Recognizes inappropriate content in images and video containing concepts: gore, drug, explicit, suggestive, and safe.
Identifies a variety of concepts in images and video including objects, themes, and more. Trained with over 10,000 concepts and 20M images.
Detects logos and location in images for both catalogue/white background and in the wild. Recognizes concepts from popular automotive, beverages, and fashion brands.
Multilingual text classification model of concepts: toxic, insult, obscene, identity_hate, severe_toxic, and threat; in the selected language.
Detects a variety of common objects and the location and generates regions of an image that may contain that object.
Generates English captions from images. Ideal for auto-generating captions and creating metadata at scale.
Text translation model from a romance language to English using sentence piece-based segmentation
An OCR model for detecting and recognizing English text in images that are more complex than scans of a page.
A workflow for obtaining the sentiment of an audio.
A language-aware optical character recognition workflow