import spaces import os import gradio as gr import numpy as np import torch from PIL import Image import trimesh import random from transformers import AutoModelForImageSegmentation from torchvision import transforms from huggingface_hub import hf_hub_download, snapshot_download import subprocess import shutil import base64 import logging import requests # Set up logging logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s') logger = logging.getLogger(__name__) # Install additional dependencies try: subprocess.run("pip install spandrel==0.4.1 --no-deps", shell=True, check=True) except Exception as e: logger.error(f"Failed to install spandrel: {str(e)}") raise DEVICE = "cuda" if torch.cuda.is_available() else "cpu" DTYPE = torch.float16 logger.info(f"Using device: {DEVICE}") DEFAULT_FACE_NUMBER = 100000 MAX_SEED = np.iinfo(np.int32).max TRIPOSG_REPO_URL = "https://github.com/VAST-AI-Research/TripoSG.git" MV_ADAPTER_REPO_URL = "https://github.com/huanngzh/MV-Adapter.git" RMBG_PRETRAINED_MODEL = "checkpoints/RMBG-1.4" TRIPOSG_PRETRAINED_MODEL = "checkpoints/TripoSG" TMP_DIR = os.path.join(os.path.dirname(os.path.abspath(__file__)), "tmp") os.makedirs(TMP_DIR, exist_ok=True) TRIPOSG_CODE_DIR = "./triposg" if not os.path.exists(TRIPOSG_CODE_DIR): logger.info(f"Cloning TripoSG repository to {TRIPOSG_CODE_DIR}") os.system(f"git clone {TRIPOSG_REPO_URL} {TRIPOSG_CODE_DIR}") MV_ADAPTER_CODE_DIR = "./mv_adapter" if not os.path.exists(MV_ADAPTER_CODE_DIR): logger.info(f"Cloning MV-Adapter repository to {MV_ADAPTER_CODE_DIR}") os.system(f"git clone {MV_ADAPTER_REPO_URL} {MV_ADAPTER_CODE_DIR} && cd {MV_ADAPTER_CODE_DIR} && git checkout 7d37a97e9bc223cdb8fd26a76bd8dd46504c7c3d") import sys sys.path.append(TRIPOSG_CODE_DIR) sys.path.append(os.path.join(TRIPOSG_CODE_DIR, "scripts")) sys.path.append(MV_ADAPTER_CODE_DIR) sys.path.append(os.path.join(MV_ADAPTER_CODE_DIR, "scripts")) try: from image_process import prepare_image from briarmbg import BriaRMBG snapshot_download("briaai/RMBG-1.4", local_dir=RMBG_PRETRAINED_MODEL) rmbg_net = BriaRMBG.from_pretrained(RMBG_PRETRAINED_MODEL).to(DEVICE) rmbg_net.eval() from triposg.pipelines.pipeline_triposg import TripoSGPipeline snapshot_download("VAST-AI/TripoSG", local_dir=TRIPOSG_PRETRAINED_MODEL) triposg_pipe = TripoSGPipeline.from_pretrained(TRIPOSG_PRETRAINED_MODEL).to(DEVICE, DTYPE) except Exception as e: logger.error(f"Failed to load TripoSG models: {str(e)}") raise try: NUM_VIEWS = 6 from inference_ig2mv_sdxl import prepare_pipeline, preprocess_image, remove_bg from mvadapter.utils import get_orthogonal_camera, tensor_to_image, make_image_grid from mvadapter.utils.render import NVDiffRastContextWrapper, load_mesh, render mv_adapter_pipe = prepare_pipeline( base_model="stabilityai/stable-diffusion-xl-base-1.0", vae_model="madebyollin/sdxl-vae-fp16-fix", unet_model=None, lora_model=None, adapter_path="huanngzh/mv-adapter", scheduler=None, num_views=NUM_VIEWS, device=DEVICE, dtype=torch.float16, ) birefnet = AutoModelForImageSegmentation.from_pretrained( "ZhengPeng7/BiRefNet", trust_remote_code=True ).to(DEVICE) transform_image = transforms.Compose( [ transforms.Resize((1024, 1024)), transforms.ToTensor(), transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]), ] ) remove_bg_fn = lambda x: remove_bg(x, birefnet, transform_image, DEVICE) except Exception as e: logger.error(f"Failed to load MV-Adapter models: {str(e)}") raise try: if not os.path.exists("checkpoints/RealESRGAN_x2plus.pth"): hf_hub_download("dtarnow/UPscaler", filename="RealESRGAN_x2plus.pth", local_dir="checkpoints") if