Update app.py

app.py

@@ -112,14 +112,15 @@ def search(search_query):
 
 
 
-with gr.Blocks(css=".caption-text {font-size: 40px !important;}") as demo:
+with gr.Blocks() as demo:
 
-    with gr.Tab("Zero-Shot Classification"):
+    with gr.Tab("Classification"):
         labels = gr.State([])   # creates hidden component that can store a value and can be used as input/output; here, initial value is an empty list
         instructions = """## Instructions:
         1. Enter list of labels separated by commas (or select one of the examples below)
-        2. Click **Get Random Image** to grab a random image from dataset and analyze it against the labels
-        3. Click **Re-Classify Image** to re-run classification on current image after changing labels"""
+        2. Click **Get Random Image** to grab a random image from dataset
+        3. Click **Classify Image** to analyze current image against the labels (including after changing labels)
+        4. The dataset (<a href="https://github.com/unsplash/datasets" target="_blank">Unsplash Lite</a>) contains 25,000 nature-focused images."""
         gr.Markdown(instructions)
         with gr.Row(variant="compact"):
             label_text = gr.Textbox(show_label=False, placeholder="Enter classification labels").style(container=False)
@@ -134,17 +135,23 @@ with gr.Blocks(css=".caption-text {font-size: 40px !important;}") as demo:
                 im = gr.Image(interactive=False).style(height=height)
                 with gr.Row():
                     get_btn = gr.Button("Get Random Image").style(full_width=False)
-                    reclass_btn = gr.Button("Re-Classify Image").style(full_width=False)
+                    class_btn = gr.Button("Classify Image").style(full_width=False)
             cf = gr.Label()
         #submit_btn.click(fn=set_labels, inputs=label_text)
         label_text.change(fn=set_labels, inputs=label_text, outputs=labels)  # parse list if changed
         label_text.blur(fn=set_labels, inputs=label_text, outputs=labels)  # parse list if focus is moved elsewhere; ensures that list is fully parsed before classification
         label_text.submit(fn=set_labels, inputs=label_text, outputs=labels)  # parse list if user hits enter; ensures that list is fully parsed before classification
         get_btn.click(fn=rand_image, outputs=im)
-        im.change(predict, inputs=[im, labels], outputs=cf)
-        reclass_btn.click(predict, inputs=[im, labels], outputs=cf)
+        #im.change(predict, inputs=[im, labels], outputs=cf)
+        class_btn.click(predict, inputs=[im, labels], outputs=cf)
 
-    with gr.Tab("Image Captioning"):
+    with gr.Tab("Captioning"):
+        instructions = """## Instructions:
+        1. Click **Get Random Image** to grab a random image from dataset
+        1. Click **Create Caption** to generate a caption for the image
+        1. Different models can be selected:  *COCO* generally produces more straight-forward captions, but it is a smaller dataset and therefore struggles to recognize certain objects; **Conceptual Captions** is a much larger dataset but often generally overly...um...poetic results
+        1. The dataset (<a href="https://github.com/unsplash/datasets" target="_blank">Unsplash Lite</a>) contains 25,000 nature-focused images."""
+        gr.Markdown(instructions)
         with gr.Row():
             with gr.Column(variant="panel"):
                 im_cap = gr.Image(interactive=False, type='filepath').style(height=height)
@@ -157,9 +164,20 @@ with gr.Blocks(css=".caption-text {font-size: 40px !important;}") as demo:
         #im_cap.change(generate_text, inputs=im_cap, outputs=caption)
         caption_btn.click(generate_text, inputs=[im_cap, model_name], outputs=caption)
 
-    with gr.Tab("Image Search"):
+    with gr.Tab("Search"):
+        instructions = """## Instructions:
+        1. Enter a search query (or select one of the examples below)
+        2. Click **Find Images** to find images that match the query (top 5 are shown in order from left to right)
+        3. The dataset (<a href="https://github.com/unsplash/datasets" target="_blank">Unsplash Lite</a>) contains 25,000 nature-focused images."""
+        gr.Markdown(instructions)
         with gr.Column(variant="panel"):
             desc = gr.Textbox(show_label=False, placeholder="Enter description").style(container=False)
+            gr.Examples(["someone holding flowers",
+                         "someone holding blue flowers",
+                         "red fruit in a person's hands",
+                         "an aerial view of forest",
+                         "a waterfall with a rainbow"
+                        ], inputs=desc)
             search_btn = gr.Button("Find Images").style(full_width=False)
         gallery = gr.Gallery(show_label=False).style(grid=(2,2,3,5))
         search_btn.click(search,inputs=desc, outputs=gallery, postprocess=False)