| | --- |
| | library_name: tf-keras |
| | tags: |
| | - gan |
| | - dcgan |
| | - huggan |
| | - tensorflow |
| | - unconditional-image-generation |
| | --- |
| | |
| | ## Model description |
| |
|
| | Simple DCGAN implementation in TensorFlow to generate CryptoPunks. |
| |
|
| | ## Generated samples |
| | <img src="https://github.com/dimitreOliveira/cryptogans/raw/main/assets/gen_samples.png" width="350" height="350"> |
| |
|
| | Project repository: [CryptoGANs](https://github.com/dimitreOliveira/cryptogans). |
| |
|
| | ## Usage |
| |
|
| | You can play with the HuggingFace [space demo](https://huggingface.co/spaces/huggan/crypto-gan). |
| |
|
| | Or try it yourself |
| |
|
| | ```python |
| | import tensorflow as tf |
| | import matplotlib.pyplot as plt |
| | from huggingface_hub import from_pretrained_keras |
| | |
| | seed = 42 |
| | n_images = 36 |
| | codings_size = 100 |
| | generator = from_pretrained_keras("huggan/crypto-gan") |
| | |
| | def generate(generator, seed): |
| | noise = tf.random.normal(shape=[n_images, codings_size], seed=seed) |
| | generated_images = generator(noise, training=False) |
| | |
| | fig = plt.figure(figsize=(10, 10)) |
| | for i in range(generated_images.shape[0]): |
| | plt.subplot(6, 6, i+1) |
| | plt.imshow(generated_images[i, :, :, :]) |
| | plt.axis('off') |
| | plt.savefig("samples.png") |
| | |
| | generate(generator, seed) |
| | ``` |
| |
|
| | ## Training data |
| |
|
| | For training, I used the 10000 CryptoPunks images. |
| |
|
| | ## Model Plot |
| |
|
| | <details> |
| | <summary>View Model Plot</summary> |
| |
|
| |  |
| |
|
| | </details> |
