mcp-compatible

README.md

@@ -4,7 +4,7 @@ emoji: 🏢
 colorFrom: indigo
 colorTo: blue
 sdk: gradio
-sdk_version: 5.34.2
+sdk_version: 4.44.1
 app_file: app.py
 pinned: false
 license: mit

app.py

@@ -1,114 +1,16 @@
-import os
-os.environ['SPCONV_ALGO'] = 'native'
-os.environ['ATTN_BACKEND'] = 'xformers'
-os.environ['SPARSE_ATTN'] = 'xformers'
-
-import subprocess, sys, tempfile, ctypes
-
-try:
-    import gradio_litmodel3d  # noqa: F401
-except ImportError:
-    subprocess.check_call(
-        [sys.executable, "-m", "pip", "install", "--no-deps",
-         "gradio_litmodel3d==0.0.1"],
-    )
-
-import spaces
-
-CUDA_HOME = "/cuda-image/usr/local/cuda-13.0"
-CUDA_LIBDIR = os.path.join(CUDA_HOME, "lib64")
-REPO_DIR = os.path.dirname(os.path.abspath(__file__))
-
-
-@spaces.GPU(duration=600)
-def first_gpu_setup():
-    need_nvdiffrast = need_dgr = False
-    try:
-        import nvdiffrast  # noqa: F401
-    except ImportError:
-        need_nvdiffrast = True
-    try:
-        import diff_gaussian_rasterization  # noqa: F401
-    except ImportError:
-        need_dgr = True
-    if not (need_nvdiffrast or need_dgr):
-        print("CUDA extensions already installed; skipping build.")
-        return
-
-    if not os.path.exists(os.path.join(CUDA_HOME, "bin", "nvcc")):
-        raise RuntimeError(
-            f"nvcc not found at {CUDA_HOME}/bin/nvcc on the ZeroGPU worker. "
-            "The new-hardware CUDA path may have moved; please update CUDA_HOME."
-        )
-
-    patch_dir = tempfile.mkdtemp(prefix="torch_cuda_patch_")
-    with open(os.path.join(patch_dir, "sitecustomize.py"), "w") as f:
-        f.write(
-            "try:\n"
-            "    import torch.utils.cpp_extension as _c\n"
-            "    _c._check_cuda_version = lambda *a, **k: None\n"
-            "except Exception:\n"
-            "    pass\n"
-        )
-
-    env = os.environ.copy()
-    env["CUDA_HOME"] = CUDA_HOME
-    env["CUDA_PATH"] = CUDA_HOME
-    env["PATH"] = os.path.join(CUDA_HOME, "bin") + os.pathsep + env.get("PATH", "")
-    env["PYTHONPATH"] = patch_dir + os.pathsep + env.get("PYTHONPATH", "")
-    env["TORCH_CUDA_ARCH_LIST"] = "12.0"
-
-    subprocess.check_call(
-        [sys.executable, "-m", "pip", "install", "--no-deps",
-         "setuptools", "wheel", "ninja"],
-    )
-
-    if need_nvdiffrast:
-        subprocess.check_call(
-            [sys.executable, "-m", "pip", "install", "--no-build-isolation",
-             os.path.join(REPO_DIR, "extensions", "nvdiffrast")],
-            env=env,
-        )
-
-    if need_dgr:
-        mip_dir = tempfile.mkdtemp(prefix="mip_splatting_")
-        subprocess.check_call(
-            ["git", "clone", "--recursive", "--depth=1",
-             "https://github.com/autonomousvision/mip-splatting.git", mip_dir],
-        )
-        subprocess.check_call(
-            [sys.executable, "-m", "pip", "install", "--no-build-isolation",
-             os.path.join(mip_dir, "submodules", "diff-gaussian-rasterization")],
-            env=env,
-        )
-
-
-first_gpu_setup()
-
-ctypes.CDLL(os.path.join(CUDA_LIBDIR, "libcudart.so.13"), mode=ctypes.RTLD_GLOBAL)
-os.environ["LD_LIBRARY_PATH"] = CUDA_LIBDIR + os.pathsep + os.environ.get("LD_LIBRARY_PATH", "")
-
 import gradio as gr
+import spaces
 from gradio_litmodel3d import LitModel3D
 
