CVxTz

from dotenv import load_dotenv
import os

load_dotenv()

import google.generativeai as genai
from data_model import (
    TouristLocation,
    ClimateType,
    ActivityType,

CVxTz

# First install https://github.com/yakupadakli/python-unsplash
# Unsplash API https://unsplash.com/documentation
import json
import os

from unsplash.api import Api
from unsplash.auth import Auth

with open('tokens.json', 'r') as f:
    data = json.load(f)

CVxTz

def get_model(n_class=5):
    inp = Input(shape=(187, 1))
    x = Convolution1D(64, kernel_size=5, activation=activations.relu, padding="valid")(
        inp
    )
    x = MaxPool1D(pool_size=4)(x)
    x = Convolution1D(64, kernel_size=3, activation=activations.relu, padding="valid")(
        x
    )
    x = MaxPool1D(pool_size=4)(x)

CVxTz

import numpy as np


def kl_loss(y_true, y_pred):
    return np.sum(y_true * np.log(y_true / y_pred))


y_true = np.array([0.1, 0.5, 0.2, 0.15, 0.05])
y_pr_1 = np.array([0.15, 0.4, 0.2, 0.2, 0.05])
y_pr_2 = np.array([0.5, 0.1, 0.1, 0.2, 0.1])

CVxTz

for i, step_size in enumerate(step_sizes):
    
    X_used = X[used_samples, ...]
    Y_used = Y[used_samples, ...]

    X_unused = X[unused_samples, ...]
    Y_unused = Y[unused_samples, ...]

    model = get_model()

CVxTz

for i, step_size in enumerate(step_sizes):
    
    X_used = X[used_samples, ...]
    Y_used = Y[used_samples, ...]

    X_unused = X[unused_samples, ...]
    Y_unused = Y[unused_samples, ...]

    model = get_model()

CVxTz

X_used = X[used_samples, ...]
Y_used = Y[used_samples, ...]

X_unused = X[unused_samples, ...]
Y_unused = Y[unused_samples, ...]

model = get_model()

model.fit(X_used, Y_used, epochs=45, verbose=1, batch_size=32)

CVxTz

step_sizes = [128]*20 # Succesive batch sizes

unused_samples = list(range(X.shape[0])) # Indexes of samples that were not used for training yet
step = np.random.choice(unused_samples, size=512).tolist() # First batch 
used_samples = step # Indexes of samples that were used for training
unused_samples = list(set(unused_samples) - set(step))

CVxTz

def get_graph_embedding_model(n_nodes):
    in_1 = Input((1,))
    in_2 = Input((1,))
    emb = Embedding(n_nodes, 100, name="node1")
    x1 = emb(in_1)
    x2 = emb(in_2)
    x1 = Flatten()(x1)
    x1 = Dropout(0.1)(x1)
    x2 = Flatten()(x2)
    x2 = Dropout(0.1)(x2)

CVxTz

def get_features_only_model(n_features, n_classes):
    in_ = Input((n_features,))
    x = Dense(10, activation="relu", kernel_regularizer=l1(0.001))(in_)
    x = Dropout(0.5)(x)
    x = Dense(n_classes, activation="softmax")(x)

    model = Model(in_, x)

    model.compile(loss="sparse_categorical_crossentropy", metrics=['acc'], optimizer="adam")