cerebras.modelzoo.data.multimodal.datasets.features

Classes

Bbox	Source: indicates how the box was made:
ImageLabels	ImageLabels(ClassLabel: str, ClassIntID: str, ClassID: str = None, Confidence: float = None, IsTrainable: bool = None, Source: str = None)
ObjectDetectionFeaturesDict	ImageID: Name of image
VQAAnswer	VQAAnswer(answer_id: int, answer: str, answer_confidence: str, answer_language: str)
VQAFeaturesDict	ImageID: Name of image
VQAQuestion	VQAQuestion(question_id: int, question: str, question_language: str)

previous

cerebras.modelzoo.data.multimodal.datasets.VQAv2

next

cerebras.modelzoo.data.multimodal.datasets.features.Bbox