Update README.md

README.md

@@ -6,15 +6,9 @@ tags:
 configs:
 - config_name: karantaocr-bench
   data_files:
-  - split: dictionary
+  - split: diacritics
     path:
-    - dictionary.jsonl
-  - split: religion
-    path:
-    - religion.jsonl
-  - split: no_type
-    path:
-    - no_type.jsonl
+    - diacritics.jsonl
 language:
 - yo
 - ig
@@ -35,50 +29,46 @@ task_categories:
 <img src="https://cdn-uploads.huggingface.co/production/uploads/604b97e27032db3f5e8d6e8e/0IwaKLOehSDEFF3zYY_Wp.png" alt="Karanta OCR Logo" width="100"/>
 </div>
 
-**KarantaOCR-Bench** is a unit-test–style evaluation dataset, similar to [olmOCR-bench](https://huggingface.co/datasets/allenai/olmOCR-bench). It consists of **1,890 PDF documents** and **4,426 test cases** spanning multiple document types.
-
-KarantaOCR-Bench is designed specifically to evaluate document text extraction for **African languages**, covering a diverse range of document formats and languages commonly under-represented in existing OCR benchmarks.
+**KarantaOCR-Bench** is a unit-test–style evaluation dataset, similar to [olmOCR-bench](https://huggingface.co/datasets/allenai/olmOCR-bench). It consists of **70 PDF documents** and **300 test cases** spanning multiple document types.
+All the tests were manually verified by us.
+KarantaOCR-Bench is designed specifically to evaluate document text extraction for **Documents with diacritics and special characters**, covering a diverse range of document formats and languages commonly under-represented in existing OCR benchmarks.
 
 ### Data Schema
 
 Below is a schema of each test case and what it means:
 
-- pdf [str]: This is the name of the pdf file in the `pdf` directory
-- page [str]: This is the page of the document for this test case
-- id [str]: Unique ID for this test case
-- type [str]: This is the test type; a description of each test type can be found in the next section
-- max_diffs [int]: This is the maximum allowable difference in the OCR outputs and test case for it to pass
-- text [str]: For "text-present" tests, this is the text span that we search for in the OCR outputs and for "text-absent" tests this is the text that we do not expect to see in the output
-- case_sensitive [bool]: This indicates whether the test case is case sensitive or not
-- first_n [int]: When provided, it means we should check for the text in the first n characters
-- last_n [int]: When provided, it means we should check for the text in the last n characters
-- checked [str]: This indicates whether or not we have verified this test case manually
-- url [str]: The original file name of the document
-- before [str]: This only applies to text-order tests and this is the text span that should exist before
-- after [str]:This only applies to text-order tests and this is the text span that should exist after
+- `pdf [str]`: This is the name of the pdf file in the `pdf` directory
+- `page [str]`: This is the page of the document for this test case
+- `id [str]`: Unique ID for this test case
+- `type [str]`: This is the test type; a description of each test type can be found in the next section
+- `max_diffs [int]`: This is the maximum allowable difference in the OCR outputs and test case for it to pass
+- `text [str]`: This is the text span that we search for in the OCR outputt.
+- `case_sensitive [bool]`: This indicates whether the test case is case sensitive or not
+- `checked [str]`: This indicates whether or not we have verified this test case manually
+- `url [str]`: The original file name of the document
+- `before [str]`: This only applies to text-order tests and this is the text span that should exist before
+- `after [str]`:This only applies to text-order tests and this is the text span that should exist after
 
 
 ### Test Case Types
 
-KarantaOCR-Bench contains three types of test cases:
-
-* **Text Order:**
-  Text Order tests verify the relative ordering of two distinct spans of text within a document. The test checks that the *before text* appears earlier in the OCR output than the *after text*. If either the *before text* or *after text* is missing, the test is marked as failed.
+KarantaOCR-Bench is designed to augment OlmoOCR test with an additional test category for Diacritics and Special Chars:
 
 * **Presence:**
   Presence tests check whether a provided span of text exists in the OCR output. Matches are evaluated with a tolerance defined by the `max_diffs` field, allowing for limited character-level differences.
 
-* **Absent:**
-  Since the primary goal of KarantaOCR is to extract clean text suitable for LLM training, Absent tests ensure that undesired content—such as headers, footers, or page numbers—is excluded from the OCR output. These tests help maintain continuity across pages and prevent boilerplate text from polluting downstream training data.
-
 ## Results
 
-Coming soon ...
+| JSONL File      | KarantaOCR (3B) | RoLMOCR (7B) | NanoNetsOCR-2 (3B)  | OLMOCR-2 (7B)  |
+| --------------- | ---------- | -------- | ------------- | -------- | 
+| Diacritics & Special Characters | **61.6** | 53.6 | 55.1 | 55.5 |
 
 
 ## How to Run the Benchmark on your Model
 
-Coming soon ...
+We are working on a comprehensive guide on how to evaluate your models on this benchmark.
+
+Coming Soon...
 
 
 ## Citation Information