Upload dataset

README.md

@@ -12,7 +12,7 @@ tags:
 - benchmark
 pretty_name: Azerbaijani Handwritten OCR Benchmark
 dataset_info:
-  config_name: lines
+- config_name: lines
   features:
   - name: image
     dtype: image
@@ -30,6 +30,37 @@ dataset_info:
     num_examples: 442
   download_size: 10255786
   dataset_size: 10277269.0
+- config_name: pages
+  features:
+  - name: image
+    dtype: image
+  - name: page_id
+    dtype: string
+  - name: source_image
+    dtype: string
+  - name: image_width
+    dtype: int32
+  - name: image_height
+    dtype: int32
+  - name: num_lines
+    dtype: int32
+  - name: lines
+    sequence:
+    - name: line_id
+      dtype: int32
+    - name: bbox
+      sequence: int32
+      length: 4
+    - name: text
+      dtype: string
+  - name: full_text
+    dtype: string
+  splits:
+  - name: test
+    num_bytes: 11571721.0
+    num_examples: 29
+  download_size: 11566491
+  dataset_size: 11571721.0
 configs:
 - config_name: default
   data_files:
@@ -39,6 +70,10 @@ configs:
   data_files:
   - split: test
     path: lines/test-*
+- config_name: pages
+  data_files:
+  - split: test
+    path: pages/test-*
 ---
 
 # Azerbaijani Handwritten OCR Benchmark

pages/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c43f8cf9db90d4997ba417a4a9baa9be01bb90bc7289cba7845a229310e76fb
+size 11566491