekistan Holy See (Vatican City State) Saint Vincent and the Grenadines Bolivarian Republic of Venezuela Virgin Islands British Virgin Islands U.S. VietNam Vanuatu Wallis and Futuna Samoa Yemen South Africa Zambia Zimbabwe checkbox

Model Card for Aya-23-8B

Note: This is an older version of Aya. The latest version is Aya Expanse 8B which is available here. We also have multimodal variant, Aya Vision 8B which is available here.

Try Aya Expanse and Aya Vision:

You can try out latest Aya models before downloading the weights in our hosted Hugging Face Space here.

Model Summary

Aya 23 is an open weights research release of an instruction fine-tuned model with highly advanced multilingual capabilities. Aya 23 focuses on pairing a highly performant pre-trained Command family of models with the recently released Aya Collection. The result is a powerful multilingual large language model serving 23 languages.

This model card corresponds to the 8-billion version of the Aya 23 model. We also released a 35-billion version which you can find here.

We cover 23 languages: Arabic, Chinese (simplified & traditional), Czech, Dutch, English, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Persian, Polish, Portuguese, Romanian, Russian, Spanish, Turkish, Ukrainian, and Vietnamese

Developed by: Cohere Labs and Cohere

• Point of Contact: Cohere Labs: cohere.for.ai
• License: CC-BY-NC, requires also adhering to Cohere Lab's Acceptable Use Policy
• Model: aya-23-8B
• Model Size: 8 billion parameters

Usage

Please install transformers from the source repository that includes the necessary changes for this model

# pip install transformers==4.41.1
from transformers import AutoTokenizer, AutoModelForCausalLM

model_id = "CohereLabs/aya-23-8B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

# Format message with the command-r-plus chat template
messages = [{"role": "user", "content": "Anneme onu ne kadar sevdiğimi anlatan bir mektup yaz"}]
input_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
## <BOS_TOKEN><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Anneme onu ne kadar sevdiğimi anlatan bir mektup yaz<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>

gen_tokens = model.generate(
    input_ids, 
    max_new_tokens=100, 
    do_sample=True, 
    temperature=0.3,
    )

gen_text = tokenizer.decode(gen_tokens[0])
print(gen_text)

Example Notebook

This notebook showcases a detailed use of Aya 23 (8B) including inference and fine-tuning with QLoRA.

Model Details

Input: Models input text only.

Output: Models generate text only.

Model Architecture: Aya-23-8B is an auto-regressive language model that uses an optimized transformer architecture. After pretraining, this model is fine-tuned (IFT) to follow human instructions.

Languages covered: The model is particularly optimized for multilinguality and supports the following languages: Arabic, Chinese (simplified & traditional), Czech, Dutch, English, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Persian, Polish, Portuguese, Romanian, Russian, Spanish, Turkish, Ukrainian, and Vietnamese

Context length: 8192

Evaluation

Please refer to the Aya 23 technical report for further details about the base model, data, instruction tuning, and evaluation.

Model Card Contact

For errors or additional questions about details in this model card, contact labs@cohere.com

Terms of Use

We hope that the release of this model will make community-based research efforts more accessible, by releasing the weights of a highly performant multilingual model to researchers all over the world. This model is governed by a CC-BY-NC License with an acceptable use addendum, and also requires adhering to Cohere Lab's Acceptable Use Policy.

Try the model today

You can try Aya 23 in the Cohere playground here. You can also use it in our dedicated Hugging Face Space here.

Citation info

@misc{aryabumi2024aya,
      title={Aya 23: Open Weight Releases to Further Multilingual Progress}, 
      author={Viraat Aryabumi and John Dang and Dwarak Talupuru and Saurabh Dash and David Cairuz and Hangyu Lin and Bharat Venkitesh and Madeline Smith and Kelly Marchisio and Sebastian Ruder and Acyr Locatelli and Julia Kreutzer and Nick Frosst and Phil Blunsom and Marzieh Fadaee and Ahmet Üstün and Sara Hooker},
      year={2024},
      eprint={2405.15032},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}