Spaces:
Runtime error
Runtime error
Upload app.py
Browse files
app.py
CHANGED
|
@@ -9,9 +9,9 @@ from PIL import Image
|
|
| 9 |
device = 'cpu'
|
| 10 |
|
| 11 |
# Load the pretrained model, feature extractor, and tokenizer
|
| 12 |
-
model = VisionEncoderDecoderModel.from_pretrained("NourFakih/Vit-GPT2-COCO2017Flickr-
|
| 13 |
-
feature_extractor = ViTImageProcessor.from_pretrained("NourFakih/Vit-GPT2-COCO2017Flickr-
|
| 14 |
-
tokenizer = AutoTokenizer.from_pretrained("NourFakih/Vit-GPT2-COCO2017Flickr-
|
| 15 |
|
| 16 |
def predict(image, max_length=64, num_beams=4):
|
| 17 |
# Process the input image
|
|
@@ -36,14 +36,7 @@ def process_images(image_files):
|
|
| 36 |
except Exception as e:
|
| 37 |
print(f"Skipping file {image_file}: {e}")
|
| 38 |
|
| 39 |
-
|
| 40 |
-
csv_file_path = 'image_captions.csv'
|
| 41 |
-
with open(csv_file_path, mode='w', newline='') as file:
|
| 42 |
-
writer = csv.writer(file)
|
| 43 |
-
writer.writerow(['Image Name', 'Caption'])
|
| 44 |
-
writer.writerows(captions)
|
| 45 |
-
|
| 46 |
-
return csv_file_path
|
| 47 |
|
| 48 |
def process_zip_files(zip_file_paths):
|
| 49 |
# Create a directory to extract images
|
|
@@ -67,24 +60,37 @@ def process_zip_files(zip_file_paths):
|
|
| 67 |
except Exception as e:
|
| 68 |
print(f"Skipping file {file}: {e}")
|
| 69 |
|
| 70 |
-
|
| 71 |
-
|
|
|
|
| 72 |
with open(csv_file_path, mode='w', newline='') as file:
|
| 73 |
writer = csv.writer(file)
|
| 74 |
writer.writerow(['Image Name', 'Caption'])
|
| 75 |
writer.writerows(captions)
|
| 76 |
-
|
| 77 |
return csv_file_path
|
| 78 |
|
| 79 |
-
def
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 88 |
|
| 89 |
css = '''
|
| 90 |
h1#title {
|
|
@@ -113,14 +119,31 @@ with demo:
|
|
| 113 |
gr.Markdown('''<h1 id="title">Image Caption 🖼️</h1>''')
|
| 114 |
gr.Markdown('''Made by : No. Fa.''')
|
| 115 |
|
|
|
|
|
|
|
|
|
|
| 116 |
with gr.Row():
|
| 117 |
with gr.Column(scale=1):
|
| 118 |
-
|
| 119 |
-
|
| 120 |
with gr.Column(scale=3):
|
| 121 |
-
|
| 122 |
-
|
| 123 |
-
|
| 124 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 125 |
|
| 126 |
-
demo.launch()
|
|
|
|
| 9 |
device = 'cpu'
|
| 10 |
|
| 11 |
# Load the pretrained model, feature extractor, and tokenizer
|
| 12 |
+
model = VisionEncoderDecoderModel.from_pretrained("NourFakih/Vit-GPT2-COCO2017Flickr-01").to(device)
|
| 13 |
+
feature_extractor = ViTImageProcessor.from_pretrained("NourFakih/Vit-GPT2-COCO2017Flickr-01")
|
| 14 |
+
tokenizer = AutoTokenizer.from_pretrained("NourFakih/Vit-GPT2-COCO2017Flickr-01")
|
| 15 |
|
| 16 |
def predict(image, max_length=64, num_beams=4):
|
| 17 |
# Process the input image
|
|
|
|
| 36 |
except Exception as e:
|
| 37 |
print(f"Skipping file {image_file}: {e}")
|
| 38 |
|
| 39 |
+
return captions
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
|
| 41 |
def process_zip_files(zip_file_paths):
|
| 42 |
# Create a directory to extract images
|
|
|
|
| 60 |
except Exception as e:
|
| 61 |
print(f"Skipping file {file}: {e}")
|
| 62 |
|
| 63 |
+
return captions
|
| 64 |
+
|
| 65 |
+
def save_captions_to_csv(captions, csv_file_path='image_captions.csv'):
