motion_generate_tools is a Blender addon for generate motion using the following Model.
motion_generate_toolsは以下のモデルを使ってモーションを生成するためのBlenderアドオンです。
Download motion_generate_tools from the github release page
The easiest way to install the motion_generate_tools is to do so through directly through Blender:
- Open the Blender User Preferences menu and select the Add-ons tab (Edit > Preferences... > Add-ons)
- Click the Install... button at the top of the Add-ons menu. This will open a file menu where you can select the
motion_generate_tools-v0.0.0.zip
file. - After installing the addon .zip file, Blender will filter the addons list to only show the motion_generate_tools addon. Click the checkbox next to the title to enable the addon.
- Click Required Python Modules: Update Python Modules button at the addon Preferences.
- If you have a NVIDIA GPU that supports CUDA, you can choose with CUDA button.
- Click CLIP Model: Download CLIP Model button at the addon Preferences.
- Restart Blender. A Blender restart is maybe required to complete the installation.
install_motion_generate_tools.mp4
- Blender 3.3 LTS or later
- Open the
smpl_model_20210421-bones.blend
file - Select the SMPLX-neutral armature
- Goto 3D Viewport > Sidebar > Animation > Motion Generator Panel
- Enter a Text Condition
a person jumps. a person throws. a man is locking his hans behind his back and sweeping his legs right and left, in a dance like motion. a person is walking in a counterclockwise circle.
- Press Generate Motion button
- Only SMPL armature is available.
- The motion to be edited is fixed.
This addon is based on code from the following repository:
- MDM: Human Motion Diffusion Model
mdm/data_loaders/
mdm/diffusion/
mdm/model/
mdm/utils/
This addon depends on the following datasets:
- MDM: Human Motion Diffusion Model
data/save/
- HumanML3D: 3D Human Motion-Language Dataset
data/dataset/HumanML3D
- SMPL: A Skinned Multi-Person Linear Model
data/body_models/smpl
We need people to test the installation and give us feedback!
- Discord Server: https://discord.gg/zRgUkuaPWw
This code is distributed under the GPLv3.
Note that our code depends on other libraries, including MDM, CLIP, SMPL, SMPL-X, PyTorch3D, and uses datasets that each have their own respective licenses that must also be followed.