model = keras.Sequential([ keras.layers.Dense(128, activation='relu', input_shape=(784,)), keras.layers.Dense(10, activation='softmax') ])

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


## Core Concepts

### Keras Functional API

```python
from tensorflow import keras
from tensorflow.keras import layers

inputs = keras.Input(shape=(784,))
x = layers.Dense(256, activation='relu')(inputs)
x = layers.Dropout(0.2)(x)
x = layers.Dense(128, activation='relu')(x)
outputs = layers.Dense(10, activation='softmax')(x)

model = keras.Model(inputs, outputs, name='classifier')

Custom Training

@tf.function
def train_step(x, y):
    with tf.GradientTape() as tape:
        predictions = model(x, training=True)
        loss = loss_fn(y, predictions)

    gradients = tape.gradient(loss, model.trainable_variables)
    optimizer.apply_gradients(zip(gradients, model.trainable_variables))
    return loss

for epoch in range(epochs):
    for x_batch, y_batch in dataset:
        loss = train_step(x_batch, y_batch)

Common Patterns

Data Pipeline

dataset = tf.data.Dataset.from_tensor_slices((x_train, y_train))
dataset = dataset.shuffle(buffer_size=1024)
dataset = dataset.batch(32)
dataset = dataset.prefetch(tf.data.AUTOTUNE)

Model Saving

# SavedModel format
model.save('saved_model/my_model')

# Load
loaded_model = keras.models.load_model('saved_model/my_model')

# TensorFlow Lite
converter = tf.lite.TFLiteConverter.from_saved_model('saved_model/my_model')
tflite_model = converter.convert()

Best Practices

Do:

Use @tf.function for performance
Use tf.data for data pipelines
Enable mixed precision training
Profile with TensorBoard

Don't:

Use Python loops in tf.function
Create tensors inside training loops
Ignore eager vs graph mode
Skip model validation

Troubleshooting

tensorflow

TensorFlow

When to Use

Quick Start

More from g1joshi/agent-skills

template

mariadb

claude

javascript

typescript

python