# Attribution and modification notice

Original model: [Datalab Chandra OCR 2](https://huggingface.co/datalab-to/chandra-ocr-2), revision
`af93b47dba1b47b6640c86ccf487ed2260ab9a09`.

Original model authorship, architecture, training, trademarks, and related rights remain with Datalab and the applicable
contributors. This repository is an unofficial conversion and is not endorsed by Datalab.

Changes made for this distribution:

- converted the source checkpoint to GGUF with `ggml-org/llama.cpp` revision
  `8f5ab832ca7d8a7b4f23687693fb8b0ecbc227e7`;
- retained BF16 tensor precision;
- omitted the unused MTP draft head from the main GGUF;
- exported the vision projector as a separate BF16 GGUF;
- added conversion, integrity, launch, and AMD RDNA4 validation documentation.

The original modified OpenRAIL license is reproduced unmodified in [`LICENSE`](LICENSE) and applies to these derivative
files. No additional permission to use Datalab trademarks is granted.