not os.path.exists("checkpoints/big-lama.pt"): subprocess.run("wget -P checkpoints/ https://github.com/Sanster/models/releases/download/add_big_lama/big-lama.pt", shell=True, check=True) except Exception as e: logger.error(f"Failed to download checkpoints: {str(e)}") raise def get_random_hex(): random_bytes = os.urandom(8) random_hex = random_bytes.hex() return random_hex @spaces.GPU(duration=5) def run_full(image: str, seed: int = 0, num_inference_steps: int = 50, guidance_scale: float = 7.5, simplify: bool = True, target_face_num: int = DEFAULT_FACE_NUMBER, req=None): try: image_seg = prepare_image(image, bg_color=np.array([1.0, 1.0, 1.0]), rmbg_net=rmbg_net) outputs = triposg_pipe( image=image_seg, generator=torch.Generator(device=triposg_pipe.device).manual_seed(seed), num_inference_steps=num_inference_steps, guidance_scale=guidance_scale ).samples[0] logger.info("Mesh extraction done") mesh = trimesh.Trimesh(outputs[0].astype(np.float32), np.ascontiguousarray(outputs[1])) if simplify: logger.info("Starting mesh simplification") from utils import simplify_mesh mesh = simplify_mesh(mesh, target_face_num) save_dir = os.path.join(TMP_DIR, "examples") os.makedirs(save_dir, exist_ok=True) mesh_path = os.path.join(save_dir, f"polygenixai_{get_random_hex()}.glb") mesh.export(mesh_path) logger.info(f"Saved mesh to {mesh_path}") torch.cuda.empty_cache() height, width = 1920, 1080 # Set resolution for YouTube Shorts, TikTok, Reels cameras = get_orthogonal_camera( elevation_deg=[0, 0, 0, 0, 89.99, -89.99], distance=[1.8] * NUM_VIEWS, left=-0.55, right=0.55, bottom=-0.55, top=0.55, azimuth_deg=[x - 90 for x in [0, 90, 180, 270, 180, 180]], device=DEVICE, ) ctx = NVDiffRastContextWrapper(device=DEVICE, context_type="cuda") mesh = load_mesh(mesh_path, rescale=True, device=DEVICE) render_out = render( ctx, mesh, cameras, height=height, width=width, render_attr=False, normal_background=0.0, ) control_images = ( (render_out.pos + 0.5).clamp(0, 1) # Use only position map, remove normal map .permute(0, 3, 1, 2) .to(DEVICE) ) image = Image.open(image) image = remove_bg_fn(image) image = preprocess_image(image, height, width) pipe_kwargs = {} if seed != -1 and isinstance(seed, int): pipe_kwargs["generator"] = torch.Generator(device=DEVICE).manual_seed(seed) images = mv_adapter_pipe( "high quality", height=height, width=width, num_inference_steps=15, guidance_scale=3.0, num_images_per_prompt=NUM_VIEWS, control_image=control_images, control_conditioning_scale=1.0, reference_image=image, reference_conditioning_scale=1.0, negative_prompt="watermark, ugly, deformed, noisy, blurry, low contrast", cross_attention_kwargs={"scale": 1.0}, **pipe_kwargs, ).images torch.cuda.empty_cache() os.makedirs(save_dir, exist_ok=True) mv_image_path = os.path.join(save_dir, f"mv_adapter_{get_random_hex()}.png") make_image_grid(images, rows=1).save(mv_image_path) from texture import TexturePipeline, ModProcessConfig texture_pipe = TexturePipeline( upscaler_ckpt_path="checkpoints/RealESRGAN_x2plus.pth", inpaint_ckpt_path="checkpoints/big-lama.pt", device=DEVICE, ) textured_glb_path = texture_pipe( mesh_path=mesh_path, save_dir=save_dir, save_name=f"polygenixai_texture_mesh_{get_random_hex()}.glb", uv_unwarp=True, uv_size=4096, rgb_path=mv_image_path, rgb_process_config=ModProcessConfig(view_upscale=True, inpaint_mode="view"), camera_azimuth_deg=[x - 90 for x in [0, 90, 180, 270, 180, 180]], ) return image_seg, mesh_path, textured_glb_path except