Update app.py

app.py

@@ -1,201 +1,56 @@
 import gradio as gr
-import numpy as np
-import random
-
-import spaces
-from diffusers import DiffusionPipeline, FlowMatchEulerDiscreteScheduler
 import torch
+from PIL import Image
+from transformers import BlipProcessor, BlipForConditionalGeneration
 
+# 1. 장치 설정
 device = "cuda" if torch.cuda.is_available() else "cpu"
-model_repo_id = "tensorart/stable-diffusion-3.5-large-TurboX"
-
-if torch.cuda.is_available():
-    torch_dtype = torch.float16
-else:
-    torch_dtype = torch.float32
-
-pipe = DiffusionPipeline.from_pretrained(model_repo_id, torch_dtype=torch_dtype)
-
-pipe.scheduler = FlowMatchEulerDiscreteScheduler.from_pretrained(model_repo_id, subfolder="scheduler", shift=5)
-
-pipe = pipe.to(device)
-
-MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 1024
-
-@spaces.GPU(duration=65)
-def infer(
-    prompt,
-    negative_prompt="",
-    seed=42,
-    randomize_seed=False,
-    width=1024,
-    height=1024,
-    guidance_scale=1.5,
-    num_inference_steps=8,
-    progress=gr.Progress(track_tqdm=True),
-):
-    if randomize_seed:
-        seed = random.randint(0, MAX_SEED)
-
-    generator = torch.Generator().manual_seed(seed)
-
-    image = pipe(
-        prompt=prompt,
-        negative_prompt=negative_prompt,
-        guidance_scale=guidance_scale,
-        num_inference_steps=num_inference_steps,
-        width=width,
-        height=height,
-        generator=generator,
-    ).images[0]
-
-    return image, seed
-
-
-examples = [
-        "A capybara wearing a suit holding a sign that reads Hello World",
-        "A serene mountain lake at sunset with cherry blossoms floating on the water",
-        "A magical crystal dragon with iridescent scales in a glowing forest",
-        "A Victorian steampunk teapot with intricate brass gears and rose gold accents",
-        "A futuristic neon cityscape with flying cars and holographic billboards",
-        "A red panda painter creating a masterpiece with tiny paws in an art studio",
-]
-
-css = """
-body {
-    background: linear-gradient(135deg, #f9e2e6 0%, #e8f3fc 50%, #e2f9f2 100%);
-    background-attachment: fixed;
-    min-height: 100vh;
-}
-
-#col-container {
-    margin: 0 auto;
-    max-width: 640px;
-    background-color: rgba(255, 255, 255, 0.85);
-    border-radius: 16px;
-    box-shadow: 0 8px 16px rgba(0, 0, 0, 0.1);
-    padding: 24px;
-    backdrop-filter: blur(10px);
-}
-
-.gradio-container {
-    background: transparent !important;
-}
-
-.gr-button-primary {
-    background: linear-gradient(90deg, #6b9dfc, #8c6bfc) !important;
-    border: none !important;
-    transition: all 0.3s ease;
-}
-
-.gr-button-primary:hover {
-    transform: translateY(-2px);
-    box-shadow: 0 5px 15px rgba(108, 99, 255, 0.3);
-}
-
-.gr-form {
-    border-radius: 12px;
-    background-color: rgba(255, 255, 255, 0.7);
-}
-
-.gr-accordion {
-    border-radius: 12px;
-    overflow: hidden;
-}
-
-h1 {
-    background: linear-gradient(90deg, #6b9dfc, #8c6bfc);
-    -webkit-background-clip: text;
-    -webkit-text-fill-color: transparent;
-    font-weight: 800;
-}
-"""
-
-with gr.Blocks(theme="apriel", css=css) as demo:
-    with gr.Column(elem_id="col-container"):
-        gr.Markdown(" # TensorArt Stable Diffusion 3.5 Large TurboX")
-        gr.Markdown("[8-step distilled turbo model](https://huggingface.co/tensorart/stable-diffusion-3.5-large-TurboX)")
-        with gr.Row():
-            prompt = gr.Text(
-                label="Prompt",
-                show_label=False,
-                max_lines=1,
-                placeholder="Enter your prompt",
-                container=False,
-            )
-
-            run_button = gr.Button("Run", scale=0, variant="primary")
-
-        result = gr.Image(label="Result", show_label=False)
-
-        with gr.Accordion("Advanced Settings", open=False):
-            negative_prompt = gr.Text(
-                label="Negative prompt",
-                max_lines=1,
-                placeholder="Enter a negative prompt",
-            )
-
-            seed = gr.Slider(
-                label="Seed",
-                minimum=0,
-                maximum=MAX_SEED,
-                step=1,
-                value=0,
-            )
-
-            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
-
-            with gr.Row():
-                width = gr.Slider(
-                    label="Width",
-                    minimum=512,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=1024, 
-                )
-
-                height = gr.Slider(
-                    label="Height",
-                    minimum=512,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=1024,
-                )
-
-            with gr.Row():
-                guidance_scale = gr.Slider(
-                    label="Guidance scale",
-                    minimum=0.0,
-                    maximum=7.5,
-                    step=0.1,
-                    value=1.5,
-                )
-
-                num_inference_steps = gr.Slider(
-                    label="Number of inference steps",
-                    minimum=1,
-                    maximum=50,
-                    step=1,
-                    value=8, 
-                )
-
-        gr.Examples(examples=examples, inputs=[prompt], outputs=[result, seed], fn=infer, cache_examples=True, cache_mode="lazy")
-    gr.on(
-        triggers=[run_button.click, prompt.submit],
-        fn=infer,
-        inputs=[
-            prompt,
-            negative_prompt,
-            seed,
-            randomize_seed,
-            width,
-            height,
-            guidance_scale,
-            num_inference_steps,
-        ],
-        outputs=[result, seed],
-    )
 
