Create app.py

app.py

@@ -0,0 +1,60 @@
+import streamlit as st
+from PIL import Image
+import jax
+import jax.numpy as jnp  # JAX NumPy
+import numpy as np
+from flax import linen as nn  # Linen API
+from huggingface_hub import HfFileSystem
+from flax.serialization import msgpack_restore, from_state_dict
+import time
+
+LATENT_DIM = 100
+
+class Generator(nn.Module):
+  @nn.compact
+  def __call__(self, latent, training=True):
+    x = nn.Dense(features=32)(latent)
+    # x = nn.BatchNorm(not training)(x)
+    x = nn.relu(x)
+    x = nn.Dense(features=2*2*256)(x)
+    x = nn.BatchNorm(not training)(x)
+    x = nn.relu(x)
+    x = nn.Dropout(0.5, deterministic=not training)(x)
+    x = x.reshape((x.shape[0], 2, 2, -1))
+    x4o = nn.ConvTranspose(features=3, kernel_size=(2, 2), strides=(2, 2))(x)
+    x4 = nn.ConvTranspose(features=128, kernel_size=(2, 2), strides=(2, 2))(x)
+    x4 = LocalResponseNorm()(x4)
+    # x4 = nn.BatchNorm(not training)(x4)
+    x8 = nn.relu(x4)
+    # x8 = nn.Dropout(0.5, deterministic=not training)(x8)
+    x8o = nn.ConvTranspose(features=3, kernel_size=(2, 2), strides=(2, 2))(x8)
+    x8 = nn.ConvTranspose(features=64, kernel_size=(2, 2), strides=(2, 2))(x8)
+    x8 = LocalResponseNorm()(x8)
+    # x8 = nn.BatchNorm(not training)(x8)
+    x16 = nn.relu(x8)
+    # x16 = nn.Dropout(0.5, deterministic=not training)(x16)
+    x16o = nn.ConvTranspose(features=3, kernel_size=(2, 2), strides=(2, 2))(x16)
+    x16 = nn.ConvTranspose(features=32, kernel_size=(2, 2), strides=(2, 2))(x16)
+    x16 = LocalResponseNorm()(x16)
+    # x16 = nn.BatchNorm(not training)(x16)
+    x32 = nn.relu(x16)
+    # x32 = nn.Dropout(0.5, deterministic=not training)(x32)
+    x32o = nn.ConvTranspose(features=3, kernel_size=(2, 2), strides=(2, 2))(x32)
+    return (nn.tanh(x32o), nn.tanh(x16o), nn.tanh(x8o), nn.tanh(x4o))
+
+generator = Generator()
+variables = generator.init(jax.random.PRNGKey(0), jnp.zeros([1, LATENT_DIM]), training=False)
+
+fs = HfFileSystem()
+with fs.open("PrakhAI/AIPlane/g_checkpoint.msgpack", "rb") as f:
+  g_state = from_state_dict(variables, msgpack_restore(f.read()))
+
+def sample_latent(key):
+  return jax.random.normal(key, shape=(1, LATENT_DIM))
+
+if st.button('Generate Plane'):
+  latents = sample_latent(jax.random.PRNGKey(int(1_000_000 * time.time())))
+  g_out = generator.apply({'params': g_state['params'], 'batch_stats': g_state['batch_stats']}, latents, training=False)
+  img = ((np.array(g_out)+1)*255./2.).astype(np.uint8)[0]
+  st.image(Image.fromarray(np.repeat(img, repeats=3, axis=2)))
+  st.write("The model's details are at https://huggingface.co/PrakhAI/AIPlane/blob/main/README.md")