Upload app.py

app.py

@@ -0,0 +1,97 @@
+# -*- coding: utf-8 -*-
+"""caption.ipynb
+
+Automatically generated by Colaboratory.
+
+Original file is located at
+    https://colab.research.google.com/drive/17BgQi1eU254RKp6BKOdC-Kfr1LqIwKmj
+
+## Image Caption Generator 
+
+
+In Colab, Pytorch comes preinstalled and same goes with PIL for Image. You will only need to install **transformers** from Huggingface.
+"""
+
+#!pip install transformers
+
+#from google.colab import drive
+#drive.mount('/content/drive')
+
+from transformers import VisionEncoderDecoderModel, ViTFeatureExtractor, AutoTokenizer
+import torch
+from PIL import Image
+
+model = VisionEncoderDecoderModel.from_pretrained("nlpconnect/vit-gpt2-image-captioning")
+feature_extractor = ViTFeatureExtractor.from_pretrained("nlpconnect/vit-gpt2-image-captioning")
+tokenizer = AutoTokenizer.from_pretrained("nlpconnect/vit-gpt2-image-captioning")
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model.to(device)
+
+
+
+max_length = 16
+num_beams = 4
+gen_kwargs = {"max_length": max_length, "num_beams": num_beams}
+def predict_step(image_paths):
+  images = []
+  for image_path in image_paths:
+    i_image = Image.open(image_path)
+    if i_image.mode != "RGB":
+      i_image = i_image.convert(mode="RGB")
+
+    images.append(i_image)
+
+  pixel_values = feature_extractor(images=images, return_tensors="pt").pixel_values
+  pixel_values = pixel_values.to(device)
+
+  output_ids = model.generate(pixel_values, **gen_kwargs)
+
+  preds = tokenizer.batch_decode(output_ids, skip_special_tokens=True)
+  preds = [pred.strip() for pred in preds]
+  return preds
+
+#predict_step(['/content/drive/MyDrive/caption generator/horses.png'])
+
+
+
+def caption_generator(img):
+  pixel_values = vit_feature_extractor(images=img, return_tensors="pt").pixel_values
+  encoder_outputs = generated_ids = model.generate(pixel_values.to('cpu'),num_beams=5)
+  generated_sentences = tokenizer.batch_decode(encoder_outputs, skip_special_tokens=True)
+
+
+
+import gradio as gr
+
+inputs = [
+    gr.inputs.Image(type="pil", label="Original Image")
+]
+
+outputs = [
+    gr.outputs.Textbox(label = 'Caption')
+]
+
+title = "Image Captioning using ViT + GPT2"
+description = "ViT and GPT2 are used to generate Image Caption for the uploaded image. COCO Dataset was used for training. This image captioning model might have some biases that we couldn't figure during our stress testing, so if you find any bias (gender, race and so on) please use `Flag` button to flag the image with bias"
+article = " <a href='https://huggingface.co/sachin/vit2distilgpt2'>Model Repo on Hugging Face Model Hub</a>"
+examples = [
+    ["horses.png"],
+    ["persons.jpeg"],
+    ['football_player'']
+
+]
+
+gr.Interface(
+    predict_step,
+    inputs,
+    outputs,
+    title=title,
+    description=description,
+    article=article,
+    examples=examples,
+    theme="huggingface",
+).launch(debug=True, enable_queue=True)
+
+
+