add contoh input gambar dan teks

app.py

@@ -9,27 +9,9 @@ import torch
 from peft import LoraConfig
 from transformers import AutoProcessor, BitsAndBytesConfig, IdeficsForVisionText2Text
 
-# Project description
-description = """
-# Kalbe Farma - Visual Question Answering (VQA) for Medical Imaging
-
-## Overview
-The project addresses the challenge of accurate and efficient medical imaging analysis in healthcare, aiming to reduce human error and workload for radiologists. The proposed solution involves developing advanced AI models for Visual Question Answering (VQA) to assist healthcare professionals in analyzing medical images quickly and accurately. These models will be integrated into a user-friendly web application, providing a practical tool for real-world healthcare settings.
-
-## Dataset
-The model is trained using the [Hugging face](https://huggingface.co/datasets/flaviagiammarino/vqa-rad/viewer).
-
-Reference: [ScienceDirect](https://www.sciencedirect.com/science/article/abs/pii/S0933365723001252)
-
-## Model Architecture
-
-![Model Architecture](img/Model-Architecture.png)
-
-Reference: [ScienceDirect](https://www.sciencedirect.com/science/article/abs/pii/S0933365723001252)
-
-## Demo
-Please select the example below or upload 4 pairs of mammography exam results.
-"""
+# Baca konten HTML dari file index.html
+with open('index.html', encoding='utf-8') as file:
+    html_content = file.read()
 
 DEVICE = torch.device("cuda")
 
@@ -51,14 +33,14 @@ if USE_QLORA or USE_LORA:
             bnb_4bit_quant_type="nf4",
             bnb_4bit_compute_dtype=torch.float16
         )
-    # Model
+
+    # Model yang akan digunakan
     model = Idefics2ForConditionalGeneration.from_pretrained(
-        "jihadzakki/idefics2-8b-vqarad-delta",
+        # "jihadzakki/idefics2-8b-vqarad-delta",
         torch_dtype=torch.float16,
         quantization_config=bnb_config
     )
 
-
 processor = AutoProcessor.from_pretrained(
     "HuggingFaceM4/idefics2-8b",
 )
@@ -135,9 +117,7 @@ with gr.Blocks(
         secondary_hue=gr.themes.colors.red,
     )
 ) as VisualQAApp:
-    gr.Markdown(description, elem_classes="title")  # Display the project description
-
-    gr.Markdown("## Demo")
+    gr.HTML(html_content)  # Display the HTML content
 
     with gr.Row():
         with gr.Column():
@@ -158,9 +138,9 @@ with gr.Blocks(
     )
 
     with gr.Row():
-        retry_button = gr.Button("Retry")
-        undo_button = gr.Button("Undo")
-        clear_button = gr.Button("Clear")
+        retry_button = gr.Button("🔄 Retry")
+        undo_button = gr.Button("↩️ Undo")
+        clear_button = gr.Button("🗑️ Clear")
 
         retry_button.click(
             retry_last,
@@ -188,9 +168,23 @@ with gr.Blocks(
             outputs=[history_gallery]
         )
 
+    gr.Markdown("## Contoh Input dengan Teks")
+    with gr.Row():
+        with gr.Column():
+            gr.Examples(
+                examples=[
+                    ["sample_data/images/Gambar-Otak-Slake.jpg", "What modality is used to take this image?"],
+                    ["sample_data/images/Gambar-Otak-Slake2.jpg", "Which part of the body does this image belong to?"]
+                ],
+                inputs=[image_input, question_input],
+                outputs=[answer_output, history_state],
+                label="Upload image",
+                elem_id="Prompt"
+            )
+
     with gr.Accordion("Help", open=False):
         gr.Markdown("**Upload image**: Select the chest X-ray image you want to analyze.")
-        gr.Markdown("**Enter your question**: Type the question you have about the image, such as 'Is there any sign of pneumonia?'")
+        gr.Markdown("**Enter your question**: Type the question you have about the image, such as 'What modality is used to take this image?'")
         gr.Markdown("**Submit**: Click the submit button to get the prediction from the model.")
 
     with gr.Accordion("User Preferences", open=False):