Furkan Gözükara
patreon
Wan2GP 1-Click Automatic Installer for Windows, RunPod and Massed Compute with Literally All GPUs Support with Flash Attention + xFormers + Torch 2.8 and CUDA 12.9
🕑 Added 2025-09-27 00:21:53 +0000 UTCPatreon exclusive posts index to find our scripts easily, Patreon scripts updates history to see which updates arrived to which scripts and amazing Patreon special generative scripts list that you can use in any of your task.
Join discord to get help, chat, discuss and also tell me your discord username to get your special rank : SECourses Discord
Please also Star, Watch and Fork our Stable Diffusion & Generative AI GitHub repository and join our Reddit subreddit and follow me on LinkedIn (my real profile)
=======
Latest installer zip file : Wan2GP_v2.zip
Official repo to learn more : https://github.com/deepbeepmeep/Wan2GP
My installer will make a venv and install this app with Torch 2.8, CUDA 12.9, Flash Attention 2.8.3, Sage Attention 2.2.0, Triton 3.4, xFormers 0.0.33
However, i noticed that not all models are supporting Sage Attention therefore I modified Windows_Start_App.bat to start with Flash Attention
Same thing happens for RunPod and Massed Compute
Moreover, the repo used onnx scripts are not supporting RTX 5000 series so I made the app use CPU, this only impacts extraction of poses etc and thus the speed loss is totally negligible
I have tested Wan 2.2 Animate and the VRAM usage is extremely optimized
However I have 0 idea about optimal settings and I may research later
It also automatically downloads all necessary models however I didn't see it downloads and applies speed LoRAs
So you have to download and apply speed LoRAs and set settings accordingly
30 September 2025 Update V2
numpy and onnxruntime versions downgraded for more compability
Windows Requirements
Python 3.10.11, FFmpeg, CUDA 12.9, cuDNN 9.12, C++ Tools, MSVC and Git
If you get any errors follow below video and its source link
https://www.patreon.com/posts/click-to-open-post-used-in-tutorial-111553210
Massed Compute (Recommend Cloud) :
Please register via this link : https://vm.massedcompute.com/signup?linkId=lp_034338&sourceId=secourses&tenantId=massed-compute
Use our coupon SECourses
Our coupon works on all GPUs now
H100 has amazing price and speed but you can use like RTX A6000 ADA as well
Full details here : https://www.patreon.com/posts/26671823
Then select our image SECourses from Creator dropdown
Then follow Massed_Compute_Instructions_READ.txt
Same as my any other Massed Compute installer script
Example tutorial for learn how to install and use Massed Compute
(Starts at 12:58) : https://youtu.be/KW-MHmoNcqo?si=G1WbG-Qw4ujWvOtG&t=778
RunPod (Cloud):
Please register via this link : https://runpod.io?ref=1aka98lq
Then follow Runpod_Instructions_READ.txt
Same as my any other RunPod installer script
Use the template written in Runpod_Instructions_READ.txt file
Example tutorial for learn how to install and use RunPod
(starts at 22:03) : https://youtu.be/KW-MHmoNcqo?si=QN8X8Sjn13ZYu-EU&t=1323
Complete List of Supported Models in WanGP
🎬 Wan 2.1 Text-to-Video Models
Wan2.1 Text2video 1.3B
Wan2.1 Text2video 14B
Wan2.1 Text2video FusioniX 14B - Enhanced with FusioniX LoRA
Wan2.1 Text2video Self-Forcing 14B - 4-8 steps generation
MoviiGen 1080p 14B - Experimental 1080p
🎬 Wan 2.2 Text-to-Video Models
Wan2.2 Text2video 14B - Latest Wan architecture
Wan2.2 TextImage2video 5B - Text + image input
Wan2.2 FastWan TextImage2video 5B - Accelerated version
🖼️ Wan 2.1 Image-to-Video Models
Wan2.1 Image2video 480p 14B - Standard resolution
Wan2.1 Image2video 720p 14B - High resolution
Wan2.1 Image2video 480p FusioniX 14B - Enhanced quality
Fun InP image2video 1.3B - Low VRAM option
Fun InP image2video 14B - High quality option
🖼️ Wan 2.2 Image-to-Video Models
Wan2.2 Image2video 14B - Latest architecture
Wan2.2 Multitalk 14B - Multi-person talking heads
🎭 Wan Animation & Character Models
Wan2.2 Animate - Body/facial motion transfer
Multitalk 480p - Multi-person lip sync
Multitalk 720p - High-res multi-person
Infinitetalk Single Speaker 480p - Single person talking
Infinitetalk Multi Speakers 480p - Multiple speakers
Fantasy Talking 720p - Fantasy character animation
🎨 Wan VACE (ControlNet) Models
Vace ControlNet 1.3B - Advanced control, low VRAM
Vace ControlNet 14B - Advanced control, high quality
Vace FusioniX 14B - Enhanced with FusioniX
