ImageNet classification w/ MobileNetV2
Segment hair, skin, clothes and accessories
150-class ADE20K scene segmentation w/ EfficientViT
Monocular depth estimation w/ Depth Anything V2
4x image super-resolution w/ Real-ESRGAN x4
Real-time object detection w/ YOLO26n
Real-time pose landmark detection w/ YOLO26n-pose
Image background removal w/ RMBG v1.4
Real-time background removal w/ MODNet
Real-time object detection w/ YOLO12n
Real-time depth estimation w/ Depth Anything V2
Text Generation
Classification: MobileNet, ResNet, EfficientNet Lite
Text-to-Image: Text Encoder, UNet, VAE Decoder, Safety Checker
Image Segmentation: SAM
Automatic Speech Recognition: Whisper Encoder, Decoder
Text-to-Image: Synthesize photorealistic images from a text prompt in a single network evaluation
Music-Source-Separation: Split a stereo music track into drums, bass, other, and vocals stems
A browser-based playground for experimenting with WebNN
Evaluate, review and modify WebNN sample codes
Classification: MobileNet, SqueezeNet, ResNet, EfficientNet Lite
Use LeNet to classify handwritten digits from the MNIST data set
Applying the STARRY NIGHT and other styles of Van Gogh to image
Detecting instances of semantic objects in digital images and videos
A NSNet2 baseline impl. of deep learning-based noise suppression
An RNNoise baseline impl. of deep learning-based noise suppression
Partitioning image into semantically meaningful parts
Detecting facial landmarks like eyes, nose, mouth, etc.
Detecting faces of participants using object detection
