The PAI Python SDK provides a rich collection of example Notebooks to help you quickly learn how to develop and deploy models on PAI.
Model development
Notebook source | Example description |
Develop a model using the MNIST dataset and the PAI PyTorch runtime image. | |
Develop and deploy models using the models and algorithms from PAI QuickStart. | |
Visualize the training procedure using the PAI TensorBoard service. | |
Submit a PyTorch distributed training job using the PAI Python SDK. | |
Save and resume training jobs using checkpoints saved during the training procedure. |
Model deployment
Notebook source | Example description |
Deploy a model inference service using a custom image. | |
Deploy and invoke an asynchronous inference service using the PAI Python SDK. |
Using ModelScope
Notebook source | Example description |
Deploy a ModelScope model using the PAI Python SDK. | |
Fine-tune and deploy a large language model (LLM) using Swift. | |
Fine-tune and deploy a ViT image classification model using the library. |
Using HuggingFace
Notebook source | Example description |
Deploy a HuggingFace model using the PAI Python SDK. | |
Fine-tune and deploy a Bidirectional Encoder Representations from Transformers (BERT) model using Transformer. |