Closed EloniX-X closed 12 months ago
You did not install Deep-Daze completely. You need to download a version of TensorFlow before V2 because Deep Daze doesn't support V2 You will also need CUDA 11.1
Microsoft Visual C++ > Download Here Python 3.8 > Download Here TensorFlow python package > Download Here CUDA 11.1 > Download Here
Make a new folder where ever you choose, drag and drop the TensorFlow python package (The .whl) file then run these commands
cd C:\the folder path
pip install tensorflow_gpu-2.4.0-cp38-cp38-win_amd64.whl
After you have done that type this in a new command line The next lines of code are to create a virtual environment and upgrades the different things you need to run the command.
python -m venv --system-site-packages .\venv
.\venv\Scripts\activate
pip install --upgrade pip
pip install torch==1.8.0+cu111 torchvision==0.9.0+cu111 torchaudio===0.8.0 -f https://download.pytorch.org/whl/torch_stable.html
pip install --upgrade CLIP
pip install --upgrade wheel
pip install deep-daze
pip install --upgrade tensorflow
(CLIP, wheel, and torch are just to verify that you have the most up-to-date installation)
I might be missing some information so let me know how all these goes and im here to help!
Does that mean something like
%tensorflow_version 1.x
%pip install torch==1.8.0+cu111 torchvision==0.9.0+cu111 torchaudio===0.8.0 -f https://download.pytorch.org/whl/torch_stable.html
%pip install --upgrade CLIP
%pip install deep-daze --upgrade
would be required in the colab notebook now? @lucidrains
Edit: also, if you do get around to changing it, there's a bug in the display code now that deep-daze saves jpg instead of png.
for epoch in trange(20, desc = 'epochs'):
for i in trange(ITERATIONS, desc = 'iteration'):
model.train_step(epoch, i)
if i % model.save_every != 0:
continue
filename = TEXT.replace(' ', '_')
image = Image(f'./{filename}.jpg') # This line
display(image)
Does that mean something like
%tensorflow_version 1.x %pip install torch==1.8.0+cu111 torchvision==0.9.0+cu111 torchaudio===0.8.0 -f https://download.pytorch.org/whl/torch_stable.html %pip install --upgrade CLIP %pip install deep-daze --upgrade
would be required in the colab notebook now? @lucidrains
Edit: also, if you do get around to changing it, there's a bug in the display code now that deep-daze saves jpg instead of png.
for epoch in trange(20, desc = 'epochs'): for i in trange(ITERATIONS, desc = 'iteration'): model.train_step(epoch, i) if i % model.save_every != 0: continue filename = TEXT.replace(' ', '_') image = Image(f'./{filename}.jpg') # This line display(image)
Yes, those were the steps I took, and the upgrade commands I put all needed to be updated, and either way it's just making sure you have to most up-to-date versions.
.