sync with remote

app.py

@@ -10,6 +10,37 @@ from tensorflow.keras.losses import SparseCategoricalCrossentropy
 from io import StringIO
 import datetime
 
+import tensorboard
+from tensorboard import program
+
+try:
+    # Check if a GPU is available
+    gpu = len(tf.config.list_physical_devices('GPU')) > 0
+
+    if gpu:
+        st.write("GPU is available!")  # Inform the user
+        # Set TensorFlow to use the GPU if available (optional, usually automatic)
+        # You can specify which GPU if you have multiple:
+        # tf.config.set_visible_devices(tf.config.list_physical_devices('GPU')[0], 'GPU') # Use the first GPU
+        # or
+        # tf.config.experimental.set_memory_growth(tf.config.list_physical_devices('GPU')[0], True) # Memory growth for the first GPU
+        # or
+        # strategy = tf.distribute.MirroredStrategy(["GPU:0", "GPU:1"]) # Use multiple GPUs
+
+    else:
+        st.write("GPU is not available. Using CPU.")
+        os.environ["CUDA_VISIBLE_DEVICES"] = "-1"  # Force CPU usage (optional)
+
+except RuntimeError as e:
+    st.write(f"Error checking GPU: {e}")
+    os.environ["CUDA_VISIBLE_DEVICES"] = "-1" # Force CPU usage if there is a runtime error
+
+def run_tensorboard(log_dir):
+    # Start TensorBoard
+    tb = program.TensorBoard()
+    tb.configure(argv=[None, '--logdir', log_dir])
+    url = tb.launch()
+    return url
 
 # Constants for dataset information
 TRAIN_FILE = "train_images.tfrecords"
@@ -184,7 +215,112 @@ show_predictions(test_dataset)
 logdir = os.path.join("logs", datetime.datetime.now().strftime("%Y%m%d-%H%M%S"))
 tensorboard_callback = tf.keras.callbacks.TensorBoard(logdir, histogram_freq=1)
 
