jkorstad commited on
Commit
5bcf31a
·
verified ·
1 Parent(s): b2dd8bf

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +15 -5
app.py CHANGED
@@ -1,21 +1,31 @@
1
  import gradio as gr
2
- from transformers import AutoProcessor, AutoModelForCausalLM
3
  from pdf2image import convert_from_path
4
  import base64
5
  import io
6
- import spaces
7
  from PIL import Image
8
 
9
  # Load the OCR model and processor from Hugging Face
10
- processor = AutoProcessor.from_pretrained("allenai/olmOCR-7B-0225-preview")
11
- model = AutoModelForCausalLM.from_pretrained("allenai/olmOCR-7B-0225-preview")
 
 
 
 
 
 
 
 
 
12
 
13
- @spaces.GPU
14
  def process_pdf(pdf_file):
15
  """
16
  Process the uploaded PDF file, extract text from each page, and generate HTML
17
  to display each page's image and text with copy buttons.
18
  """
 
 
 
19
  # Check if a PDF file was uploaded
20
  if pdf_file is None:
21
  return "<p>Please upload a PDF file.</p>"
 
1
  import gradio as gr
2
+ from transformers import AutoProcessor, AutoModelForVision2Seq
3
  from pdf2image import convert_from_path
4
  import base64
5
  import io
 
6
  from PIL import Image
7
 
8
  # Load the OCR model and processor from Hugging Face
9
+ try:
10
+ processor = AutoProcessor.from_pretrained("allenai/olmOCR-7B-0225-preview")
11
+ model = AutoModelForVision2Seq.from_pretrained("allenai/olmOCR-7B-0225-preview")
12
+ except ImportError as e:
13
+ processor = None
14
+ model = None
15
+ print(f"Error loading model: {str(e)}. Please ensure PyTorch is installed.")
16
+ except ValueError as e:
17
+ processor = None
18
+ model = None
19
+ print(f"Error with model configuration: {str(e)}")
20
 
 
21
  def process_pdf(pdf_file):
22
  """
23
  Process the uploaded PDF file, extract text from each page, and generate HTML
24
  to display each page's image and text with copy buttons.
25
  """
26
+ if processor is None or model is None:
27
+ return "<p>Error: Model could not be loaded. Check environment setup (PyTorch may be missing) or model compatibility.</p>"
28
+
29
  # Check if a PDF file was uploaded
30
  if pdf_file is None:
31
  return "<p>Please upload a PDF file.</p>"