# DharmaOCR: Specialized Small Language Models for Structured OCR that outperform Open-Source and Commercial Baselines

arXiv: https://arxiv.org/abs/2604.14314 · HTML: https://mlsanity.com/paper/2604.14314/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Diffusion and multimodal model alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 26 of its 58 references that are in the map.
- First posted on arXiv 15 Apr 2026.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Diffusion and multimodal model alignment](https://mlsanity.com/alignment/diffusion-and-multimodal-model-alignment/)

## Builds on (most cited first, 26 of 26)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [SGDR: Stochastic Gradient Descent with Warm Restarts](https://mlsanity.com/paper/1608.03983/) (2016)
- [The Curious Case of Neural Text Degeneration](https://mlsanity.com/paper/1904.09751/) (2019)
- [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://mlsanity.com/paper/1912.13318/) (2019)
- [An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition](https://mlsanity.com/paper/1507.05717/) (2015)
- [Neural Text Generation with Unlikelihood Training](https://mlsanity.com/paper/1908.04319/) (2019)
- [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://mlsanity.com/paper/2012.14740/) (2020)
- [Break the Sequential Dependency of LLM Inference Using Lookahead Decoding](https://mlsanity.com/paper/2402.02057/) (2024)
- [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://mlsanity.com/paper/2004.09602/) (2020)
- [olmOCR: Unlocking Trillions of Tokens in PDFs with Vision Language Models](https://mlsanity.com/paper/2502.18443/) (2025)
- [A Theoretical Analysis of the Repetition Problem in Text Generation](https://mlsanity.com/paper/2012.14660/) (2020)
- [olmOCR 2: Unit Test Rewards for Document OCR](https://mlsanity.com/paper/2510.19817/) (2025)
- [MBQ: Modality-Balanced Quantization for Large Vision-Language Models](https://mlsanity.com/paper/2412.19509/) (2024)
- [Repetition In Repetition Out: Towards Understanding Neural Text Degeneration from the Data Perspective](https://mlsanity.com/paper/2310.10226/) (2023)
- [Principled Data Selection for Alignment: The Hidden Risks of Difficult Examples](https://mlsanity.com/paper/2502.09650/) (2025)
- [Understanding the Repeat Curse in Large Language Models from a Feature Perspective](https://mlsanity.com/paper/2504.14218/) (2025)
- [Queue management for slo-oriented large language model serving](https://mlsanity.com/paper/2407.00047/) (2024)
- [Relating Neural Text Degeneration to Exposure Bias](https://mlsanity.com/paper/2109.08705/) (2021)
- [Comparing Specialised Small and General Large Language Models on Text Classification: 100 Labelled Samples to Achieve Break-Even Performance](https://mlsanity.com/paper/2402.12819/) (2024)
- [Mitigating the Language Mismatch and Repetition Issues in LLM-based Machine Translation via Model Editing](https://mlsanity.com/paper/2410.07054/) (2024)
- [Multimodal LLMs for OCR, OCR Post-Correction, and Named Entity Recognition in Historical Documents](https://mlsanity.com/paper/2504.00414/) (2025)
- [Beyond Reward Margin: Rethinking and Resolving Likelihood Displacement in Diffusion Models via Video Generation](https://mlsanity.com/paper/2511.19049/) (2025)
- [What LLMs Think When You Don't Tell Them What to Think About?](https://mlsanity.com/paper/2602.01689/) (2026)

## Abstract (arXiv)

This manuscript introduces DharmaOCR Full and Lite, a pair of specialized small language models (SSLMs) for structured OCR that jointly optimize transcription quality, generation stability, and inference cost. It also presents DharmaOCR-Benchmark, a benchmark that covers printed, handwritten, and legal/administrative documents, and proposes a unified evaluation protocol that measures fidelity and structure while explicitly tracking text degeneration as a first-class benchmark metric (alongside unit cost). Beyond reporting degeneration rates, the manuscript empirically shows degeneration is not merely a quality failure, since it materially worsens production performance by increasing response time, reducing throughput, and inflating computational cost due to abnormally long generations. To the best of the author's knowledge, as a methodological contribution, this is the first application of Direct Preference Optimization (DPO) for OCR, explicitly using degenerate generations as rejected examples to penalize looping behavior. Combined with Supervised Fine-Tuning (SFT) for enforcing a strict JSON schema (header, margin, footer, and text), DPO consistently reduces degeneration rate across model families (up to 87.6% relative) while preserving or improving extraction quality. The resulting models, namely, DharmaOCR Full (7B) and DharmaOCR Lite (3B), set a new state-of-the-art on DharmaOCR-Benchmark, outperforming each open-source and commercial baseline model evaluated regarding extraction quality, reaching 0.925 and 0.911 scores with 0.40% and 0.20% degeneration rates. AWQ quantization reduced up to 22% per-page cost with negligible quality loss, enabling a strong quality-cost trade-off in comparison to proprietary OCR APIs and open-source alternatives.