File size: 2,838 Bytes
c50541c d8a8dd1 c50541c ad94d7e c50541c ad94d7e c50541c ad94d7e 58ed9ad ad94d7e c50541c ad94d7e c50541c ad94d7e c50541c ad94d7e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 |
import streamlit as st
from PIL import Image
import jax
import numpy as np
import jax.numpy as jnp # JAX NumPy
from flax.training import train_state # Useful dataclass to keep train state
from flax import linen as nn # Linen API
from huggingface_hub import HfFileSystem
from flax.serialization import msgpack_restore, from_state_dict
import os
import tensorflow as tf
hf_key = text_input = st.text_input("Access token")
class CNN(nn.Module):
@nn.compact
def __call__(self, x):
x = nn.Conv(features=32, kernel_size=(3, 3))(x)
x = nn.relu(x)
x = nn.avg_pool(x, window_shape=(2, 2), strides=(2, 2))
x = nn.Conv(features=64, kernel_size=(3, 3))(x)
x = nn.relu(x)
x = nn.avg_pool(x, window_shape=(2, 2), strides=(2, 2))
x = x.reshape((x.shape[0], -1)) # flatten
x = nn.Dense(features=256)(x)
x = nn.relu(x)
x = nn.Dense(features=16)(x)
x = nn.relu(x)
x = nn.Dense(features=2)(x)
return x
cnn = CNN()
params = cnn.init(jax.random.PRNGKey(0), jnp.ones([2, 50, 50, 3]))['params']
fs = HfFileSystem(token=hf_key)
with fs.open("PrakhAI/CatVsDog/checkpoint.msgpack", "rb") as f:
params = from_state_dict(params, msgpack_restore(f.read())["params"])
uploaded_files = st.file_uploader("Input images of cats or dogs (examples in files)", type=['jpg','png','tif'], accept_multiple_files=True)
if len(uploaded_files) == 0:
st.write("Please upload an image!")
else:
for uploaded_file in uploaded_files:
img = Image.open(uploaded_file)
st.image(img)
input = jnp.array(tf.cast(tf.image.resize(tf.convert_to_tensor(img), [50, 50]), tf.float32) / 255.)
st.write("Input shape: " + input.shape)
st.write("Model Prediction: " + cnn.apply({"params": params}, input))
st.write("Model Prediction type: " + type(cnn.apply({"params": params}, input)))
st.write("Model Prediction type dir: " + dir(cnn.apply({"params": params}, input)))
def gridify(kernel, grid, kernel_size, scaling=5, padding=1):
scaled_and_padded = np.pad(np.repeat(np.repeat(kernel, repeats=scaling, axis=0), repeats=scaling, axis=1), ((padding,),(padding,),(0,),(0,)), 'constant', constant_values=(-1,))
grid = np.pad(np.array(scaled_and_padded.reshape((kernel_size[0]*scaling+2*padding, kernel_size[1]*scaling+2*padding, 3, grid[0], grid[1])).transpose(3,0,4,1,2).reshape(grid[0]*(kernel_size[0]*scaling+2*padding), grid[1]*(kernel_size[1]*scaling+2*padding), 3)+1)*127., ((padding,),(padding,),(0,)), 'constant', constant_values=(0,))
st.image(Image.fromarray(grid.astype(np.uint8), mode="RGB"))
with st.expander("See first convolutional layer"):
gridify(params["Conv_0"]["kernel"], grid=(4,8), kernel_size=(3,3))
with st.expander("See second convolutional layer"):
print(params["Conv_1"]["kernel"].shape)
gridify(params["Conv_1"]["kernel"], grid=(64,96), kernel_size=(3,3))
|