Hyperbots Team Presents SAVIOR OCR Model at Best of WACV 2026

Hyperbots Inc., a leading Agentic AI platform for Finance & Accounting, today announced that its AI research team presented SAVIOR: Sample-efficient Adaptation of Vision-Language Models for OCR Representation at Best of WACV 2026, a virtual meetup hosted by Voxel51 on May 1, 2026. Akshata Bhat, AI/ML Research Engineer at Hyperbots, presented the work to the global computer vision community.

Best of WACV is a Voxel51 series that brings the authors of standout WACV papers together to present their research live. WACV, the IEEE/CVF Winter Conference on Applications of Computer Vision, is one of the world’s leading computer vision conferences. SAVIOR was first presented at the VisionDocs workshop at WACV 2026 in Tucson, Arizona, and the Best of WACV session extended its reach to a wider audience of AI, machine learning, and document AI practitioners. The session also featured talks from New York University and Ecliptor.

OCR continues to be a major bottleneck in finance operations. While modern Vision-Language Models perform well on standard benchmarks, they often struggle with the documents common in enterprise workflows, including invoices with vertical text, logo-embedded vendor names, degraded scans, and complex multi-column layouts. These patterns are rarely represented in public datasets, yet they account for a large share of real-world OCR failures.

SAVIOR approaches the problem differently. Instead of relying on massive amounts of generic training data, it focuses on curating the documents that represent the exact visual structures where pretrained models fail. The resulting dataset, SAVIOR-Train, contains just 2,234 documents selected for these challenging patterns.

Using SAVIOR-Train, Hyperbots fine-tuned Qwen2.5-VL-Instruct and achieved a word-level recall of 0.93. When the OCR output was used for downstream document question answering, the model achieved an F1 score of 0.89, outperforming GPT-4o, Mistral-OCR, PaddleOCR-VL, and DeepSeek-OCR despite using a fraction of the training data.

The research also introduces SAVIOR-Bench, a 509-document benchmark built from real-world finance and accounting workflows, and PaIRS (Pairwise Relational Similarity), a structure-aware evaluation metric that measures whether OCR systems preserve spatial relationships between words and fields, something traditional text-based metrics often miss.

“Most OCR benchmarks don’t reflect what actually breaks in enterprise finance workflows,” said Niyati Chhaya, Co-Founder & VP AI at Hyperbots Inc. “SAVIOR shows that carefully curated data targeting real production failure modes can outperform much larger systems trained on generic datasets. This is how we approach AI at Hyperbots: focused on production accuracy, operational reliability, and practical deployment.”

SAVIOR-trained models also maintain strong performance under 4-bit quantization, making them suitable for memory-constrained and latency-sensitive enterprise environments. The quantized 3B model achieved throughput exceeding 1,085 tokens per second on an A100 GPU, enabling scalable deployment in high-volume document pipelines.

By presenting SAVIOR at Best of WACV 2026, Hyperbots continues its investment in production-focused AI research for enterprise finance. The company’s AI team holds more than 50 patents across artificial intelligence, document understanding, and enterprise software.

Full Research Paper: https://openaccess.thecvf.com/content/WACV2026W/VisionDocs/papers/Bhat_SAVIOR_Sample-efficient_Adaptation_of_Vision-Language_Models_for_OCR_Representation_WACVW_2026_paper.pdf

About Hyperbots Inc.
Hyperbots Inc. transforms Finance & Accounting operations using proprietary AI agents that autonomously read, reconcile, validate, and post financial data across complex workflows. Its specialized AI Co-Pilots operate on a finance-specific foundational architecture delivering intelligence across language, vision, reasoning, interpretation, recommendation, prediction, redaction, and exception handling. Hyperbots integrates with major ERPs and maintains ISO 27001, SOC 1 Type 2, and SOC 2 Type 2 certifications.

Contact: support@hyperbots.com
Website: https://www.hyperbots.com
Request a Demo: https://www.hyperbots.com/request-demo