Upload 7 files

.gitattributes

@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+Image1.jpg filter=lfs diff=lfs merge=lfs -text
+Image2.png filter=lfs diff=lfs merge=lfs -text
+Image3.png filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,36 @@
+import streamlit as st
+from functions import predict_step
+from itertools import cycle
+
+
+def image_uploader():
+    with st.form("uploader"):
+        images = st.file_uploader("Upload Images",accept_multiple_files=True,type=["png","jpg","jpeg"])
+        submitted = st.form_submit_button("Submit")
+        if submitted:
+            predicted_captions = predict_step(images,False)
+            for i,caption in enumerate(predicted_captions):
+                st.write(str(i+1)+'. '+caption)
+
+def images_url():
+    with st.form("url"):
+        urls = st.text_input('Enter URL of Images followed by comma for multiple URLs')
+        images = urls.split(',')
+        submitted = st.form_submit_button("Submit")
+        if submitted:
+            predicted_captions = predict_step(images,True)
+            for i,caption in enumerate(predicted_captions):
+                st.write(str(i+1)+'. '+caption)
+
+def main():
+    st.set_page_config(page_title="Image Captioning", page_icon="🖼️")
+    st.title("Image Caption")
+
+    st.subheader("Upload your own Images")
+    image_uploader()
+
+    st.subheader("Enter Image URLs")
+    images_url()
+
+if __name__ == '__main__':
+    main()

functions.py

@@ -0,0 +1,48 @@
+from PIL import Image
+from tqdm import tqdm
+from transformers import VisionEncoderDecoderModel, ViTFeatureExtractor, AutoTokenizer
+import torch 
+from PIL import Image
+from tqdm import tqdm
+import urllib.request
+from itertools import cycle
+import os
+
+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
+num_return_sequences = 3  # Number of captions to generate for each image
+gen_kwargs = {"max_length": max_length, "num_beams": num_beams, "num_return_sequences": num_return_sequences}
+
+
+def predict_step(images_list,is_url):
+    images = []
+    for image in tqdm(images_list):
+        if is_url:
+            urllib.request.urlretrieve(image, "file.jpg")
+            i_image = Image.open("file.jpg")
+            
+        else:
+            i_image = Image.open(image)
+
+        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, clean_up_tokenization_spaces=True)
+    preds = [pred.strip() for pred in preds]
+    if is_url:
+        os.remove('file.jpg')
+    return preds

readme.txt

@@ -0,0 +1,17 @@
+TASK
+● Create an AI tool that creates captions based on the image provided by the user. Should also have
+the option to generate multiple captions based on the image.
+● Provide an interface where the user can come and upload images and get AI generated captions. ●
+You are to free use the library of your choice
+● Use the following images as test cases - Link
+Note - Try to use pre-trained models from websites like huggingface.
+
+To Put it out there
+Your assignment will also be evaluated on the following criteria
+● Code quality, the less the code the better.
+● Time taken to submit, the less the better.
+● Response times, the faster the better.
+
+Project Link: 
+
+https://drive.google.com/drive/folders/1Ekn8HzzHbo0oULYo6o8aSeju1hDdgJnV?usp=share_link

requirements.txt

@@ -0,0 +1,12 @@
+fastapi==0.78.0
+pandas==1.5.0
+pydantic==1.10.2
+scikit-learn==1.1.2
+servicefoundry
+mlfoundry
+streamlit==1.13.0
+uvicorn==0.18.3
+xgboost==1.6.2
+torch
+transformers
+tqdm