API Reference

PaddlePaddle (PArallel Distributed Deep LEarning) is an efficient, flexible, and extensible deep learning framework, commits to making the innovation and application of deep learning technology easier.

In this version, PaddlePaddle has made many optimizations to the APIs. You can refer to the following table to understand the API directory structure and description of the latest version of PaddlePaddle.In addition, you can refer to PaddlePaddle’s GitHub for details, or read Release Notes to learn about the features of the new version.

The API directory structure of PaddlePaddle is as follows:

Directory

Functions and Included APIs

paddle.*

The aliases of commonly used APIs are reserved in the paddle root directory, which currently include all the APIs in the paddle.tensor, paddle.framework and paddle.device directories.

paddle.tensor

APIs related to tensor operations such as creating zeros, matrix operation matmul, transforming concat, computing add, and finding argmax.

paddle.framework

PaddlePaddle universal APIs and dynamic graph APIs such as no_grad, save and load.

paddle.amp

Paddle automatic mixed precision strategy, including auto_cast, GradScaler, etc.

paddle.audio

Audio API, including feature extraction and dataset operations.

paddle.autograd

Auto grad APIs, including backward, PyLayer, etc.

paddle.callbacks

Paddle log callback APIs, including ModelCheckpoint, ProgBarLogger, etc.

paddle.device

Device management related APIs, such as set_device, get_device, etc.

paddle.distributed

Distributed related basic APIs.

paddle.distributed.fleet

Distributed related high-level APIs.

paddle.distribution

Probability distribution class APIs, including various common probability distribution.

paddle.fft

Fast Fourier Transform related APIs, such as fft, fft2, etc.

paddle.geometric

APIs for Graph Learning, including message passing, graph sampling, etc.

paddle.hub

Model extension API, including list, load, help, etc.

paddle.incubate

APIs for incubating new features.

paddle.io

APIs related to data input and output such as Dataset, and DataLoader.

paddle.inference

APIs related to inference for predictions.

paddle.jit

Dynamic graph to static graph APIs, including to_static, not_to_static, save, load, etc.

paddle.linalg

Linear algebra related APIs, such as det, svd, etc.

paddle.metric

APIs related to evaluation computation such as Accuracy and Auc.

paddle.nn

Networking-related APIs such as Linear, Conv2D, CrossEntropyLoss, RNN,and ReLU, etc.

paddle.onnx

APIs related to convert paddle model to ONNX,such as export

paddld.optimizer

APIs related to optimization algorithms such as SGD, Adagrad, and Adam.

paddle.optimizer.lr

APIs related to learning rate decay, such as NoamDecay, StepDecay, PiecewiseDecay, etc.

paddle.profiler

Performance profiler APIs for the Paddle framework, providing functionality to display and analyze performance for model training and inference processes.

paddle.quantization

Quantization related APIs.

paddle.regularizer

Regularization APIs, including L1Decay, L2Decay, etc.

paddle.signal

APIs for signal processing.

paddle.sparse

The Sparse domain API.

paddle.static

Basic framework related APIs under static graph, such as Variable, Program, Executor, etc.

paddle.static.nn

Special APIs for networking under static graph such as full connect layer fc and control flow while_loop/cond.

paddle.sysconfig

System configuration APIs, such as get_include, get_lib.

paddle.text

The NLP domain API currently includes data sets related to the NLP domain, such as Imdb and Movielens.

paddle.utils

Utils APIs, including CppExtension, CUDAExtension.

paddle.vision

Vision domain APIs such as datasets Cifar10, data processing ColorJitter, and commonly used models like resnet.