|
| 66 |
with open(csv_file_path, mode='w', newline='') as file:
|
| 67 |
writer = csv.writer(file)
|
| 68 |
writer.writerow(['Image Name', 'Caption'])
|
| 69 |
writer.writerows(captions)
|
|
|
|
| 70 |
return csv_file_path
|
| 71 |
|
| 72 |
+
def gr_process_zip(zip_files):
|
| 73 |
+
captions = process_zip_files([zip_file.name for zip_file in zip_files])
|
| 74 |
+
return save_captions_to_csv(captions, 'zip_image_captions.csv')
|
| 75 |
+
|
| 76 |
+
def gr_process_images(image_files):
|
| 77 |
+
captions = process_images([image_file.name for image_file in image_files])
|
| 78 |
+
return save_captions_to_csv(captions, 'image_captions.csv')
|
| 79 |
+
|
| 80 |
+
def combine_csv_files(file1, file2, output_file='combined_captions.csv'):
|
| 81 |
+
with open(output_file, mode='w', newline='') as outfile:
|
| 82 |
+
writer = csv.writer(outfile)
|
| 83 |
+
writer.writerow(['Image Name', 'Caption'])
|
| 84 |
+
|
| 85 |
+
for file in [file1, file2]:
|
| 86 |
+
if os.path.exists(file):
|
| 87 |
+
with open(file, mode='r') as infile:
|
| 88 |
+
reader = csv.reader(infile)
|
| 89 |
+
next(reader) # Skip header row
|
| 90 |
+
for row in reader:
|
| 91 |
+
writer.writerow(row)
|
| 92 |
+
|
| 93 |
+
return output_file
|
| 94 |
|
| 95 |
css = '''
|
| 96 |
h1#title {
|
|
|
|
| 119 |
gr.Markdown('''<h1 id="title">Image Caption 🖼️</h1>''')
|
| 120 |
gr.Markdown('''Made by : No. Fa.''')
|
| 121 |
|
| 122 |
+
zip_files = gr.State([])
|
| 123 |
+
image_files = gr.State([])
|
| 124 |
+
|
| 125 |
with gr.Row():
|
| 126 |
with gr.Column(scale=1):
|
| 127 |
+
new_zip_files = gr.File(label="Upload Zip Files", type="filepath", file_count="multiple")
|
| 128 |
+
new_image_files = gr.File(label="Upload Images", type="filepath", file_count="multiple")
|
| 129 |
with gr.Column(scale=3):
|
| 130 |
+
output_zip_file = gr.File(label="Download Zip Captions")
|
| 131 |
+
output_image_file = gr.File(label="Download Image Captions")
|
| 132 |
+
combined_file = gr.File(label="Download Combined Captions")
|
| 133 |
+
|
| 134 |
+
add_files_btn = gr.Button("Add More Files")
|
| 135 |
+
generate_zip_captions_btn = gr.Button("Generate Zip Captions")
|
| 136 |
+
generate_image_captions_btn = gr.Button("Generate Image Captions")
|
| 137 |
+
combine_files_btn = gr.Button("Combine CSV Files")
|
| 138 |
+
|
| 139 |
+
def add_files(zip_files, image_files, new_zip_files, new_image_files):
|
| 140 |
+
zip_files.extend(new_zip_files)
|
| 141 |
+
image_files.extend(new_image_files)
|
| 142 |
+
return zip_files, image_files
|
| 143 |
+
|
| 144 |
+
add_files_btn.click(add_files, inputs=[zip_files, image_files, new_zip_files, new_image_files], outputs=[zip_files, image_files])
|
| 145 |
+
generate_zip_captions_btn.click(fn=gr_process_zip, inputs=zip_files, outputs=output_zip_file)
|
| 146 |
+
generate_image_captions_btn.click(fn=gr_process_images, inputs=image_files, outputs=output_image_file)
|
| 147 |
+
combine_files_btn.click(fn=combine_csv_files, inputs=[output_zip_file, output_image_file], outputs=combined_file)
|
| 148 |
|
| 149 |
+
demo.launch(share=True)
|