Skip to content

adityahbapat/nft-generative-models-backend

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 

Repository files navigation

Creating NFT using generative models backend

The idea is to provide user with images with their specifications, give them the option to upload their NFT in ipfs get the link and then give them option to sell their NFT at leading NFT Marketplaces like OpenSea all with a hassle free no code experience.

Generative Models used for image creation:

  • Stable Diffusion (Text to Image)
  • Neural Style Transfer (One Image to Another)
  • Deep Dream (visualizes the patterns learned by a neural network)

GPU required

  • The GPU configuration uhhhhh, almost made me cry :)
  • On windows in virtual environment terminal run Set-ExecutionPolicy -Scope CurrentUser -ExecutionPolicy Unrestricted
  • pip list to see all packages
  • pip cache purge to clear pip cache
  • Please make sure of your tensorflow version and if it supports Cuda tf.test.is_built_with_cuda()
  • I uninstalled my previous tensorflow version, upgraded pip and reinstalled version 2.10.0 for python version 3.10.6
  • with this tf version, keras_cv version 0.0.25 is compatible, any other version will lead to error pip uninstall -y $(pip freeze | Select-String tensor | ForEach-Object { $_.ToString().Split('==')[0] }) same for keras pip install tensorflow==2.10.0
  • Finally run code to check gpu availability:
import tensorflow as tf

print("TensorFlow version:", tf.__version__)
# Get the list of available GPUs
gpus = tf.config.list_physical_devices('GPU')
if gpus:
  for gpu in gpus:
    print("GPU detected:", gpu)
else:
  print("No GPU detected")

# Check if TensorFlow is using the GPU
print("Is TensorFlow using GPU?", tf.test.is_built_with_cuda())
print("Is GPU available for TensorFlow?", tf.test.is_gpu_available()) 

Releases

No releases published

Packages

No packages published