Exception as e: logger.error(f"Error in run_full: {str(e)}") raise def gradio_generate(image: str, seed: int = 0, num_inference_steps: int = 50, guidance_scale: float = 7.5, simplify: bool = True, target_face_num: int = DEFAULT_FACE_NUMBER): try: logger.info("Starting gradio_generate") api_key = os.getenv("POLYGENIX_API_KEY", "your-secret-api-key") request = gr.Request() if not request.headers.get("x-api-key") == api_key: logger.error("Invalid API key") raise ValueError("Invalid API key") if image.startswith("data:image"): logger.info("Processing base64 image") base64_string = image.split(",")[1] image_data = base64.b64decode(base64_string) temp_image_path = os.path.join(TMP_DIR, f"input_{get_random_hex()}.png") with open(temp_image_path, "wb") as f: f.write(image_data) else: temp_image_path = image if not os.path.exists(temp_image_path): logger.error(f"Image file not found: {temp_image_path}") raise ValueError("Invalid or missing image file") image_seg, mesh_path, textured_glb_path = run_full(temp_image_path, seed, num_inference_steps, guidance_scale, simplify, target_face_num, req=None) session_hash = os.path.basename(os.path.dirname(textured_glb_path)) logger.info(f"Generated model at /files/{session_hash}/{os.path.basename(textured_glb_path)}") return {"file_url": f"/files/{session_hash}/{os.path.basename(textured_glb_path)}"} except Exception as e: logger.error(f"Error in gradio_generate: {str(e)}") raise def start_session(req: gr.Request): try: save_dir = os.path.join(TMP_DIR, str(req.session_hash)) os.makedirs(save_dir, exist_ok=True) logger.info(f"Started session, created directory: {save_dir}") except Exception as e: logger.error(f"Error in start_session: {str(e)}") raise def end_session(req: gr.Request): try: save_dir = os.path.join(TMP_DIR, str(req.session_hash)) shutil.rmtree(save_dir) logger.info(f"Ended session, removed directory: {save_dir}") except Exception as e: logger.error(f"Error in end_session: {str(e)}") raise def get_random_seed(randomize_seed, seed): try: if randomize_seed: seed = random.randint(0, MAX_SEED) logger.info(f"Generated seed: {seed}") return seed except Exception as e: logger.error(f"Error in get_random_seed: {str(e)}") raise def download_image(url: str, save_path: str) -> str: try: logger.info(f"Downloading image from {url}") response = requests.get(url, stream=True) response.raise_for_status() with open(save_path, "wb") as f: for chunk in response.iter_content(chunk_size=8192): f.write(chunk) logger.info(f"Saved image to {save_path}") return save_path except Exception as e: logger.error(f"Failed to download image from {url}: {str(e)}") raise @spaces.GPU() @torch.no_grad() def run_segmentation(image): try: logger.info("Running segmentation") if isinstance(image, dict): image_path = image.get("path") or image.get("url") if not image_path: logger.error("Invalid image input: no path or URL provided") raise ValueError("Invalid image input: no path or URL provided") if image_path.startswith("http"): temp_image_path = os.path.join(TMP_DIR, f"input_{get_random_hex()}.png") image_path = download_image(image_path, temp_image_path) elif isinstance(image, str) and image.startswith("http"): temp_image_path = os.path.join(TMP_DIR, f"input_{get_random_hex()}.png") image_path = download_image(image, temp_image_path) else: image_path = image if not isinstance(image, (str, bytes)) or (isinstance(image, str) and not os.path.exists(image)): logger.error(f"Invalid image type or path: {type(image)}") raise ValueError(f"Expected str (path/URL), bytes, or