Added more examples

app.py

@@ -156,15 +156,43 @@ model.eval()
 
 
 # Gradio
-examples = [
+examples_in_domain = [
     [
         "./imgs/examples/Animalia_Chordata_Aves_Charadriiformes_Laridae_Larus_marinus/80645_39.76079_-74.10316.jpg",
-        "Where can I find the shore birds (Animalia Chordata Aves Charadriiformes Laridae Larus marinus) in this image? Please output segmentation mask and explain why.",
+        "Where can I find the shore birds (Animalia Chordata Aves Charadriiformes Laridae Larus marinus) in this image? Please output segmentation mask.",
     ],    
     [
         "./imgs/examples/Animalia_Chordata_Mammalia_Rodentia_Caviidae_Hydrochoerus_hydrochaeris/28871_-12.80255_-69.29999.jpg",
-        "Where can I find the capybaras (Animalia Chordata Mammalia Rodentia Caviidae Hydrochoerus hydrochaeris) in this image? Please output segmentation mask.",
+        "Where can I find the capybaras (Animalia Chordata Mammalia Rodentia Caviidae Hydrochoerus hydrochaeris) in this image? Please output segmentation mask and explain why.",
     ],  
+    [
+        "./imgs/examples/Animalia_Arthropoda_Malacostraca_Decapoda_Ocypodidae_Ocypode_quadrata/277303_38.72364_-75.07749.jpg",
+        "Where can I find the crabs (Animalia Arthropoda Malacostraca Decapoda Ocypodidae Ocypode quadrata) in this image? Please output segmentation mask.",
+    ],
+    [
+        "./imgs/examples/Animalia_Chordata_Mammalia_Rodentia_Sciuridae_Marmota_marmota/388246_45.49036_7.14796.jpg",
+        "Where can I find the marmots (Animalia Chordata Mammalia Rodentia Sciuridae Marmota marmota) in this image? Please output segmentation mask and explain why.",
+    ],
+    [
+        "./imgs/examples/Animalia_Chordata_Reptilia_Squamata_Varanidae_Varanus_salvator/410613_5.35573_100.28948.jpg",
+        "Where can I find monitor lizard (Animalia Chordata Reptilia Squamata Varanidae Varanus salvator) in this image? Please output segmentation mask.",
+    ],
+
+]
+
+examples_out_domain = [
+    [
+        "./imgs/examples/Animalia_Chordata_Mammalia_Carnivora_Phocidae_Mirounga_angustirostris/27423_35.64005_-121.17595.jpg",
+        "Where can I find the seals (Animalia Chordata Mammalia Carnivora Phocidae Mirounga angustirostris) in this image? Please output segmentation mask and explain why.",
+    ],
+    [
+        "./imgs/examples/Animalia_Chordata_Mammalia_Carnivora_Canidae_Canis_aureus/1528408_13.00422_80.23033.jpg",
+        "Where can I find the wolves (Animalia Chordata Mammalia Carnivora Canidae Canis aureus) in this image? Please output segmentation mask.",
+    ],
+    [
+        "./imgs/examples/Animalia_Chordata_Mammalia_Carnivora_Canidae_Urocyon_littoralis/304160_34.0144_-119.54417.jpg",
+        "Where can I find the raccoons (Animalia Chordata Mammalia Carnivora Canidae Urocyon littoralis) in this image? Please output segmentation mask and explain why.",
+    ],
 ]
 output_labels = ["Segmentation Output"]
 
@@ -173,9 +201,7 @@ title = "LISA-AVS: LISA 7B Model Finetuned on AVS-Bench Dataset"
 description = """
 <font size=4>
 This is an adapted version of the online demo for <a href='https://github.com/dvlab-research/LISA' target='_blank'>LISA</a>, where we finetune from scratch the LISA model (7B) with data from <a href='https://search-tta.github.io/' target='_blank'>AVS-Bench (Search-TTA)</a>. \n
-**Note**:  <br>
-&ensp;(a) If multiple users are using it at the same time, they will enter a queue, which may delay some time. \n
-&ensp;(b) Different prompts can lead to significantly varied results. Please **standardize** your input text prompts to **avoid ambiguity**, and pay attention to whether the **punctuations** of the input are correct. \n
+**Note**: Different prompts can lead to significantly varied results. Please **standardize** your input text prompts to **avoid ambiguity**, and pay attention to whether the **punctuations** of the input are correct. \n
 **Usage**: <br>
 &ensp;(1) To let LISA-AVS **segment something**, input prompt like: "Where can I find the <em>Common Name</em> (<em>Full Taxonomy Name</em>) in this image? Please output segmentation mask."; <br>
 &ensp;(2) To let LISA-AVS **output an explanation**, input prompt like: "Where can I find the <em>Common Name</em> (<em>Full Taxonomy Name</em>) in this image? Please output segmentation mask and explain why."; <br>
@@ -335,22 +361,43 @@ def inference(input_image, input_str):
     return output_image, output_str
 
 
-demo = gr.Interface(
-    inference,
-    inputs=[
-        gr.Image(type="filepath", label="Input Image"),
-        gr.Textbox(lines=1, placeholder=None, label="Text Instruction"),
-    ],
-    outputs=[
-        gr.Image(type="pil", label="Segmentation Output"),
-        gr.Textbox(lines=1, placeholder=None, label="Text Output"),
-    ],
-    title=title,
-    description=description,
-    article=article,
-    examples=examples,
-    allow_flagging="auto",
-)
+with gr.Blocks() as demo:
+    # Title, description and article
+    gr.Markdown(f"# {title}")
+    gr.Markdown(description)
+    # gr.Markdown(articSle)
+
+    # Input and output columns side-by-side
+    with gr.Row():
+        with gr.Column():  # Left column – inputs
+            inp_image = gr.Image(type="filepath", label="Input Image")
+            inp_text = gr.Textbox(lines=1, placeholder=None, label="Text Instruction")
+            run_btn = gr.Button("Run", variant="primary")
+        with gr.Column():  # Right column – outputs
+            out_seg = gr.Image(type="pil", label="Segmentation Output")
+            out_text = gr.Textbox(lines=1, label="Text Output")
+
+    # Bind the button to inference
+    run_btn.click(fn=inference, inputs=[inp_image, inp_text], outputs=[out_seg, out_text])
+
+    # ---------------- Example Galleries ----------------
+    gr.Markdown("### In-Domain Taxonomy")
+    gr.Examples(
+        examples=examples_in_domain,
+        inputs=[inp_image, inp_text],
+        outputs=[out_seg, out_text],
+        fn=inference,
+        cache_examples=False,
+    )
+
+    gr.Markdown("### Out-Domain Taxonomy")
+    gr.Examples(
+        examples=examples_out_domain,
+        inputs=[inp_image, inp_text],
+        outputs=[out_seg, out_text],
+        fn=inference,
+        cache_examples=False,
+    )
 
 demo.queue()
 demo.launch()