Update app.py

app.py

@@ -2,7 +2,7 @@ import gradio as gr
 from langchain.vectorstores import Chroma
 from langchain_community.document_loaders import PyPDFLoader
 from langchain_community.embeddings import HuggingFaceEmbeddings
-from transformers import LayoutLMv3Processor, AutoModelForTokenClassification
+from transformers import LayoutLMv3Processor, AutoModelForSeq2SeqLM
 from langchain.chains import RetrievalQA
 from langchain.prompts import PromptTemplate
 from pdf2image import convert_from_path
@@ -11,15 +11,15 @@ import os
 class LayoutLMv3OCR:
     def __init__(self):
         self.processor = LayoutLMv3Processor.from_pretrained("microsoft/layoutlmv3-base")
-        self.model = AutoModelForTokenClassification.from_pretrained("microsoft/layoutlmv3-base")
+        self.model = AutoModelForSeq2SeqLM.from_pretrained("microsoft/layoutlmv3-base")
 
     def extract_text(self, pdf_path):
         images = convert_from_path(pdf_path)
         text_pages = []
         for image in images:
             inputs = self.processor(images=image, return_tensors="pt")
-            outputs = self.model(**inputs)
-            text = self.processor.batch_decode(outputs.logits, skip_special_tokens=True)[0]
+            outputs = self.model.generate(**inputs)
+            text = self.processor.batch_decode(outputs, skip_special_tokens=True)[0]
             text_pages.append(text)
         return text_pages
 
@@ -41,14 +41,21 @@ def process_pdf_and_query(pdf_path, question):
     return response
 
 def chatbot_response(pdf, question):
+    # Speichern der hochgeladenen Datei auf dem lokalen Dateisystem
     pdf_path = "uploaded_pdf.pdf"
-    pdf.save(pdf_path)
+    
+    # Schreibe die PDF-Datei in eine lokale Datei
+    with open(pdf_path, "wb") as f:
+        f.write(pdf.read())
+    
     extracted_text = ocr_tool.extract_text(pdf_path)
     answer = process_pdf_and_query(pdf_path, question)
+    
+    # Lösche die gespeicherte PDF-Datei nach der Verarbeitung
     os.remove(pdf_path)
+    
     return answer
 
-# Ändere 'inputs' und 'outputs' zur neuen Gradio API
 pdf_input = gr.File(label="PDF-Datei hochladen")
 question_input = gr.Textbox(label="Frage eingeben")
 response_output = gr.Textbox(label="Antwort")
@@ -62,4 +69,4 @@ interface = gr.Interface(
 )
 
 if __name__ == "__main__":
-    interface.launch()
+    interface.launch(share=True)