First commit

app.py

@@ -0,0 +1,96 @@
+import gradio as gr
+import torch
+import numpy as np
+from diffusers import StableDiffusionXLImg2ImgPipeline
+from transformers import DPTFeatureExtractor, DPTForDepthEstimation
+from PIL import Image, ImageEnhance, ImageOps
+
+device = "cpu"  # or "cuda" if you have a GPU
+torch_dtype = torch.float32
+
+print("Loading SDXL Img2Img model...")
+pipe = StableDiffusionXLImg2ImgPipeline.from_pretrained(
+    "stabilityai/stable-diffusion-xl-base-1.0",
+    torch_dtype=torch_dtype
+).to(device)
+
+print("Loading bas-relief LoRA weights with PEFT...")
+pipe.load_lora_weights(
+    "KappaNeuro/bas-relief",
+    weight_name="BAS-RELIEF.safetensors",
+    peft_backend="peft"
+)
+
+print("Loading DPT Depth Model...")
+feature_extractor = DPTFeatureExtractor.from_pretrained("Intel/dpt-large")
+depth_model = DPTForDepthEstimation.from_pretrained("Intel/dpt-large").to(device)
+
+
+def enhance_depth_map(depth_arr: np.ndarray) -> Image.Image:
+    d_min, d_max = depth_arr.min(), depth_arr.max()
+    depth_stretched = (depth_arr - d_min) / (d_max - d_min + 1e-8)
+    depth_stretched = (depth_stretched * 255).astype(np.uint8)
+
+    depth_pil = Image.fromarray(depth_stretched)
+    depth_pil = ImageOps.autocontrast(depth_pil)
+
+    enhancer = ImageEnhance.Sharpness(depth_pil)
+    depth_pil = enhancer.enhance(2.0)
+
+    return depth_pil
+
+
+def generate_bas_relief_and_depth(input_image: Image.Image):
+    # Redimensionar a imagem para o tamanho esperado
+    input_image = input_image.resize((512, 512))
+
+    # Prompt fixo para ativar o LoRA
+    prompt = "BAS-RELIEF"
+
+    print("Gerando imagem no estilo baixo-relevo...")
+    result = pipe(
+        prompt=prompt,
+        image=input_image,
+        strength=0.7,  # Controla a intensidade da transformação
+        num_inference_steps=15,
+        guidance_scale=7.5
+    )
+    generated_image = result.images[0]
+
+    print("Calculando mapa de profundidade...")
+    inputs = feature_extractor(generated_image, return_tensors="pt").to(device)
+    with torch.no_grad():
+        outputs = depth_model(**inputs)
+        predicted_depth = outputs.predicted_depth
+
+    prediction = torch.nn.functional.interpolate(
+        predicted_depth.unsqueeze(1),
+        size=generated_image.size[::-1],
+        mode="bicubic",
+        align_corners=False
+    ).squeeze()
+
+    depth_map_pil = enhance_depth_map(prediction.cpu().numpy())
+
+    return generated_image, depth_map_pil
+
+
+title = "Conversor para Baixo-relevo (SDXL + LoRA) com Mapa de Profundidade"
+description = (
+    "Carrega stable-diffusion-xl-base-1.0 no CPU, aplica LoRA de 'KappaNeuro/bas-relief' "
+    "para transformar imagens em baixo-relevo e calcula o mapa de profundidade correspondente."
+)
+
+iface = gr.Interface(
+    fn=generate_bas_relief_and_depth,
+    inputs=gr.Image(label="Imagem de Entrada", type="pil"),
+    outputs=[
+        gr.Image(label="Imagem em Baixo-relevo"),
+        gr.Image(label="Mapa de Profundidade")
+    ],
+    title=title,
+    description=description
+)
+
+if __name__ == "__main__":
+    iface.launch()

requirements.txt

@@ -0,0 +1,9 @@
+open3d
+peft
+accelerate
+diffusers>=0.20.0
+transformers>=4.30.0
+torch
+gradio
+Pillow
+safetensors