BanglaWild: An In-the-Wild Bengali Scene Text Recognition Benchmark for OCR and Vision-Language Models 文章

ArXiv CS.CV2026-08-05PAPERen作者: Sadab Shiper, Tawsif Tashwar Dipto, Mir Md Inzamam, Eshat Tanzeem
大语言模型
查看原文 →
关系图谱

BanglaWild: An In-the-Wild Bengali Scene Text Recognition Benchmark for OCR and Vision-Language Models · 相关技术

相关技术

CuMAConvConfidence-Guided Early StoppingConditional average treatment effectsComputerized Adaptive TestingClinical Vision-Language ModelsCarbon-Aware Governance GatesCalibrated Entropy ScoreCOMCOGCANONBotBoABenchmarkBANGAutoregressive large language modelsAutomatic Depression DetectionAugmented Latent IntrinsicsAudio Mamba EncoderAttention-based multiple instance learningAttention ResidualsArtificial Language LearningAgentic Resource DiscoveryAdaptive Batch ScalingATEARGANSIANOANNAI-based Visually Impaired Assistance@Attribute3D Segment Anything ModelAbstractACT