Add instructions of installing dependencies using pip

* Enable to support Windows and venv.
* Simplify the installation.

Author: gongminmin

README.md

@@ -52,7 +52,11 @@
     ```
 
 2. Install the dependencies:
-    
+
+   There are two ways of installing dependencies.
+
+    1. Using a script (Linux & Conda only)
+
     **Before running the following command there are somethings to note:**
     - By adding `--new-env`, a new conda environment named `trellis` will be created. If you want to use an existing conda environment, please remove this flag.
     - By default the `trellis` environment will use pytorch 2.4.0 with CUDA 11.8. If you want to use a different version of CUDA (e.g., if you have CUDA Toolkit 12.2 installed and do not want to install another 11.8 version for submodule compilation), you can remove the `--new-env` flag and manually install the required dependencies. Refer to [PyTorch](https://pytorch.org/get-started/previous-versions/) for the installation command.
@@ -83,6 +87,40 @@
         --demo                  Install all dependencies for demo
     ```
 
+    2. Using pip (Windows & venv):
+
+    **Note:** Preinstall Python 3.10 and CUDA SDK 12.4. These versions are crucial. Other combinations may cause issues and will be discussed later.
+
+    Steps:
+
+    - **Create a virtual environment in the `Venv` directory:**
+    ```cmd
+    python -m v Venv
+    ```
+
+    - **Activate the virtual environment:**
+
+    ```cmd
+    Venv\Scripts\activate
+    ```
+
+    - **Install package using pip:** This may take a while.
+    ```cmd
+    pip install -r requirements.txt
+    ```
+
+    - **Install extra packages with dependencies on PyTorch:**
+    ```cmd
+    pip install -r requirements_extra.txt
+    ```
+
+    **Known Issues**
+
+      1. For Python 3.11, there are issues when installing `diffoctreerast` and `diff-gaussian-rasterization`. The error reported is that the module `torch` couldn't be found.
+      2. The package `spconv` does not support Python 3.12 (yet).
+
+      Additional work may be necessary to resolve these issues.
+
 <!-- Pretrained Models -->
 ## 🤖 Pretrained Models
 

requirements.txt

@@ -0,0 +1,35 @@
+--extra-index-url https://download.pytorch.org/whl/cu124
+torch==2.5.1+cu124
+torchvision==0.20.1+cu124
+xformers==0.0.28.post3
+
+pillow==10.2.0
+imageio==2.37.0
+imageio-ffmpeg==0.6.0
+tqdm==4.67.1
+easydict==1.13
+opencv-python-headless==4.11.0.86
+scipy==1.15.1
+ninja==1.11.1.3
+rembg==2.0.61
+onnxruntime==1.20.1
+trimesh==4.6.0
+xatlas==0.0.9
+pyvista==0.44.2
+pymeshfix==0.17.0
+igraph==0.11.8
+transformers==4.48.1
+
+git+https://github.com/EasternJournalist/utils3d.git@9a4eb15e4021b67b12c460c7057d642626897ec8
+
+https://github.com/bdashore3/flash-attention/releases/download/v2.7.1.post1/flash_attn-2.7.1.post1+cu124torch2.5.1cxx11abiFALSE-cp310-cp310-win_amd64.whl
+
+--find-links https://nvidia-kaolin.s3.us-east-2.amazonaws.com/torch-2.5.1_cu124.html
+kaolin
+
+git+https://github.com/NVlabs/nvdiffrast.git
+
+spconv-cu120==2.3.6
+
+gradio==4.44.1
+gradio_litmodel3d==0.0.1

requirements_extra.txt

@@ -0,0 +1,2 @@
+git+https://github.com/JeffreyXiang/diffoctreerast.git
+git+https://github.com/autonomousvision/mip-splatting.git#subdirectory=submodules/diff-gaussian-rasterization/