FileData dict, got {type(image)}") image = prepare_image(image_path, bg_color=np.array([1.0, 1.0, 1.0]), rmbg_net=rmbg_net) logger.info("Segmentation complete") return image except Exception as e: logger.error(f"Error in run_segmentation: {str(e)}") raise @spaces.GPU(duration=5) @torch.no_grad() def image_to_3d( image, seed: int, num_inference_steps: int, guidance_scale: float, simplify: bool, target_face_num: int, req: gr.Request ): try: logger.info("Running image_to_3d") if isinstance(image, dict): image_path = image.get("path") or image.get("url") if not image_path: logger.error("Invalid image input: no path or URL provided") raise ValueError("Invalid image input: no path or URL provided") image = Image.open(image_path) elif not isinstance(image, Image.Image): logger.error(f"Invalid image type: {type(image)}") raise ValueError(f"Expected PIL Image or FileData dict, got {type(image)}") outputs = triposg_pipe( image=image, generator=torch.Generator(device=triposg_pipe.device).manual_seed(seed), num_inference_steps=num_inference_steps, guidance_scale=guidance_scale ).samples[0] logger.info("Mesh extraction done") mesh = trimesh.Trimesh(outputs[0].astype(np.float32), np.ascontiguousarray(outputs[1])) if simplify: logger.info("Starting mesh simplification") try: from utils import simplify_mesh mesh = simplify_mesh(mesh, target_face_num) except ImportError as e: logger.error(f"Failed to import simplify_mesh: {str(e)}") raise save_dir = os.path.join(TMP_DIR, str(req.session_hash)) os.makedirs(save_dir, exist_ok=True) mesh_path = os.path.join(save_dir, f"polygenixai_{get_random_hex()}.glb") mesh.export(mesh_path) logger.info(f"Saved mesh to {mesh_path}") torch.cuda.empty_cache() return mesh_path except Exception as e: logger.error(f"Error in image_to_3d: {str(e)}") raise @spaces.GPU(duration=5) @torch.no_grad() def run_texture(image: Image, mesh_path: str, seed: int, req: gr.Request): try: logger.info("Running texture generation") height, width = 1920, 1080 # Set resolution for YouTube Shorts, TikTok, Reels cameras = get_orthogonal_camera( elevation_deg=[0, 0, 0, 0, 89.99, -89.99], distance=[1.8] * NUM_VIEWS, left=-0.55, right=0.55, bottom=-0.55, top=0.55, azimuth_deg=[x - 90 for x in [0, 90, 180, 270, 180, 180]], device=DEVICE, ) ctx = NVDiffRastContextWrapper(device=DEVICE, context_type="cuda") mesh = load_mesh(mesh_path, rescale=True, device=DEVICE) render_out = render( ctx, mesh, cameras, height=height, width=width, render_attr=False, normal_background=0.0, ) control_images = ( (render_out.pos + 0.5).clamp(0, 1) # Use only position map, remove normal map .permute(0, 3, 1, 2) .to(DEVICE) ) image = Image.open(image) image = remove_bg_fn(image) image = preprocess_image(image, height, width) pipe_kwargs = {} if seed != -1 and isinstance(seed, int): pipe_kwargs["generator"] = torch.Generator(device=DEVICE).manual_seed(seed) images = mv_adapter_pipe( "high quality", height=height, width=width, num_inference_steps=15, guidance_scale=3.0, num_images_per_prompt=NUM_VIEWS, control_image=control_images, control_conditioning_scale=1.0, reference_image=image, reference_conditioning_scale=1.0, negative_prompt="watermark, ugly, deformed, noisy, blurry, low contrast", cross_attention_kwargs={"scale": 1.0}, **pipe_kwargs, ).images torch.cuda.empty_cache() save_dir = os.path.join(TMP_DIR, str(req.session_hash)) os.makedirs(save_dir, exist_ok=True) mv_image_path = os.path.join(save_dir, f"mv_adapter_{get_random_hex()}.png") make_image_grid(images, rows=1).save(mv_image_path) from texture