IDM VTON : Virtual Try On APP Automatic Installers for Windows, RunPod, Massed Compute and a free Kaggle Account notebook - Can transfer objects too

Furkan Gözükara

your linux machine probably lacking some of the necessary libraries like gcc. you have to setup it exactly as we setup windows : https://youtu.be/DrhUHnYfwC0

eyal giron

~/AI/idm_vton_v21/IDM-VTON on  main ⌚ 10:04:48 $ python3 gradio_demo/app_VTON.py --share True --fixed_vae Traceback (most recent call last): File "/home/eyal/AI/idm_vton_v21/IDM-VTON/gradio_demo/app_VTON.py", line 25, in from preprocess.humanparsing.run_parsing import Parsing File "/home/eyal/AI/idm_vton_v21/IDM-VTON/./preprocess/humanparsing/run_parsing.py", line 5, in import onnxruntime as ort File "/home/eyal/AI/idm_vton_v21/venv/lib/python3.10/site-packages/onnxruntime/__init__.py", line 56, in raise import_capi_exception File "/home/eyal/AI/idm_vton_v21/venv/lib/python3.10/site-packages/onnxruntime/__init__.py", line 23, in from onnxruntime.capi._pybind_state import ExecutionMode # noqa: F401 File "/home/eyal/AI/idm_vton_v21/venv/lib/python3.10/site-packages/onnxruntime/capi/_pybind_state.py", line 32, in from .onnxruntime_pybind11_state import * # noqa ImportError: /home/eyal/AI/idm_vton_v21/venv/lib/python3.10/site-packages/onnxruntime/capi/onnxruntime_pybind11_state.cpython-310-x86_64-linux-gnu.so: cannot enable executable stack as shared object requires: Invalid argument i can't launch this on my linux machine

Furkan Gözükara

it is not an error ignore it

Muhammad Umar Jan

Hey Sir! i got this error The config attributes {'decay': 0.9999, 'inv_gamma': 1.0, 'min_decay': 0.0, 'optimization_step': 37000, 'power': 0.6666666666666666, 'update_after_step': 0, 'use_ema_warmup': False} were passed to UNet2DConditionModel, but are not expected and will be ignored. Please verify your config.json configuration file.

Furkan Gözükara

i think can be ignored

NP

v19 is generating results but shows : FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(dtype=dtype):

Furkan Gözükara

yes Kaggle blocked the Gradio sadly. I reported this error but no solution. not even ngrok works anymore

C Unnikrishnan B Nambiar

Furkan Gözükara

great

Dmitry

I already left a message saying that the problem was solved, but for some reason this was not displayed in patreon. The problem turned out to be in an old installation script that I once downloaded. It was also version 17, so I thought it was relevant. But it wasn't working. I re-downloaded the installation script and everything worked.

Furkan Gözükara

yes i see your error. now what you need to do is edit Windows_Start_VTON_app.bat and remove set HF_HUB_ENABLE_HF_TRANSFER=1 after removing this set HF_HUB_ENABLE_HF_TRANSFER=1 from bat file save and start again

Dmitry

OK, I'll try to reinstall, the second one is there, but there is only 4 GB of video memory. The problem occurs when I upload an image and click the try-on button. The model download starts and at this point the app returns an error after a short download: To create a public link, set `share=True` in `launch()`. auto_crop_upload: Original image stored with resolution: (1050, 1400) auto_crop_upload: Cropping done. Crop region: 0.0 0.0 1050.0 1400.0 start_tryon: Input dict from ImageEditor: {'background': , 'layers': [], 'composite': } start_tryon: Full Input dict content: {'background': , 'layers': [], 'composite': } diffusion_pytorch_model.bin: 3%|█▎ | 336M/12.0G [00:22<05:14, 37.0MB/s]To continue, press any key . . . I'll reinstall it now and try again, I will also send you the installation logs

Furkan Gözükara

i just made a fresh install and tried on 3090 ti no issues. can you make a fresh install and send me logs? monstermmorpg@gmail.com also do you have second gpu etc?

Dmitry

RTX 3090ti

Furkan Gözükara

what is your GPU model? i am making a fresh install now to test

Dmitry

It starts, but then returns an error. I didn't get any errors when installing the app. It is possible that the necessary models were not loaded into the folders at startup? I have cuda 12.4 11.8, python 3.10, 3.11, 3.12 installed I'm confused by this line in the log, I think it's in it, but I can't figure out what to do: "function 'cadam32bit_grad_fp32' not found" PS https://youtu.be/DrhUHnYfwC0 With this video, I generally began my acquaintance with AI). This was my first step, thanks to which all your applications work perfectly for me UPD I noticed that I don't have the cuDNN 8.9 version installed, maybe this is the case, I'll check it out and let you know

Furkan Gözükara

hi is it working? also did you follow this url ? https://youtu.be/DrhUHnYfwC0 looks to me like no issues try it

Dmitry

Modify this file and add --share if you want Gradio share Options VRAM usages are as below. Fastest is FP16 and no cpu offloading (1+1) 1+1 = 16.5 GB, 1+2 = 12 GB , 1+3 = 10 GB, 2+1 = 15 GB, 2+2 = 10.3 GB , 2+3 = 8.2 GB Please select an option: 1. Default - Faster but higher VRAM 2. Enable CPU Offloading - Reduces VRAM but slightly slower same quality Enter your choice (1-2): 1 Please select an option: 1. Default Precision - Higher quality higher VRAM 2. FP8 Precision - Slightly lower quality medium VRAM 3. FP4 Precision - More slightly lower quality lower VRAM Enter your choice (1-3): 1 function 'cadam32bit_grad_fp32' not found INFO: Files cannot be found using the specified templates. * Running on local URL: http://127.0.0.1:7860 To create a public link, set `share=True` in `launch()`. To continue, press any key . . .