Aana SDK supports three types of model integration methods in a modular design:
1. Native support model
- speech processing: Whisper (transcription/translation)
- language model: vLLM (Efficient LLM Reasoning)
- visual model:: BLIP-2 (image description generation)
2. Transformers Ecology
pass (a bill or inspection etc) aana[transformers] Extension pack support:
- 20,000+ Pre-Trained Models on Hugging Face
- Custom model weight loading
- Efficient fine-tuning of PEFT parameters
3. Customized integration
Developers can inherit the BaseDeployment Class Implementation:
- ONNX/TensorRT Accelerated Models
- The Big Model of Privatized Deployment
- Traditional machine learning models (to be encapsulated as Python Callable)
All integration models enjoy unified resource management, API exposure and monitoring capabilities.
This answer comes from the articleAana SDK: An Open Source Tool for Easy Deployment of Multimodal AI ModelsThe
May not be reproduced without permission:AI productivity tools " What are the major AI model integrations supported by the Aana SDK?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil