README updates

This commit is contained in:
Jake Poznanski 2025-06-17 19:58:06 +00:00
parent 69524cb305
commit 9d260791a0
2 changed files with 37 additions and 35 deletions

View File

@ -49,18 +49,30 @@ We also ship a comprehensive benchmark suite covering over 7,000 test cases acro
<thead>
<tr>
<th align="left"><strong>Model</strong></th>
<th align="center">AR</th>
<th align="center">OSM</th>
<th align="center">TA</th>
<th align="center">OS</th>
<th align="center">HF</th>
<th align="center">MC</th>
<th align="center">LTT</th>
<th align="center">ArXiv</th>
<th align="center">Old Scans Math</th>
<th align="center">Tables</th>
<th align="center">Old Scans</th>
<th align="center">Headers and Footers</th>
<th align="center">Multi column</th>
<th align="center">Long tiny text</th>
<th align="center">Base</th>
<th align="center">Overall Score</th>
<th align="center">Overall</th>
</tr>
</thead>
<tbody>
<tr>
<td align="left">Marker v1.7.5 (base)</td>
<td align="center">76.0</td>
<td align="center">57.9</td>
<td align="center">57.6</td>
<td align="center">27.8</td>
<td align="center">84.9</td>
<td align="center">72.9</td>
<td align="center">84.6</td>
<td align="center">99.1</td>
<td align="center">70.1 ± 1.1</td>
</tr>
<tr>
<td align="left">MinerU v1.3.10</td>
<td align="center">75.4</td>
@ -75,43 +87,32 @@ We also ship a comprehensive benchmark suite covering over 7,000 test cases acro
</tr>
<tr>
<td align="left">Mistral OCR API</td>
<td align="center">77.2</td>
<td align="center"><strong>77.2</strong></td>
<td align="center">67.5</td>
<td align="center">60.6</td>
<td align="center">29.3</td>
<td align="center">93.6</td>
<td align="center">71.3</td>
<td align="center">77.1</td>
<td align="center">99.4</td>
<td align="center"><strong>99.4</strong></td>
<td align="center">72.0 ± 1.1</td>
</tr>
<tr>
<td align="left">Marker v1.7.5 (base)</td>
<td align="center">76.0</td>
<td align="center">57.9</td>
<td align="center">57.6</td>
<td align="center">27.8</td>
<td align="center">84.9</td>
<td align="center">72.9</td>
<td align="center">84.6</td>
<td align="center">99.1</td>
<td align="center">70.1 ± 1.1</td>
</tr>
<tr>
<td align="left">olmOCR v0.1.68 (pipeline.py)</td>
<td align="center">75.6</td>
<td align="center">75.1</td>
<td align="center">70.2</td>
<td align="center"><strong>44.5</strong></td>
<td align="center">93.4</td>
<td align="center"><strong>79.4</strong></td>
<td align="center">81.7</td>
<td align="center">99.0</td>
<td align="center"><strong>77.4 ± 1.0</strong></td>
<td align="left">olmOCR v0.1.75 (Anchored)</td>
<td align="center">74.9</td>
<td align="center">71.2</td>
<td align="center">71.0</td>
<td align="center">42.2</td>
<td align="center">94.5</td>
<td align="center"><strong>78.3</strong></td>
<td align="center">73.3</td>
<td align="center">98.3</td>
<td align="center"><strong>75.5 ± 1.0</strong></td>
</tr>
</tbody>
</table>
### Installation
Requirements:

View File

@ -121,7 +121,7 @@ to run it against your own OCR tools. Your tool just needs to support Markdown o
<td align="left">Gemini Flash 2 (Anchored)</td>
<td align="center">54.5</td>
<td align="center">56.1</td>
<td align="center">72.1</td>
<td align="center"><strong>72.1</strong></td>
<td align="center">34.2</td>
<td align="center">64.7</td>
<td align="center">61.5</td>
@ -158,7 +158,7 @@ to run it against your own OCR tools. Your tool just needs to support Markdown o
<td align="center">71.5</td>
<td align="center">71.4</td>
<td align="center">71.4</td>
<td align="center">42.8</td>
<td align="center"><strong>42.8</strong></td>
<td align="center">94.1</td>
<td align="center">77.7</td>
<td align="center">71.0</td>
@ -172,7 +172,7 @@ to run it against your own OCR tools. Your tool just needs to support Markdown o
<td align="center">71.0</td>
<td align="center">42.2</td>
<td align="center">94.5</td>
<td align="center">78.3</td>
<td align="center"><strong>78.3</strong></td>
<td align="center">73.3</td>
<td align="center">98.3</td>
<td align="center"><strong>75.5 ± 1.0</strong></td>
@ -180,6 +180,7 @@ to run it against your own OCR tools. Your tool just needs to support Markdown o
</tbody>
</table>
<sup><sub>There was a small drop in scores from olmOCR v0.1.68 (77.4), which is due to two factors. One, is that we have adjusted our benchmark code to not include
any "fallback" mechanism when measuring benchmark scores (though it still exists when you run olmocr.pipeline). Second, there is a small drop in scores as we have updated
from sglang 0.4.2 to vllm 0.9.1. In net, we think the upgrade to vllm is the right choice, given that sglang 0.4.6 had even lower scores by one point, and vllm comes with a