mem: reduce PaddleOCR rec_batch_num from 6 to 1

Paddle's native inference engine allocates 500 MiB memory arena chunks
during text recognition, proportional to batch size. With the default
rec_batch_num=6, four 500 MiB chunks are allocated simultaneously.

Setting rec_batch_num=1 reduces this to a single chunk, cutting peak
memory on the PaddleOCR code path by ~1,265 MiB (-42.6%).

Latency benchmark (55 text regions, CPU, 5 runs):
- rec_batch_num=6: 39.1s +/- 3.5s
- rec_batch_num=1: 37.0s +/- 2.0s
No throughput regression — on CPU, batch processing is sequential.

Author: KRRT7

unstructured/partition/utils/ocr_models/paddle_ocr.py

@@ -48,6 +48,7 @@ def load_agent(self, language: str):
                 lang=language,
                 enable_mkldnn=True,
                 show_log=False,
+                rec_batch_num=1,
             )
         except AttributeError:
             paddle_ocr = PaddleOCR(
@@ -56,6 +57,7 @@ def load_agent(self, language: str):
                 lang=language,
                 enable_mkldnn=False,
                 show_log=False,
+                rec_batch_num=1,
             )
         return paddle_ocr