Update app.py

app.py

@@ -1,9 +1,13 @@
-#!/usr/bin/env python
-# coding: utf-8
-
-# In[ ]:
+# -*- coding: utf-8 -*-
+"""Image Captioning with ViT+GPT2
+Automatically generated by Colaboratory.
+Original file is located at
+    https://colab.research.google.com/drive/1P3O0gO5AUqSmM8rE9dxy2tXJ-9jkhxHz
+"""
 
+#! pip install transformers -q
 
+#! pip install gradio -q
 
 from PIL import Image
 from transformers import VisionEncoderDecoderModel, ViTFeatureExtractor, PreTrainedTokenizerFast
@@ -15,6 +19,21 @@ vit_feature_extractor = ViTFeatureExtractor.from_pretrained("google/vit-base-pat
 
 tokenizer = PreTrainedTokenizerFast.from_pretrained("distilgpt2")
 
+# url = 'https://d2gp644kobdlm6.cloudfront.net/wp-content/uploads/2016/06/bigstock-Shocked-and-surprised-boy-on-t-113798588-300x212.jpg'
+
+# with Image.open(requests.get(url, stream=True).raw) as img:
+#     pixel_values = vit_feature_extractor(images=img, return_tensors="pt").pixel_values
+
+#encoder_outputs = model.generate(pixel_values.to('cpu'),num_beams=5)
+
+#generated_sentences = tokenizer.batch_decode(encoder_outputs, skip_special_tokens=True)
+
+#generated_sentences
+
+#naive text processing 
+#generated_sentences[0].split('.')[0]
+
+# inference function
 
 def vit2distilgpt2(img):
   pixel_values = vit_feature_extractor(images=img, return_tensors="pt").pixel_values
@@ -23,7 +42,7 @@ def vit2distilgpt2(img):
 
   return(generated_sentences[0].split('.')[0])
 
-
+#!wget https://media.glamour.com/photos/5f171c4fd35176eaedb36823/master/w_2560%2Cc_limit/bike.jpg
 
 import gradio as gr
 
@@ -54,5 +73,4 @@ gr.Interface(
     article=article,
     examples=examples,
     theme="huggingface",
-).launch(debug=True, enable_queue=True)
-
+).launch(debug=True, enable_queue=True)