From 0ce582250d53e37884762f4e357023396fdf58bd Mon Sep 17 00:00:00 2001
From: coolneng <akasroua@gmail.com>
Date: Sun, 6 Jun 2021 00:20:03 +0200
Subject: [PATCH] Implement the training loop and metrics evaluation
---
src/model.py | 60 +++++++++++++++++++++++++++++++++++++++++++++-------
1 file changed, 52 insertions(+), 8 deletions(-)
diff --git a/src/model.py b/src/model.py
index 6c43f2c..c73104a 100644
--- a/src/model.py
+++ b/src/model.py
@@ -1,28 +1,35 @@
+from random import seed
+
from tensorflow.keras import Model, Sequential, layers
+from tensorflow.keras.callbacks import TensorBoard
+from tensorflow.keras.losses import sparse_categorical_crossentropy
+from tensorflow.keras.optimizers import Adam
from tensorflow.keras.regularizers import l2
+from tensorflow.random import set_seed
-from constants import BASES
+from constants import *
+from preprocessing import dataset_creation
-def build_model(hyper_parameters) -> Model:
+def build_model() -> Model:
"""
- Builds the CNN model
+ Build the CNN model
"""
- return Sequential(
+ model = Sequential(
[
# Two convolutions + maxpooling blocks
layers.Conv1D(
filters=16,
kernel_size=5,
activation="relu",
- kernel_regularizer=l2(hyper_parameters.l2),
+ kernel_regularizer=l2(L2),
),
layers.MaxPool1D(pool_size=3, strides=1),
layers.Conv1D(
filters=16,
kernel_size=3,
activation="relu",
- kernel_regularizer=l2(hyper_parameters.l2),
+ kernel_regularizer=l2(L2),
),
layers.MaxPool1D(pool_size=3, strides=1),
# Flatten the input volume
@@ -31,16 +38,53 @@ def build_model(hyper_parameters) -> Model:
layers.Dense(
units=16,
activation="relu",
- kernel_regularizer=l2(hyper_parameters.l2),
+ kernel_regularizer=l2(L2),
),
layers.Dropout(rate=0.3),
layers.Dense(
units=16,
activation="relu",
- kernel_regularizer=l2(hyper_parameters.l2),
+ kernel_regularizer=l2(L2),
),
layers.Dropout(rate=0.3),
# Output layer with softmax activation
layers.Dense(units=len(BASES), activation="softmax"),
]
)
+ model.compile(
+ optimizer=Adam(LEARNING_RATE),
+ loss=sparse_categorical_crossentropy,
+ metrics=["accuracy"],
+ )
+ return model
+
+
+def show_metrics(model, eval_dataset, test_dataset) -> None:
+ """
+ Show the model metrics
+ """
+ eval_metrics = model.evaluate(eval_dataset, verbose=0)
+ test_metrics = model.evaluate(test_dataset, verbose=0)
+ print(f"Final eval metrics - loss: {eval_metrics[0]} - accuracy: {eval_metrics[1]}")
+ print(f"Final test metrics - loss: {test_metrics[0]} - accuracy: {test_metrics[1]}")
+
+
+def run(data_file, label_file, seed_value=42) -> None:
+ """
+ Create a dataset, a model and runs training and evaluation on it
+ """
+ seed(seed_value)
+ set_seed(seed)
+ train_data, eval_data, test_data = dataset_creation(data_file, label_file)
+ tensorboard = TensorBoard(log_dir=LOG_DIR, histogram_freq=1, profile_batch=0)
+ model = build_model()
+ print("Training the model")
+ model.fit(
+ train_data,
+ epochs=EPOCHS,
+ validation_data=eval_data,
+ callbacks=[tensorboard],
+ verbose=0,
+ )
+ print("Training complete. Obtaining final metrics...")
+ show_metrics(model, eval_data, test_data)