+import os
 import shutil
+os.environ['SPCONV_ALGO'] = 'native'
 from typing import *
 import torch
 import numpy as np
 import imageio
 from easydict import EasyDict as edict
 from PIL import Image
-
-# xformers on Blackwell (sm_120) picks Flash-Attn-3 (Hopper-only) and crashes
-# with "invalid argument". Force it to use Cutlass kernels instead.
-import xformers.ops as _xops
-_orig_mea = _xops.memory_efficient_attention
-_cutlass_op = (_xops.fmha.cutlass.FwOp, _xops.fmha.cutlass.BwOp)
-def _mea_cutlass(*args, **kwargs):
-    kwargs.setdefault("op", _cutlass_op)
-    return _orig_mea(*args, **kwargs)
-_xops.memory_efficient_attention = _mea_cutlass
-
 from trellis.pipelines import TrellisImageTo3DPipeline
 from trellis.representations import Gaussian, MeshExtractResult
 from trellis.utils import render_utils, postprocessing_utils
@@ -131,17 +33,13 @@ def end_session(req: gr.Request):
 
 def preprocess_image(image: Image.Image) -> Image.Image:
     """
-    Preprocess the input image for 3D generation.
-    
-    This function is called when a user uploads an image or selects an example.
-    It applies background removal and other preprocessing steps necessary for
-    optimal 3D model generation.
+    Preprocess the input image.
 
     Args:
-        image (Image.Image): The input image from the user
+        image (Image.Image): The input image.
 
     Returns:
-        Image.Image: The preprocessed image ready for 3D generation
+        Image.Image: The preprocessed image.
     """
     processed_image = pipeline.preprocess_image(image)
     return processed_image
@@ -149,16 +47,13 @@ def preprocess_image(image: Image.Image) -> Image.Image:
 
 def preprocess_images(images: List[Tuple[Image.Image, str]]) -> List[Image.Image]:
     """
-    Preprocess a list of input images for multi-image 3D generation.
-    
-    This function is called when users upload multiple images in the gallery.
-    It processes each image to prepare them for the multi-image 3D generation pipeline.
+    Preprocess a list of input images.
     
     Args:
-        images (List[Tuple[Image.Image, str]]): The input images from the gallery
+        images (List[Tuple[Image.Image, str]]): The input images.
         
     Returns:
-        List[Image.Image]: The preprocessed images ready for 3D generation
+        List[Image.Image]: The preprocessed images.
     """
     images = [image[0] for image in images]
     processed_images = [pipeline.preprocess_image(image) for image in images]
@@ -207,23 +102,13 @@ def unpack_state(state: dict) -> Tuple[Gaussian, edict, str]:
 
 def get_seed(randomize_seed: bool, seed: int) -> int:
     """
-    Get the random seed for generation.
-    
-    This function is called by the generate button to determine whether to use
-    a random seed or the user-specified seed value.
-    
-    Args:
-        randomize_seed (bool): Whether to generate a random seed
-        seed (int): The user-specified seed value
-        
-    Returns:
-        int: The seed to use for generation
+    Get the random seed.
     """
     return np.random.randint(0, MAX_SEED) if randomize_seed else seed
 
 
-@spaces.GPU(duration=120)
-def generate_and_extract_glb(
+@spaces.GPU
+def image_to_3d(
     image: Image.Image,
     multiimages: List[Tuple[Image.Image, str]],
     is_multiimage: bool,
@@ -233,12 +118,10 @@ def generate_and_extract_glb(
     slat_guidance_strength: float,
     slat_sampling_steps: int,
     multiimage_algo: Literal["multidiffusion", "stochastic"],
-    mesh_simplify: float,
-    texture_size: int,
     req: gr.Request,
-) -> Tuple[dict, str, str, str]:
+) -> Tuple[dict, str]:
     """
-    Convert an image to a 3D model and extract GLB file.
+    Convert an image to a 3D model.
 
     Args:
         image (Image.Image): The input image.
@@ -250,18 +133,12 @@ def generate_and_extract_glb(
         slat_guidance_strength (float): The guidance strength for structured latent generation.
         slat_sampling_steps (int): The number of sampling steps for structured latent generation.
         multiimage_algo (Literal["multidiffusion", "stochastic"]): The algorithm for multi-image generation.
-        mesh_simplify (float): The mesh simplification factor.
-        texture_size (int): The texture resolution.
 
