README.md

Pre-trained models based on TensorLayerX. TensorLayerX is a multi-backend AI framework, which can run on almost all operation systems and AI hardwares, and support hybrid-framework programming. The currently version supports TensorFlow, MindSpore, PaddlePaddle, PyTorch, OneFlow and Jittor as the backends.

Quick Start

Installation

pip

# install from pypi
pip3 install tlxzoo

docker

docker pull xiaolong0612/tlxzoo:0.0.1
# run docker
cd /workspace/tlxzoo
python demo/vision/image_classification/vgg/predict.py

train

python demo/vision/image_classification/vgg/train.py

pretrained model

https://pan.baidu.com/s/1FZiHSthgX2FqynBP9cJ6_A?pwd=8dbs

predict

python demo/vision/image_classification/vgg/predict.py

Scenes and Models

TLXZOO currently covers 4 fields and 12 tasks. They ares listed below:

Vision

Object Detection

• DETR
• PPYOLOE

Image Classification

• ResNet
• VGG
• EfficientNet

Image Segmentation

• UNet

Face Recognition

• RetinaFace
• ArcFace

Human Pose Estimation

• HRNet

Facial Landmark Detection

• PFLD

OCR

• TrOCR

Text

Text Classification

• BERT
• T5

Text Conditional Generation

• T5

Text Token Classidication

• BERT
• T5

Speech

Automatic Speech Recognition

• wav2vec

Graph

Node Classification

• GCN