Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -76,19 +76,25 @@ tokenizer = MBartTokenizer.from_pretrained(
|
|
76 |
'facebook/mbart-large-50'
|
77 |
)
|
78 |
processortext2 = CustomOCRProcessor(image_processor,tokenizer)
|
79 |
-
import os
|
80 |
|
|
|
81 |
huggingface_token = os.environ.get("HUGGINGFACE_TOKEN")
|
82 |
-
|
|
|
|
|
|
|
|
|
|
|
83 |
st.title("Image OCR with musadac/vilanocr")
|
|
|
84 |
uploaded_file = st.file_uploader("Choose an image", type=["jpg", "jpeg", "png"])
|
85 |
-
|
86 |
if uploaded_file is not None:
|
|
|
87 |
img = Image.open(uploaded_file).convert("RGB")
|
88 |
pixel_values = processortext2(img.convert("RGB"), return_tensors="pt").pixel_values
|
89 |
|
90 |
with torch.no_grad():
|
91 |
-
generated_ids =
|
92 |
|
93 |
result = processortext2.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
94 |
st.write("OCR Result:")
|
|
|
76 |
'facebook/mbart-large-50'
|
77 |
)
|
78 |
processortext2 = CustomOCRProcessor(image_processor,tokenizer)
|
|
|
79 |
|
80 |
+
import os
|
81 |
huggingface_token = os.environ.get("HUGGINGFACE_TOKEN")
|
82 |
+
model = {}
|
83 |
+
model['single-urdu'] = VisionEncoderDecoderModel.from_pretrained("musadac/vilanocr-single-urdu", use_auth_token=huggingface_token)
|
84 |
+
model['multi-urdu'] = VisionEncoderDecoderModel.from_pretrained("musadac/ViLanOCR", use_auth_token=huggingface_token)
|
85 |
+
model['medical'] = VisionEncoderDecoderModel.from_pretrained("musadac/vilanocr-multi-medical", use_auth_token=huggingface_token)
|
86 |
+
model['chinese'] = VisionEncoderDecoderModel.from_pretrained("musadac/vilanocr-single-chinese", use_auth_token=huggingface_token)
|
87 |
+
|
88 |
st.title("Image OCR with musadac/vilanocr")
|
89 |
+
model_name = st.selectbox("Choose an OCR model", ["single-urdu", "multi-urdu", "medical","chinese" ])
|
90 |
uploaded_file = st.file_uploader("Choose an image", type=["jpg", "jpeg", "png"])
|
|
|
91 |
if uploaded_file is not None:
|
92 |
+
|
93 |
img = Image.open(uploaded_file).convert("RGB")
|
94 |
pixel_values = processortext2(img.convert("RGB"), return_tensors="pt").pixel_values
|
95 |
|
96 |
with torch.no_grad():
|
97 |
+
generated_ids = model[model_name].generate(pixel_values)
|
98 |
|
99 |
result = processortext2.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
100 |
st.write("OCR Result:")
|