     Returns:
         dict: The information of the generated 3D model.
         str: The path to the video of the 3D model.
-        str: The path to the extracted GLB file.
-        str: The path to the extracted GLB file (for download).
     """
     user_dir = os.path.join(TMP_DIR, str(req.session_hash))
-    
-    # Generate 3D model
     if not is_multiimage:
         outputs = pipeline.run(
             image,
@@ -293,43 +170,53 @@ def generate_and_extract_glb(
             },
             mode=multiimage_algo,
         )
-    
-    # Render video
     video = render_utils.render_video(outputs['gaussian'][0], num_frames=120)['color']
     video_geo = render_utils.render_video(outputs['mesh'][0], num_frames=120)['normal']
     video = [np.concatenate([video[i], video_geo[i]], axis=1) for i in range(len(video))]
     video_path = os.path.join(user_dir, 'sample.mp4')
     imageio.mimsave(video_path, video, fps=15)
-    
-    # Extract GLB
-    gs = outputs['gaussian'][0]
-    mesh = outputs['mesh'][0]
+    state = pack_state(outputs['gaussian'][0], outputs['mesh'][0])
+    torch.cuda.empty_cache()
+    return state, video_path
+
+
+@spaces.GPU(duration=90)
+def extract_glb(
+    state: dict,
+    mesh_simplify: float,
+    texture_size: int,
+    req: gr.Request,
+) -> Tuple[str, str]:
+    """
+    Extract a GLB file from the 3D model.
+
+    Args:
+        state (dict): The state of the generated 3D model.
+        mesh_simplify (float): The mesh simplification factor.
+        texture_size (int): The texture resolution.
+
+    Returns:
+        str: The path to the extracted GLB file.
+    """
+    user_dir = os.path.join(TMP_DIR, str(req.session_hash))
+    gs, mesh = unpack_state(state)
     glb = postprocessing_utils.to_glb(gs, mesh, simplify=mesh_simplify, texture_size=texture_size, verbose=False)
     glb_path = os.path.join(user_dir, 'sample.glb')
     glb.export(glb_path)
-    
-    # Pack state for optional Gaussian extraction
-    state = pack_state(gs, mesh)
-    
     torch.cuda.empty_cache()
-    return state, video_path, glb_path, glb_path
+    return glb_path, glb_path
 
 
 @spaces.GPU
 def extract_gaussian(state: dict, req: gr.Request) -> Tuple[str, str]:
     """
-    Extract a Gaussian splatting file from the generated 3D model.
-    
-    This function is called when the user clicks "Extract Gaussian" button.
-    It converts the 3D model state into a .ply file format containing
-    Gaussian splatting data for advanced 3D applications.
+    Extract a Gaussian file from the 3D model.
 
     Args:
-        state (dict): The state of the generated 3D model containing Gaussian data
-        req (gr.Request): Gradio request object for session management
+        state (dict): The state of the generated 3D model.
 
     Returns:
-        Tuple[str, str]: Paths to the extracted Gaussian file (for display and download)
+        str: The path to the extracted Gaussian file.
     """
     user_dir = os.path.join(TMP_DIR, str(req.session_hash))
     gs, _ = unpack_state(state)
@@ -355,17 +242,7 @@ def prepare_multi_example() -> List[Image.Image]:
 
 def split_image(image: Image.Image) -> List[Image.Image]:
     """
-    Split a multi-view image into separate view images.
-    
-    This function is called when users select multi-image examples that contain
-    multiple views in a single concatenated image. It automatically splits them
-    based on alpha channel boundaries and preprocesses each view.
-    
-    Args:
-        image (Image.Image): A concatenated image containing multiple views
-        
-    Returns:
-        List[Image.Image]: List of individual preprocessed view images
+    Split an image into multiple views.
     """
     image = np.array(image)
     alpha = image[..., 3]
@@ -381,9 +258,8 @@ def split_image(image: Image.Image) -> List[Image.Image]:
 with gr.Blocks(delete_cache=(600, 600)) as demo:
     gr.Markdown("""
     ## Image to 3D Asset with [TRELLIS](https://trellis3d.github.io/)
-    * Upload an image and click "Generate & Extract GLB" to create a 3D asset and automatically extract the GLB file.
-    * If you want the Gaussian file as well, click "Extract Gaussian" after generation.
-    * If the image has alpha channel, it will be used as the mask. Otherwise, we use `rembg` to remove the background.
+    * Upload an image and click "Generate" to create a 3D asset. If the image has alpha channel, it be used as the mask. Otherwise, we use `rembg` to remove the background.
+    * If you find the generated 3D asset satisfactory, click "Extract GLB" to extract the GLB file and download it.
     
