Closed wdcww closed 4 months ago
Thank you for the detailed description of the problem.
requirements.txt
while I try to add TensorRT support. I have just reverted the change. So you can now install with git pull origin main; pip -r requirements.txt
. Thanks for the reporting!.safetensors
files like BlazingDriveV11m.safetensors
(e.g., from CivitAI) like <demo/semantic_palette/checkpoints/BlazingDriveV11m.safetensor>
. Then you may use it by calling the app with python app.py --model "BlazingDriveV11m.safetensors"
.requirements.txt
. It means your pytorch-compatible diffusers
library is outdated to not have LCMScheduler
class. Please try installing again with the instruction I provided in answer (1).I just have tried installing this from scratch and it works under my environment settings. You may reopen this issue if any other problems are found. Thanks again for the reporting!
First of all, I really appreciate you replying to me in detail.
Your updated requirements.txt can indeed install packages without conflicts,but the CPU version of torch was installed for me.
Although your previous question (#6) provided PyTorch 2.0.1 + cpu, the warning reminded us: xFormers was built for PyTorch 2.0.1+cu118 with CUDA 1108, so I had to switch to the GPU version of torch.
I use this:
conda install pytorch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 pytorch-cuda=11.8 -c pytorch -c nvidia
Matching the environment is really torture for me ! ! !
you are right ! By default,
python app.py --height 512 --width 512 --port 8000
I downloaded and placed the blazingDrive_V11m.safetensors file into checkpoints/
(smd) PS E:\github_download\StreamMultiDiffusion\demo\semantic_palette> python app.py --model "blazingDrive_V11m.safetensors" --height 512 --width 512 --port 8000
Now it can run, I want to continue studying.
Thank you so much!
Ohh, thank you so much for pointing out!
You can always reopen any issue if you find any trouble.
hello, I'm a newbie and want to ask you some questions
1. I followed the installation steps in the readme.md , create a new environment from scratch.
Why does the following error appear when running"pip install -r requirements.txt"?
Does my environment require torch to be installed before I can install your package?
2. Then I changed to an environment that already had installed.
but I have " protobuf 3.20.2 pypi_0 pypi "
3. I don't know what to add to <demo/semantic_palette/checkpoints> ?
And I run app.py
?