fix: image preprocessing

app.py

@@ -28,8 +28,7 @@ image_processor = git_processor_base.image_processor
 def preprocess_image(image):
     """Преобразуем изображение для модели"""
     image_rgb = image.convert("RGB")
-    processed = image_processor(image_rgb, return_tensors="pt").pixel_values.to(device)
-    return processed.pixel_values, processed.words, processed.boxes
+    return image_processor(image_rgb, return_tensors="pt").pixel_values.to(device)
 
 
 def translate_ru2en(text):
@@ -46,13 +45,13 @@ def translate_en2ru(text):
     return translated_text
 
 
-def generate_answer_git(processor, model, image, question):
+def generate_answer_git(image, question):
 
     qa_pipeline = pipeline(
         "document-question-answering",
         model="andgrt/layoutlmv2-base-uncased_finetuned_docvqa",
     )
-    return qa_pipeline(image, question)[0]["answer"]
+    return qa_pipeline(preprocess_image(image), question)[0]["answer"]
     # pixel_values, _, _ = preprocess_image(image)
     # input_ids = processor(text=question, add_special_tokens=False).input_ids
     # input_ids = [processor.tokenizer.cls_token_id] + input_ids
@@ -69,9 +68,7 @@ def generate_answer(image, question):
     question_en = translate_ru2en(question)
     print(f"Вопрос на английском: {question_en}")
 
-    answer_en = generate_answer_git(
-        git_processor_base, git_model_base, image, question_en
-    )
+    answer_en = generate_answer_git(image, question_en)
     print(f"Ответ на английском: {answer_en}")
 
     answer_ru = translate_en2ru(answer_en)