Update app.py

app.py

@@ -1,35 +1,39 @@
-import easyocr as ocr  #OCR
-from PIL import Image #Image Processing
-import numpy as np #Image Processing 
+import torch 
+import re 
 import gradio as gr
+from transformers import AutoTokenizer, ViTFeatureExtractor, VisionEncoderDecoderModel 
 
+device='cpu'
+encoder_checkpoint = "jaimin/image_caption"
+decoder_checkpoint = "jaimin/image_caption"
+model_checkpoint = "jaimin/image_caption"
+feature_extractor = ViTFeatureExtractor.from_pretrained(encoder_checkpoint)
+tokenizer = AutoTokenizer.from_pretrained(decoder_checkpoint)
+model = VisionEncoderDecoderModel.from_pretrained(model_checkpoint).to(device)
 
-input = gr.inputs.Image(label="Upload your Image", type= 'pil', optional=True)
 
-def load_model(): 
-    reader = ocr.Reader(['en'],model_storage_directory='.')
-    return reader 
+def predict(image,max_length=64, num_beams=4):
+  image = image.convert('RGB')
+  image = feature_extractor(image, return_tensors="pt").pixel_values.to(device)
+  clean_text = lambda x: x.replace('<|endoftext|>','').split('\n')[0]
+  caption_ids = model.generate(image, max_length = max_length)[0]
+  caption_text = clean_text(tokenizer.decode(caption_ids))
+  return caption_text 
 
-def text(image):
-    reader = load_model() #load model
-    print("Model Load")
-    input_image = image.convert('RGB') #read image
-    print("image read")
-    result = reader.readtext(np.array(input_image))
-    result_text = [] #empty list for results
-    for text in result:
-        result_text.append(text[1])
-    return result_text
 
-output = gr.outputs.Textbox(type="text",label="Captions")
 
-title = "Image Captioning "
+input = gr.inputs.Image(label="Upload your Image", type = 'pil', optional=True)
+output = gr.outputs.Textbox(type="auto",label="Captions")
+examples = [f"example{i}.jpg" for i in range(1,7)]
+
+title = "Image To Text"
 
 interface = gr.Interface(
-        fn=text,
+        fn=predict,
         inputs = input,
         theme="grass",
         outputs=output,
+        examples = examples,
         title=title,
     )
 interface.launch(debug=True)