Create app.py

app.py

@@ -0,0 +1,38 @@
+import torch
+import re
+import gradio as gr
+from transformers import AutoTokenizer,ViTFeatureExtractor VisionEncoderDecoderModel
+
+device = 'cpu'
+encoder_checkpoint = 'google/vit-base-patch16-224'
+decoder_checkpoint = 'gpt2'
+model_checkpoint = 'nlpconnect/vit-gpt2-image-captioning'
+feature_extractor = ViTFeatureExtractor.from_pretrained(encoder_checkpoint)
+model = VisionEncoderDecoderModel.from_pretrained(model_checkpoint).to(device)
+
+
+def predict(image,max_length=64,num_beams=4):
+  image = image.convert('RGB')
+  image = feature_extractor(image,return_tensor='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
+  
+
+input = gr.inputs.Image(label='Image to generate caption',type = 'pil', optional=False)
+output = gr.outputs.Textbox(type="auto",label="Caption")
+
+article = "This is a Image captioning model created by Shreyas Dixit"
+title = "Image Captioning"
+
+interface = gr.Interface(
+        fn=predict,
+        inputs = input,
+        theme="grass",
+        outputs=output,
+        examples = examples,
+        title=title,
+        description=article,
+    )
+interface.launch(debug=True)