TensorFlow.js models

Classify images with labels from the ImageNet database (MobileNet).

Localize and identify multiple objects in a single image (Coco SSD).

Run semantic segmentation in the browser (DeepLab).

Detect faces in images using a Single Shot Detector architecture with a custom encoder (Blazeface).

Predict 486 3D facial landmarks to infer the approximate surface geometry of human faces.

Unified pose detection API for using one of three models that help detect atypical poses and fast body motions with real time performance.

Segment person(s) and body parts in real-time.

Palm detector and hand-skeleton finger tracking model. Predict 21 3D hand keypoints per detected hand.

Estimate a depth map for a single portrait image of a human.

Answer questions based on the content of a given passage of text using BERT.

Score the perceived impact a comment may have on a conversation, from "Very toxic" to "Very healthy" (Toxicity).

Encode text into embeddings for NLP tasks such as sentiment classification and textual similarity (Universal Sentence Encoder).

Classify 1-second audio snippets from the speech commands dataset (speech-commands).

Utility to create a classifier using the K-Nearest-Neighbors algorithm. Can be used for transfer learning.