Hugging Face's logo

Spaces:
prithivMLmods
/
Multimodal-OCR3
Running on Zero

App Files Community
1
prithivMLmods commited on
Commit
d389dfd
·
verified ·
1 Parent(s): f452786

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +3 -2
app.py CHANGED
@@ -15,7 +15,8 @@ import cv2
15
 
16
  from transformers import (
17
  Qwen2_5_VLForConditionalGeneration,
18
- AutoModelForImageTextToText, # Added for PaddleOCR-VL
 
19
  AutoProcessor,
20
  TextIteratorStreamer,
21
  )
@@ -136,7 +137,7 @@ model_v = Qwen2_5_VLForConditionalGeneration.from_pretrained(
136
  MODEL_ID_P = "PaddlePaddle/PaddleOCR-VL"
137
  SUBFOLDER_P = "PaddleOCR-VL-0.9B"
138
  processor_p = AutoProcessor.from_pretrained(MODEL_ID_P, trust_remote_code=True, subfolder=SUBFOLDER_P)
139
- model_p = AutoModelForImageTextToText.from_pretrained(
140
  MODEL_ID_P,
141
  trust_remote_code=True,
142
  subfolder=SUBFOLDER_P,
 
15
 
16
  from transformers import (
17
  Qwen2_5_VLForConditionalGeneration,
18
+ AutoModelForImageTextToText,
19
+ AutoModelForCausalLM,# Added for PaddleOCR-VL
20
  AutoProcessor,
21
  TextIteratorStreamer,
22
  )
 
137
  MODEL_ID_P = "PaddlePaddle/PaddleOCR-VL"
138
  SUBFOLDER_P = "PaddleOCR-VL-0.9B"
139
  processor_p = AutoProcessor.from_pretrained(MODEL_ID_P, trust_remote_code=True, subfolder=SUBFOLDER_P)
140
+ model_p = AutoModelForCausalLM.from_pretrained(
141
  MODEL_ID_P,
142
  trust_remote_code=True,
143
  subfolder=SUBFOLDER_P,