Instructions to use alegchenko/command-r-08-2024-awq-ru-calib with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Local Apps Settings
- vLLM
How to use alegchenko/command-r-08-2024-awq-ru-calib with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "alegchenko/command-r-08-2024-awq-ru-calib" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "alegchenko/command-r-08-2024-awq-ru-calib", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/alegchenko/command-r-08-2024-awq-ru-calib
- SGLang
How to use alegchenko/command-r-08-2024-awq-ru-calib with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "alegchenko/command-r-08-2024-awq-ru-calib" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "alegchenko/command-r-08-2024-awq-ru-calib", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "alegchenko/command-r-08-2024-awq-ru-calib" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "alegchenko/command-r-08-2024-awq-ru-calib", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use alegchenko/command-r-08-2024-awq-ru-calib with Docker Model Runner:
docker model run hf.co/alegchenko/command-r-08-2024-awq-ru-calib
AWQ квантизация модели https://huggingface.co/CohereForAI/c4ai-command-r-08-2024 полученная с помощью https://github.com/casper-hansen/AutoAWQ Для калибровки использовались ограничения на 256 пакетов длиной до 256 токенов, собранные из решений различных задач на русском и английском языке с помощью GPT4 / GPT4o из датасетов: https://huggingface.co/datasets/IlyaGusev/saiga_scored https://huggingface.co/datasets/Open-Orca/OpenOrca
Валидация модели производилась на обучающей части бенчмарка MERA https://mera.a-ai.ru/ru/leaderboard, так для задачи PARus модель набирает 0.92 что эквивалетно например 4bit квантизациям Qwen2-72B и Llama3-70B
- Downloads last month
- 2
Model tree for alegchenko/command-r-08-2024-awq-ru-calib
Base model
CohereLabs/c4ai-command-r-08-2024