|
--- |
|
inference: false |
|
library_name: transformers |
|
language: |
|
- en |
|
- fr |
|
- de |
|
- es |
|
- it |
|
- pt |
|
- ja |
|
- ko |
|
- zh |
|
- ar |
|
- el |
|
- fa |
|
- pl |
|
- id |
|
- cs |
|
- he |
|
- hi |
|
- nl |
|
- ro |
|
- ru |
|
- tr |
|
- uk |
|
- vi |
|
license: cc-by-nc-4.0 |
|
--- |
|
|
|
# Making AI Models Faster, Smaller, and Cleaner! 🥝 |
|
|
|
**Contact me** if you have suggestions for models to compress next or if you want to compress your own models! |
|
|
|
**Reduction:** |
|
We reduced the Aya's size from 11 GB to 7GB, as well as reducing inference cost and gpu usage. |
|
|
|
**How does the compression work?** |
|
The model is compressed using Quanto to 8 bits. |
|
|
|
**How does the model quality change?** |
|
Quality might vary compared to the base model. |
|
|
|
**How is the model efficiency evaluated?** |
|
Results are based on testing with specific hardware and configurations. Efficiency may vary with different settings. |
|
|
|
**What is the model format?** |
|
We use safetensors. |
|
|
|
**What calibration data was used?** |
|
Calibration data includes WikiText, if needed by the compression method. |
|
|
|
**How to compress your own models?** |
|
Request access for more compression methods and support: [Request Access](#) |
|
|
|
### About |
|
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](https://huggingface.co/CohereForAI/c4ai-command-r-plus) of models with the recently released [Aya Collection](https://huggingface.co/datasets/CohereForAI/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](https://huggingface.co/CohereForAI/aya-23-35B). |
|
|
|
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 For AI](https://cohere.for.ai) and [Cohere](https://cohere.com/) |
|
|
|
- Point of Contact: Cohere For AI: [cohere.for.ai](https://cohere.for.ai/) |
|
- License: [CC-BY-NC](https://cohere.com/c4ai-cc-by-nc-license), requires also adhering to [C4AI's Acceptable Use Policy](https://docs.cohere.com/docs/c4ai-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 |
|
|
|
```python |
|
# pip install transformers==4.41.1 |
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
model_id = "CohereForAI/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) |
|
``` |
|
|
|
|
|
## Configurations |
|
|
|
Configuration details are in `config.json`. |
|
|
|
## Credits & License |
|
|
|
The license of the model follows the original model’s license. Check the license of the base model for details. |
|
|
|
**Want to suggest or compress other models?** |
|
Send me a message! |