-if __name__ == "__main__":
-    demo.launch(mcp_server=True)
+# 2. 모델 및 프로세서 로드
+processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
+model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base").to(device)
+
+# 3. 이미지 설명 생성 함수
+def generate_caption(image):
+    if image is None:
+        return "이미지를 업로드해주세요."
+    
+    # 고속 처리를 위한 리사이즈
+    image = image.resize((384, 384))
+    
+    # 설명 생성
+    inputs = processor(images=image, return_tensors="pt").to(device)
+    output_ids = model.generate(**inputs, max_length=50)
+    caption = processor.decode(output_ids[0], skip_special_tokens=True)
+    print("✅ 생성된 설명:", caption)
+    if "Asian" in caption:
+        caption = caption.replace("Asian", "Korean")
+    print("✅ 생성된 설명:", caption)
+    return caption
+    return caption
+
+# 4. Gradio 인터페이스 구성
+with gr.Blocks(title="이미지 설명 생성기") as demo:
+    gr.Markdown("## 🖼️ 이미지를 업로드하면 설명이 자동 생성됩니다.")
+
+    with gr.Row():
+        with gr.Column():
+            image_input = gr.Image(label="입력 이미지", type="pil")
+        with gr.Column():
+            caption_output = gr.Textbox(label="생성된 설명", lines=3, show_copy_button=True)
+            # HTML로 버튼 생성
+            gr.HTML("""
+            <div style='margin-top: 10px; text-align: center;'>
+                <a href="https://huggingface.co/spaces/VIDraft/stable-diffusion-3.5-large-turboX" target="_blank">
+                    <button style='padding: 10px 20px; background-color: #ff9900; color: white; border: none; border-radius: 10px; font-size: 16px; box-shadow: 2px 2px 8px rgba(0,0,0,0.3); cursor: pointer;'>
+                        🎨 캐리커쳐 만들기
+                    </button>
+                </a>
+            </div>
+            """)
+
+    # 업로드 → 설명 자동 생성 연결
+    image_input.upload(fn=generate_caption, inputs=image_input, outputs=caption_output)
+
+# 5. 앱 실행
+demo.launch(debug=True)