-if st.button("Run Model"):
+if st.button("Train Model"):
+    # setup and run the model
+    EPOCHS = 20
+    STEPS_PER_EPOCH = len(list(parsed_training_dataset))
+    VALIDATION_STEPS = 26
+
+    model_history = model.fit(train_dataset, epochs=EPOCHS,
+                            steps_per_epoch=STEPS_PER_EPOCH,
+                            validation_steps=VALIDATION_STEPS,
+                            validation_data=test_dataset,
+                            callbacks=[DisplayCallback(), tensorboard_callback])   
+
+    # output model statistics
+    loss = model_history.history['loss']
+    val_loss = model_history.history['val_loss']
+    accuracy = model_history.history['accuracy']
+    val_accuracy = model_history.history['val_accuracy']
+
+    epochs = range(EPOCHS)
+
+    st.title('Training and Validation Loss')  # Optional title for the Streamlit app
+
+    fig, ax = plt.subplots() # Create a figure and an axes object
+
+    ax.plot(epochs, loss, 'r', label='Training loss')
+    ax.plot(epochs, val_loss, 'bo', label='Validation loss')
+    ax.set_title('Training and Validation Loss') #Set title for the axes
+    ax.set_xlabel('Epoch')
+    ax.set_ylabel('Loss Value')
+    ax.set_ylim([0, 1])
+    ax.legend()
+
+    st.pyplot(fig)  # Display the plot in Streamlit
+
+if st.button("Evaluate Model"):
+    # Evaluate the model
+    evaluation_results = model.evaluate(test_dataset, verbose=0)  # Set verbose=0 to suppress console output
+    
+    # Assuming model.metrics_names provides labels for evaluation_results
+    results_dict = dict(zip(model.metrics_names, evaluation_results))
+    
+    st.subheader("Model Evaluation Results")
+    
+    # Display each metric and its corresponding value
+    for metric, value in results_dict.items():
+        st.write(f"**{metric.capitalize()}:** {value:.4f}")
+
+if st.button("Show TensorBoard"):
+    # Create a log directory for TensorBoard
+    log_dir = "logs"
+    if not os.path.exists(log_dir):
+        os.makedirs(log_dir)
+    
+    # Run TensorBoard
+    url = run_tensorboard(log_dir)
+    
+    # Display TensorBoard in an iframe
+    st.markdown(f"<iframe src='{url}' width='100%' height='800'></iframe>", unsafe_allow_html=True)
+
+if st.button("CNN"):
+    tf.keras.backend.clear_session()
+
+    inputs = tf.keras.Input(shape=(256, 256, 1), name="InputLayer")
+
+    x = tf.keras.layers.Conv2D(filters=100, kernel_size=5, strides=2, padding="same",
+                               activation="relu", name="Conv1")(inputs)
+    x = tf.keras.layers.MaxPool2D(pool_size=2, strides=2, padding="same")(x)
+
+    x = tf.keras.layers.Conv2D(filters=200, kernel_size=5, strides=2, padding="same",
+                               activation="relu", name="Conv2")(x)
+    x = tf.keras.layers.MaxPool2D(pool_size=2, strides=2, padding="same")(x)
+
+    x = tf.keras.layers.Conv2D(filters=300, kernel_size=3, strides=1, padding="same",
+                               activation="relu", name="Conv3")(x)
+    x = tf.keras.layers.Conv2D(filters=300, kernel_size=3, strides=1, padding="same",
+                               activation="relu", name="Conv4")(x)
+
+    x = tf.keras.layers.Conv2D(filters=2, kernel_size=1, strides=1, padding="same",
+                               activation="relu", name="Conv5")(x)
+
+    outputs = tf.keras.layers.Conv2DTranspose(filters=2, kernel_size=31, strides=16, 
+                                              padding="same", activation="softmax", 
+                                              name="UpSampling")(x)
+
+    model = tf.keras.Model(inputs=inputs, outputs=outputs, name="CNN_Segmentation")
+
+    model.compile(
+        optimizer=tf.keras.optimizers.Adam(),
+        loss=tf.keras.losses.SparseCategoricalCrossentropy(),
+        metrics=['accuracy']
+    )
+
+    # Capture the model summary
+    model_summary = StringIO()
+    model.summary(print_fn=lambda x: model_summary.write(x + '\n'))
+
+    # plot the model including the sizes of the model
+    tf.keras.utils.plot_model(model, show_shapes=True)
+
+    # show a predection, as an example
+    show_predictions(test_dataset)
+
+    # Initialize new directories for new task
+    logdir = os.path.join("logs", datetime.datetime.now().strftime("%Y%m%d-%H%M%S"))
+    tensorboard_callback = tf.keras.callbacks.TensorBoard(logdir, histogram_freq=1)
+
     # setup and run the model
     EPOCHS = 20
     STEPS_PER_EPOCH = len(list(parsed_training_dataset))
@@ -194,4 +330,26 @@ if st.button("Run Model"):
                             steps_per_epoch=STEPS_PER_EPOCH,
                             validation_steps=VALIDATION_STEPS,
                             validation_data=test_dataset,
-                            callbacks=[DisplayCallback(), tensorboard_callback])    
+                            callbacks=[DisplayCallback(), tensorboard_callback])    
+
+    # output model statistics
+    loss = model_history.history['loss']
+    val_loss = model_history.history['val_loss']
+    accuracy = model_history.history['accuracy']
+    val_accuracy = model_history.history['val_accuracy']
+
+    epochs = range(EPOCHS)
+
+    st.title('Training and Validation Loss')  # Optional title for the Streamlit app
+
+    fig, ax = plt.subplots() # Create a figure and an axes object
+
+    ax.plot(epochs, loss, 'r', label='Training loss')
+    ax.plot(epochs, val_loss, 'bo', label='Validation loss')
+    ax.set_title('Training and Validation Loss') #Set title for the axes
+    ax.set_xlabel('Epoch')
+    ax.set_ylabel('Loss Value')
+    ax.set_ylim([0, 1])
+    ax.legend()
+
+    st.pyplot(fig)  # Display the plot in Streamlit

logs/20250131-092418/train/events.out.tfevents.1738286658.MIS-DIRECTOR-CE.7000.0.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4ee6fdccada8310f63d3e4e4650dfbcc43861b9681f7aa8ef92aa91703616a7
+size 72247

logs/20250131-092418/validation/events.out.tfevents.1738286684.MIS-DIRECTOR-CE.7000.1.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea6904baf3931bec88aafd431437f0d451ae235c85b413f6685c5a5d66238c70
+size 6474

logs/20250131-094419/train/events.out.tfevents.1738287859.MIS-DIRECTOR-CE.21128.0.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20ea3a216d82b8fda2bb367d5ce0b9f7a528cac609a225ced8391561de895ff3
+size 72247

logs/20250131-094419/validation/events.out.tfevents.1738287886.MIS-DIRECTOR-CE.21128.1.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7d59970d19f8e6b8cf33137a9170223ccc0e373b4a0ca96bda438900ef4c57c
+size 6474

logs/20250131-121144/train/events.out.tfevents.1738296704.MIS-DIRECTOR-CE.21828.0.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae345409300b4881d3213deda821fefa0a2492c2db21adeabdcb251eadc0a728
+size 206319

logs/20250131-121144/validation/events.out.tfevents.1738296714.MIS-DIRECTOR-CE.21828.1.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb7e2a251dc434155e8b41b76470168b66be660f51bba4f5babad3ff861f7e45
+size 6474