import TexturePipeline, ModProcessConfig texture_pipe = TexturePipeline( upscaler_ckpt_path="checkpoints/RealESRGAN_x2plus.pth", inpaint_ckpt_path="checkpoints/big-lama.pt", device=DEVICE, ) textured_glb_path = texture_pipe( mesh_path=mesh_path, save_dir=save_dir, save_name=f"polygenixai_texture_mesh_{get_random_hex()}.glb", uv_unwarp=True, uv_size=4096, rgb_path=mv_image_path, rgb_process_config=ModProcessConfig(view_upscale=True, inpaint_mode="view"), camera_azimuth_deg=[x - 90 for x in [0, 90, 180, 270, 180, 180]], ) logger.info(f"Textured model saved to {textured_glb_path}") return textured_glb_path except Exception as e: logger.error(f"Error in run_texture: {str(e)}") raise @spaces.GPU(duration=5) @torch.no_grad() def run_full_api(image, seed: int = 0, num_inference_steps: int = 50, guidance_scale: float = 7.5, simplify: bool = True, target_face_num: int = DEFAULT_FACE_NUMBER, req: gr.Request = None): try: logger.info("Running run_full_api") if isinstance(image, dict): image_path = image.get("path") or image.get("url") if not image_path: logger.error("Invalid image input: no path or URL provided") raise ValueError("Invalid image input: no path or URL provided") if image_path.startswith("http"): temp_image_path = os.path.join(TMP_DIR, f"input_{get_random_hex()}.png") image_path = download_image(image_path, temp_image_path) elif isinstance(image, str) and image.startswith("http"): temp_image_path = os.path.join(TMP_DIR, f"input_{get_random_hex()}.png") image_path = download_image(image, temp_image_path) else: image_path = image if not isinstance(image, str) or not os.path.exists(image_path): logger.error(f"Invalid image path: {image_path}") raise ValueError(f"Invalid image path: {image_path}") image_seg, mesh_path, textured_glb_path = run_full(image_path, seed, num_inference_steps, guidance_scale, simplify, target_face_num, req) session_hash = os.path.basename(os.path.dirname(textured_glb_path)) logger.info(f"Generated textured model at /files/{session_hash}/{os.path.basename(textured_glb_path)}") return {"file_url": f"/files/{session_hash}/{os.path.basename(textured_glb_path)}"} except Exception as e: logger.error(f"Error in run_full_api: {str(e)}") raise # Define Gradio API endpoint try: logger.info("Initializing Gradio API interface") api_interface = gr.Interface( fn=gradio_generate, inputs=[ gr.Image(type="filepath", label="Image"), gr.Number(label="Seed", value=0, precision=0), gr.Number(label="Inference Steps", value=50, precision=0), gr.Number(label="Guidance Scale", value=7.5), gr.Checkbox(label="Simplify Mesh", value=True), gr.Number(label="Target Face Number", value=DEFAULT_FACE_NUMBER, precision=0) ], outputs="json", api_name="/api/generate" ) logger.info("Gradio API interface initialized successfully") except Exception as e: logger.error(f"Failed to initialize Gradio API interface: {str(e)}") raise HEADER = """ # 🌌 PolyGenixAI: Craft 3D Worlds with Cosmic Precision ## Unleash Infinite Creativity with AI-Powered 3D Generation by AnvilInteractive Solutions

By AnvilInteractive Solutions

## 🚀 Launch Your Creation: 1. **Upload an Image** (clear, single-object images shine brightest) 2. **Choose a Style Filter** to infuse your unique vision 3. Click **Generate 3D Model** to sculpt your mesh 4. Click **Apply Texture** to bring your model to life 5. **Download GLB** to share your masterpiece

Powered by cutting-edge AI and multi-view technology from AnvilInteractive Solutions. Join our PolyGenixAI Community to connect with creators and spark inspiration.

