refactor(memory management): remove unused clean_memory function

The clean_memory function has been removed from pdf_parse_union_core_v2.py due to it not being used.
This change streamlines the code and prevents potential confusion regarding its purpose.

magic_pdf/libs/clean_memory.py

@@ -0,0 +1,10 @@
+# Copyright (c) Opendatalab. All rights reserved.
+import torch
+import gc
+
+
+def clean_memory():
+    if torch.cuda.is_available():
+        torch.cuda.empty_cache()
+        torch.cuda.ipc_collect()
+    gc.collect()

magic_pdf/model/pdf_extract_kit.py

@@ -3,6 +3,7 @@
 import time
 
 from magic_pdf.libs.Constants import *
+from magic_pdf.libs.clean_memory import clean_memory
 from magic_pdf.model.model_list import AtomicModel
 
 os.environ['NO_ALBUMENTATIONS_UPDATE'] = '1'  # 禁止albumentations检查更新
@@ -330,6 +331,8 @@ def __call__(self, image):
             elif int(res['category_id']) in [5]:
                 table_res_list.append(res)
 
+        clean_memory()
+
         # ocr识别
         if self.apply_ocr:
             ocr_start = time.time()

magic_pdf/pdf_parse_union_core_v2.py

@@ -7,6 +7,7 @@
 
 import torch
 
+from magic_pdf.libs.clean_memory import clean_memory
 from magic_pdf.libs.commons import fitz, get_delta_time
 from magic_pdf.libs.convert_utils import dict_to_list
 from magic_pdf.libs.drop_reason import DropReason
@@ -304,14 +305,6 @@ def parse_page_core(pdf_docs, magic_model, page_id, pdf_bytes_md5, imageWriter,
     return page_info
 
 
-def clean_memory():
-    import gc
-    if torch.cuda.is_available():
-        torch.cuda.empty_cache()
-        torch.cuda.ipc_collect()
-    gc.collect()
-
-
 def pdf_parse_union(pdf_bytes,
                     model_list,
                     imageWriter,