Update app.py to use gr.Blocks for multi-tab interface and fix ValueError; update requirements.txt for git-large-coco

app.py

@@ -109,36 +109,58 @@ def view_caption_history():
 def batch_generate_captions(image_list, caption_type: str = "descriptive", caption_length: str = "medium", prompt: str = ""):
     results = []
     for img in image_list:
-        caption = generate_caption(img, caption_type, caption_length, prompt)
-        results.append(f"Image {img.name}: {caption}")
+        # Convert file to PIL Image
+        img_pil = Image.open(img.name).convert("RGB")
+        caption = generate_caption(img_pil, caption_type, caption_length, prompt)
+        results.append(f"Image {os.path.basename(img.name)}: {caption}")
     return "\n".join(results)
 
-# Create the Gradio interface
-interface = gr.Interface(
-    fn=[generate_caption, view_caption_history, batch_generate_captions],
-    inputs=[
-        [
-            gr.Image(label="Upload Image", type="pil"),
-            gr.Dropdown(choices=["descriptive", "casual", "social media"], label="Caption Type", value="descriptive"),
-            gr.Dropdown(choices=["short", "medium", "long"], label="Caption Length", value="medium"),
-            gr.Textbox(label="Prompt", placeholder="Enter a prompt for the model")
-        ],
-        [],
-        [
-            gr.Files(label="Upload Multiple Images", file_types=["image"]),
-            gr.Dropdown(choices=["descriptive", "casual", "social media"], label="Caption Type", value="descriptive"),
-            gr.Dropdown(choices=["short", "medium", "long"], label="Caption Length", value="medium"),
-            gr.Textbox(label="Prompt", placeholder="Enter a prompt for the model")
-        ]
-    ],
-    outputs=[
-        gr.Textbox(label="Generated Caption"),
-        gr.Textbox(label="Caption History"),
-        gr.Textbox(label="Batch Caption Results")
-    ],
-    title="Image Captioning with GIT",
-    description="Upload an image or multiple images to generate captions using the Microsoft/git-large-coco model. Results are saved to captions.json."
-)
+# Create Gradio Blocks interface
+with gr.Blocks(title="Image Captioning with GIT") as demo:
+    gr.Markdown("# Image Captioning with GIT")
+    gr.Markdown("Upload an image or multiple images to generate captions using the Microsoft/git-large-coco model. Results are saved to captions.json.")
+
+    # Tab for single image captioning
+    with gr.Tab("Single Image Captioning"):
+        with gr.Row():
+            with gr.Column():
+                single_image_input = gr.Image(label="Upload Image", type="pil")
+                single_caption_type = gr.Dropdown(choices=["descriptive", "casual", "social media"], label="Caption Type", value="descriptive")
+                single_caption_length = gr.Dropdown(choices=["short", "medium", "long"], label="Caption Length", value="medium")
+                single_prompt = gr.Textbox(label="Prompt", placeholder="Enter a prompt for the model")
+                single_submit = gr.Button("Generate Caption")
+            single_output = gr.Textbox(label="Generated Caption")
+        single_submit.click(
+            fn=generate_caption,
+            inputs=[single_image_input, single_caption_type, single_caption_length, single_prompt],
+            outputs=single_output
+        )
+
+    # Tab for viewing caption history
+    with gr.Tab("Caption History"):
+        history_output = gr.Textbox(label="Caption History")
+        history_button = gr.Button("View History")
+        history_button.click(
+            fn=view_caption_history,
+            inputs=None,
+            outputs=history_output
+        )
+
+    # Tab for batch processing
+    with gr.Tab("Batch Image Captioning"):
+        with gr.Row():
+            with gr.Column():
+                batch_image_input = gr.Files(label="Upload Multiple Images", file_types=["image"])
+                batch_caption_type = gr.Dropdown(choices=["descriptive", "casual", "social media"], label="Caption Type", value="descriptive")
+                batch_caption_length = gr.Dropdown(choices=["short", "medium", "long"], label="Caption Length", value="medium")
+                batch_prompt = gr.Textbox(label="Prompt", placeholder="Enter a prompt for the model")
+                batch_submit = gr.Button("Generate Captions")
+            batch_output = gr.Textbox(label="Batch Caption Results")
+        batch_submit.click(
+            fn=batch_generate_captions,
+            inputs=[batch_image_input, batch_caption_type, batch_caption_length, batch_prompt],
+            outputs=batch_output
+        )
 
 if __name__ == "__main__":
-    interface.launch()  
+    demo.launch()