""" # Gradio web interface try: logger.info("Initializing Gradio Blocks interface") with gr.Blocks(title="PolyGenixAI", css="body { background-color: #1A1A1A; } .gr-panel { background-color: #2D2D2D; }") as demo: gr.Markdown(HEADER) with gr.Tabs(elem_classes="gr-tab"): with gr.Tab("Create 3D Model"): with gr.Row(): with gr.Column(scale=1): image_prompts = gr.Image(label="Upload Image", type="filepath", height=300, elem_classes="gr-panel") seg_image = gr.Image(label="Preview Segmentation", type="pil", format="png", interactive=False, height=300, elem_classes="gr-panel") with gr.Accordion("Style & Settings", open=True, elem_classes="gr-accordion"): style_filter = gr.Dropdown( choices=["None", "Realistic", "Fantasy", "Cartoon", "Sci-Fi", "Vintage", "Cosmic", "Neon"], label="Style Filter", value="None", info="Select a style to inspire your 3D model (optional)", elem_classes="gr-dropdown" ) seed = gr.Slider( label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=0, elem_classes="gr-slider" ) randomize_seed = gr.Checkbox(label="Randomize Seed", value=True) num_inference_steps = gr.Slider( label="Inference Steps", minimum=8, maximum=50, step=1, value=50, info="Higher steps enhance detail but increase processing time", elem_classes="gr-slider" ) guidance_scale = gr.Slider( label="Guidance Scale", minimum=0.0, maximum=20.0, step=0.1, value=7.0, info="Controls adherence to input image", elem_classes="gr-slider" ) reduce_face = gr.Checkbox(label="Simplify Mesh", value=True) target_face_num = gr.Slider( maximum=1000000, minimum=10000, value=DEFAULT_FACE_NUMBER, label="Target Face Number", info="Adjust mesh complexity for performance", elem_classes="gr-slider" ) gen_button = gr.Button("Generate 3D Model", variant="primary", elem_classes="gr-button-primary") gen_texture_button = gr.Button("Apply Texture", variant="secondary", interactive=False, elem_classes="gr-button-secondary") with gr.Column(scale=1): model_output = gr.Model3D(label="3D Model Preview", interactive=False, height=400, elem_classes="gr-panel") textured_model_output = gr.Model3D(label="Textured 3D Model", interactive=False, height=400, elem_classes="gr-panel") download_button = gr.Button("Download GLB", variant="secondary", elem_classes="gr-button-secondary") with gr.Tab("Cosmic Gallery"): gr.Markdown("### Discover Stellar Creations") gr.Examples( examples=[ f"{TRIPOSG_CODE_DIR}/assets/example_data/{image}" for image in os.listdir(f"{TRIPOSG_CODE_DIR}/assets/example_data") ], fn=run_full, inputs=[image_prompts], outputs=[seg_image, model_output, textured_model_output], cache_examples=True, ) gr.Markdown("Connect with creators in our PolyGenixAI Cosmic Community!") gen_button.click( run_segmentation, inputs=[image_prompts], outputs=[seg_image] ).then( get_random_seed, inputs=[randomize_seed, seed], outputs=[seed], ).then( image_to_3d, inputs=[ seg_image, seed, num_inference_steps, guidance_scale, reduce_face, target_face_num ], outputs=[model_output] ).then(lambda: gr.Button(interactive=True), outputs=[gen_texture_button]) gen_texture_button.click( run_texture, inputs=[image_prompts, model_output, seed], outputs=[textured_model_output] ) demo.load(start_session) demo.unload(end_session) logger.info("Gradio Blocks interface initialized successfully") except Exception as e: logger.error(f"Failed to initialize Gradio Blocks interface: {str(e)}") raise if __name__ == "__main__": try: logger.info("Launching Gradio application") demo.launch(server_name="0.0.0.0", server_port=7860, show_error=True) logger.info("Gradio application launched successfully") except Exception as e: logger.error(f"Failed to launch Gradio application: {str(e)}") raise