     ✨New: 1) Experimental multi-image support. 2) Gaussian file extraction.
     """)
@@ -413,13 +289,16 @@ with gr.Blocks(delete_cache=(600, 600)) as demo:
                     slat_guidance_strength = gr.Slider(0.0, 10.0, label="Guidance Strength", value=3.0, step=0.1)
                     slat_sampling_steps = gr.Slider(1, 50, label="Sampling Steps", value=12, step=1)
                 multiimage_algo = gr.Radio(["stochastic", "multidiffusion"], label="Multi-image Algorithm", value="stochastic")
+
+            generate_btn = gr.Button("Generate")
             
             with gr.Accordion(label="GLB Extraction Settings", open=False):
                 mesh_simplify = gr.Slider(0.9, 0.98, label="Simplify", value=0.95, step=0.01)
                 texture_size = gr.Slider(512, 2048, label="Texture Size", value=1024, step=512)
-
-            generate_btn = gr.Button("Generate & Extract GLB", variant="primary")
-            extract_gs_btn = gr.Button("Extract Gaussian", interactive=False)
+            
+            with gr.Row():
+                extract_glb_btn = gr.Button("Extract GLB", interactive=False)
+                extract_gs_btn = gr.Button("Extract Gaussian", interactive=False)
             gr.Markdown("""
                         *NOTE: Gaussian file can be very large (~50MB), it will take a while to display and download.*
                         """)
@@ -487,17 +366,26 @@ with gr.Blocks(delete_cache=(600, 600)) as demo:
         inputs=[randomize_seed, seed],
         outputs=[seed],
     ).then(
-        generate_and_extract_glb,
-        inputs=[image_prompt, multiimage_prompt, is_multiimage, seed, ss_guidance_strength, ss_sampling_steps, slat_guidance_strength, slat_sampling_steps, multiimage_algo, mesh_simplify, texture_size],
-        outputs=[output_buf, video_output, model_output, download_glb],
+        image_to_3d,
+        inputs=[image_prompt, multiimage_prompt, is_multiimage, seed, ss_guidance_strength, ss_sampling_steps, slat_guidance_strength, slat_sampling_steps, multiimage_algo],
+        outputs=[output_buf, video_output],
     ).then(
         lambda: tuple([gr.Button(interactive=True), gr.Button(interactive=True)]),
-        outputs=[extract_gs_btn, download_glb],
+        outputs=[extract_glb_btn, extract_gs_btn],
     )
 
     video_output.clear(
-        lambda: tuple([gr.Button(interactive=False), gr.Button(interactive=False), gr.Button(interactive=False)]),
-        outputs=[extract_gs_btn, download_glb, download_gs],
+        lambda: tuple([gr.Button(interactive=False), gr.Button(interactive=False)]),
+        outputs=[extract_glb_btn, extract_gs_btn],
+    )
+
+    extract_glb_btn.click(
+        extract_glb,
+        inputs=[output_buf, mesh_simplify, texture_size],
+        outputs=[model_output, download_glb],
+    ).then(
+        lambda: gr.Button(interactive=True),
+        outputs=[download_glb],
     )
     
     extract_gs_btn.click(
@@ -510,8 +398,8 @@ with gr.Blocks(delete_cache=(600, 600)) as demo:
     )
 
     model_output.clear(
-        lambda: tuple([gr.Button(interactive=False), gr.Button(interactive=False)]),
-        outputs=[download_glb, download_gs],
+        lambda: gr.Button(interactive=False),
+        outputs=[download_glb],
     )
     
 
@@ -523,4 +411,4 @@ if __name__ == "__main__":
         pipeline.preprocess_image(Image.fromarray(np.zeros((512, 512, 3), dtype=np.uint8)))    # Preload rembg
     except:
         pass
-    demo.launch(mcp_server=True)
+    demo.launch()

requirements.txt

@@ -1,5 +1,7 @@
-torch==2.8.0
-torchvision==0.23.0
+--extra-index-url https://download.pytorch.org/whl/cu121
+
+torch==2.4.0
+torchvision==0.19.0
 pillow==10.4.0
 imageio==2.36.1
 imageio-ffmpeg==0.5.1
@@ -15,6 +17,11 @@ pyvista==0.44.2
 pymeshfix==0.17.0
 igraph==0.11.8
 git+https://github.com/EasternJournalist/utils3d.git@9a4eb15e4021b67b12c460c7057d642626897ec8
-xformers
+xformers==0.0.27.post2
 spconv-cu120==2.3.6
 transformers==4.46.3
+gradio_litmodel3d==0.0.1
+pydantic==2.10.6
+https://github.com/Dao-AILab/flash-attention/releases/download/v2.7.0.post2/flash_attn-2.7.0.post2+cu12torch2.4cxx11abiFALSE-cp310-cp310-linux_x86_64.whl
+https://huggingface.co/spaces/JeffreyXiang/TRELLIS/resolve/main/wheels/diff_gaussian_rasterization-0.0.0-cp310-cp310-linux_x86_64.whl?download=true
+https://huggingface.co/spaces/JeffreyXiang/TRELLIS/resolve/main/wheels/nvdiffrast-0.3.3-cp310-cp310-linux_x86_64.whl?download=true