Instructions to use amazon/MistralLite with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use amazon/MistralLite with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="amazon/MistralLite")# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("amazon/MistralLite") model = AutoModelForCausalLM.from_pretrained("amazon/MistralLite") - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use amazon/MistralLite with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "amazon/MistralLite" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "amazon/MistralLite", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/amazon/MistralLite
- SGLang
How to use amazon/MistralLite 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 "amazon/MistralLite" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "amazon/MistralLite", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'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 "amazon/MistralLite" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "amazon/MistralLite", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use amazon/MistralLite with Docker Model Runner:
docker model run hf.co/amazon/MistralLite
Prompt template for multi-turn conversations?
#15
by apepkuss79 - opened
According to the description in Model card, the prompt template is
<|prompter|>{user_message}</s><|assistant|>
How about the prompt template for multi-turn conversations? Just concatenate the prompt templates one by one? looks like the example below?
<|prompter|>{user_message_1}</s><|assistant|>{assistant_message_1}<|prompter|>{user_message_2}</s><|assistant|>
Yes @apepkuss79 , you can try that, but the fine tuned data didn't have such multiple run conversation. So there is no guarantee it would be able to pick the message history.
You can try to use this https://python.langchain.com/docs/modules/memory/ to manage the conversation history here. Thank you!