Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2511.20478 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909924650385408 |
|---|---|
| author | Chumachenko, Kateryna Deshmukh, Amala Sanjay Seppanen, Jarno Karmanov, Ilia Chen, Chia-Chih Voegtle, Lukas Fischer, Philipp Wawrzos, Marek Motiian, Saeid Ageev, Roman Wu, Kedi Milesi, Alexandre Moosaei, Maryam Pawelec, Krzysztof Subramanian, Padmavathy Samadi, Mehrzad Yu, Xin Dear, Celina Stoddard, Sarah Diamond, Jenna Oliver, Jesse Chraghchian, Leanna Skelly, Patrick Balough, Tom Xu, Yao Scowcroft, Jane Polak Korzekwa, Daniel Hanley, Darragh Bhaskar, Sandip Roman, Timo Sapra, Karan Tao, Andrew Catanzaro, Bryan |
| author_facet | Chumachenko, Kateryna Deshmukh, Amala Sanjay Seppanen, Jarno Karmanov, Ilia Chen, Chia-Chih Voegtle, Lukas Fischer, Philipp Wawrzos, Marek Motiian, Saeid Ageev, Roman Wu, Kedi Milesi, Alexandre Moosaei, Maryam Pawelec, Krzysztof Subramanian, Padmavathy Samadi, Mehrzad Yu, Xin Dear, Celina Stoddard, Sarah Diamond, Jenna Oliver, Jesse Chraghchian, Leanna Skelly, Patrick Balough, Tom Xu, Yao Scowcroft, Jane Polak Korzekwa, Daniel Hanley, Darragh Bhaskar, Sandip Roman, Timo Sapra, Karan Tao, Andrew Catanzaro, Bryan |
| contents | We introduce Nemotron-Parse-1.1, a lightweight document parsing and OCR model that advances the capabilities of its predecessor, Nemoretriever-Parse-1.0. Nemotron-Parse-1.1 delivers improved capabilities across general OCR, markdown formatting, structured table parsing, and text extraction from pictures, charts, and diagrams. It also supports a longer output sequence length for visually dense documents. As with its predecessor, it extracts bounding boxes of text segments, as well as corresponding semantic classes. Nemotron-Parse-1.1 follows an encoder-decoder architecture with 885M parameters, including a compact 256M-parameter language decoder. It achieves competitive accuracy on public benchmarks making it a strong lightweight OCR solution. We release the model weights publicly on Huggingface, as well as an optimized NIM container, along with a subset of the training data as part of the broader Nemotron-VLM-v2 dataset. Additionally, we release Nemotron-Parse-1.1-TC which operates on a reduced vision token length, offering a 20% speed improvement with minimal quality degradation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_20478 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | NVIDIA Nemotron Parse 1.1 Chumachenko, Kateryna Deshmukh, Amala Sanjay Seppanen, Jarno Karmanov, Ilia Chen, Chia-Chih Voegtle, Lukas Fischer, Philipp Wawrzos, Marek Motiian, Saeid Ageev, Roman Wu, Kedi Milesi, Alexandre Moosaei, Maryam Pawelec, Krzysztof Subramanian, Padmavathy Samadi, Mehrzad Yu, Xin Dear, Celina Stoddard, Sarah Diamond, Jenna Oliver, Jesse Chraghchian, Leanna Skelly, Patrick Balough, Tom Xu, Yao Scowcroft, Jane Polak Korzekwa, Daniel Hanley, Darragh Bhaskar, Sandip Roman, Timo Sapra, Karan Tao, Andrew Catanzaro, Bryan Machine Learning We introduce Nemotron-Parse-1.1, a lightweight document parsing and OCR model that advances the capabilities of its predecessor, Nemoretriever-Parse-1.0. Nemotron-Parse-1.1 delivers improved capabilities across general OCR, markdown formatting, structured table parsing, and text extraction from pictures, charts, and diagrams. It also supports a longer output sequence length for visually dense documents. As with its predecessor, it extracts bounding boxes of text segments, as well as corresponding semantic classes. Nemotron-Parse-1.1 follows an encoder-decoder architecture with 885M parameters, including a compact 256M-parameter language decoder. It achieves competitive accuracy on public benchmarks making it a strong lightweight OCR solution. We release the model weights publicly on Huggingface, as well as an optimized NIM container, along with a subset of the training data as part of the broader Nemotron-VLM-v2 dataset. Additionally, we release Nemotron-Parse-1.1-TC which operates on a reduced vision token length, offering a 20% speed improvement with minimal quality degradation. |
| title | NVIDIA Nemotron Parse 1.1 |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2511.20478 |