Vace Cocktail 14B - Multiple LoRAs combined
Vace Self-Forcing 14B - Fast generation
Vace Multitalk FusioniX 14B - Talking heads with control
Vace Standin 14B - Person replacement
Wan2.2 Vace Fun 14B - Wan 2.2 + VACE
Wan2.2 Vace Fun Cocktail 14B - Enhanced cocktail
Wan2.2 Vace Experimental Cocktail 14B - Experimental mix
Wan2.2 Vace Lightning 3 Phases 14B - Multi-phase generation
👻 Wan Specialized Models
Phantom 1.3B - Person/object transfer, low VRAM
Phantom 14B - Person/object transfer, high quality
Wan2.1 Standin 14B - Person replacement
ReCamMaster 1.3B - Viewpoint change
Wan2.2 Lucy Edit 5B - Video editing
Wan2.2 FastWan Lucy Edit 5B - Fast video editing
🌌 Sky Reels Models
SkyReels2 Diffusion Forcing 1.3B - Infinite length videos
SkyReels2 Diffusion Forcing 540p 14B - Standard resolution
SkyReels2 Diffusion Forcing 720p 14B - High resolution
🎥 Hunyuan Video Models
Hunyuan Video Text2video 720p 13B - High-quality T2V
Hunyuan Video Image2video 720p 13B - High-quality I2V
Hunyuan Video Custom 720p 13B - Identity preservation
Hunyuan Video Custom Audio 720p 13B - Audio-driven
Hunyuan Video Custom Edit 720p 13B - Video editing
Hunyuan Video Avatar 720p 13B - Avatar generation
Hunyuan Video FastHunyuan 720p 13B - 6-step generation
Hunyuan Video AccVideo 720p 13B - Accelerated version
🎬 LTX Video Models
LTX Video 0.9.8 13B - Long video generation
LTX Video 0.9.8 Distilled 13B - Ultra-fast generation
🎨 Flux Models
Flux 1 Dev 12B - Standard Flux
Flux 1 Schnell 12B - Fast generation
Flux 1 Dev USO 12B - USO variant
Flux 1 Dev UMO 12B - UMO variant
Flux 1 Dev Kontext 12B - Context-aware
Flux 1 Krea Dev 12B - Krea variant
Flux 1 SRPO Dev 12B - SRPO optimization
Flux 1 SRPO USO 12B - SRPO + USO
Flux 1 Chroma 1 HD 8.9B - High-definition
🖼️ Qwen Image Models
Qwen Image 20B - Large image model
Qwen Image Edit 20B - Image editing
Qwen Image Edit Plus 20B - Enhanced editing
🎞️ Specialized Video Models
First Last Frame to Video 720p (FLF2V) 14B - Start/end frame specialist
Comments
Furkan Gözükara
our windows installer already doing that :D pip install onnxruntime==1.17.0 pip install numpy==1.26.4
Loyd Keudem
Hi guys to solve this problem : The original AttributeError: module 'onnxruntime' has no attribute 'SessionOptions' issue has been resolved by: 1. Downgrading onnxruntime from 1.23.0 to 1.17.0 2. Downgrading NumPy from 2.2.6 to 1.26.4 for compatibility I fix that by doing this : pip uninstall numpy -y pip install "numpy>=2.0,<2.3" pip install --upgrade onnxruntime It solve the problem and the gradio run amazing well
Furkan Gözükara
so first time worked and second time broken? can you run installer again and launch and try see if works?
Nobbit K
Hey, unfortunately it looks like the onyxruntime and numby issue is still persistent after a clean v2 installation. It occured after the second launch of the .bat. The first launch and relaunch worked like intended. Traceback (most recent call last): File "E:\AI local\Wan2GP_v2\Wan2GP\wgp.py", line 25, in from shared.utils.utils import convert_tensor_to_image, save_image, get_video_info, get_file_creation_date, convert_image_to_video, calculate_new_dimensions, convert_image_to_tensor, calculate_dimensions_and_resize_image, rescale_and_crop, get_video_frame, resize_and_remove_background, rgb_bw_to_rgba_mask File "E:\AI local\Wan2GP_v2\Wan2GP\shared\utils\utils.py", line 13, in from rembg import remove, new_session File "E:\AI local\Wan2GP_v2\Wan2GP\venv\lib\site-packages\rembg\__init__.py", line 5, in from .bg import remove File "E:\AI local\Wan2GP_v2\Wan2GP\venv\lib\site-packages\rembg\bg.py", line 23, in from .session_factory import new_session File "E:\AI local\Wan2GP_v2\Wan2GP\venv\lib\site-packages\rembg\session_factory.py", line 6, in from .sessions import sessions_class File "E:\AI local\Wan2GP_v2\Wan2GP\venv\lib\site-packages\rembg\sessions\__init__.py", line 5, in from .base import BaseSession File "E:\AI local\Wan2GP_v2\Wan2GP\venv\lib\site-packages\rembg\sessions\base.py", line 10, in class BaseSession: File "E:\AI local\Wan2GP_v2\Wan2GP\venv\lib\site-packages\rembg\sessions\base.py", line 13, in BaseSession def __init__(self, model_name: str, sess_opts: ort.SessionOptions, *args, **kwargs): AttributeError: module 'onnxruntime' has no attribute 'SessionOptions'
Furkan Gözükara
no not normal at all.
Walker4k
Is it normal to take 2 hours to generate a 4 sec I2V using an H100 at MC?
