Comprehensive Machine Learning Study Guide
### Keras
1. **Keras is a high-level neural networks API**, written in Python and capable
of running on top of TensorFlow.
2. Key Components:
- **Sequential Model**: For simple layer stacking.
- **Functional API**: For creating complex models like multi-input/output.
- **Model Training**: Methods like `model.fit`, `model.evaluate`, and
`model.predict`.
3. Important Layers:
- **Dense**: Fully connected layer.
- **Dropout**: Regularization layer to prevent overfitting.
- **Conv2D**: Convolution layer for image data.
- **LSTM**: Long Short-Term Memory for sequence data.
4. Example Code:
```python
from keras.models import Sequential
from keras.layers import Dense, Conv2D, Flatten, Dropout
model = Sequential([
Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 1)),
Flatten(),
Dense(128, activation='relu'),
Dropout(0.5),
Dense(10, activation='softmax')
])
model.compile(optimizer='adam', loss='categorical_crossentropy',
metrics=['accuracy'])
```
### PyTorch
1. **PyTorch is a deep learning framework** known for its flexibility and
dynamic computation graph.
2. Key Components:
- **`torch.nn`**: For building neural network layers.
- **`torch.optim`**: Optimizers like SGD and Adam.
- **Autograd**: Automatic differentiation for gradient computation.
3. Example Code:
```python
import torch
import torch.nn as nn
import torch.optim as optim
class SimpleNN(nn.Module):
def __init__(self):
super(SimpleNN, self).__init__()
self.fc1 = nn.Linear(784, 128)
self.relu = nn.ReLU()
self.fc2 = nn.Linear(128, 10)
def forward(self, x):
x = self.relu(self.fc1(x))
x = self.fc2(x)
return x
model = SimpleNN()
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.001)
```
### Key Differences Between Keras and PyTorch
1. **Ease of Use**: Keras is simpler for quick prototyping; PyTorch offers more
flexibility.
2. **Computation Graph**: PyTorch uses a dynamic computation graph, while
Keras (via TensorFlow) uses static graphs.
3. **Debugging**: PyTorch is easier to debug as it integrates seamlessly with
Python debuggers.
