Update app.py

app.py

@@ -1,16 +1,25 @@
 import gradio as gr
-from transformers import pipeline, BertTokenizer, BertModel
+from transformers import BlipProcessor, BlipForConditionalGeneration
 
-tokenizer = BertTokenizer.from_pretrained('bert-large-uncased')
-model = BertModel.from_pretrained("bert-large-uncased")
-unmasker = pipeline('fill-mask', model=model, tokenizer=tokenizer)
+# Load the model and tokenizer
+model_name = "Salesforce/blip-image-captioning-large"
+processor = BlipProcessor.from_pretrained(model_name)
+model = BlipForConditionalGeneration.from_pretrained(model_name)
 
-def fill_mask(sentence):
-    results = unmasker(sentence)
-    return [result['sequence'] for result in results]
+def generate_caption(image):
+    # Preprocess the image
+    inputs = processor(images=image, return_tensors="pt")
 
-inputs = gr.inputs.Image(label="Upload an image")
-outputs = gr.outputs.Textbox(label="Predicted sequences")
+    # Generate caption using the model
+    caption = model.generate(**inputs)
+
+    # Decode the output caption
+    decoded_caption = processor.decode(caption[0], skip_special_tokens=True)
+    return decoded_caption
 
-gr.Interface(fn=fill_mask, inputs=inputs, outputs=outputs).launch()
+# Define the Gradio interface
+inputs = gr.inputs.Image(label="Upload an image")
+outputs = gr.outputs.Textbox(label="Generated Caption")
 
+# Create the Gradio app
+gr.Interface(fn=generate_caption, inputs=inputs, outputs=outputs).launch()