18 Deep Learning Framework: Pytorch

  18 Deep Learning Framework: Pytorch

1. What is PyTorch?

PyTorch is an open-source deep learning framework developed by Meta (Facebook) AI. It’s widely used for building neural networks and machine learning models.

Key features:

1. Dynamic Computation Graphs (Define-by-Run)
• Unlike older frameworks (like TensorFlow 1.x), PyTorch builds the computation graph on the fly, which makes debugging and experimentation easier.
2. GPU Acceleration
• PyTorch can perform fast computations using CUDA-enabled NVIDIA GPUs, making training large models much faster.
3. Pythonic and Flexible
• It feels like regular Python code, which is intuitive for developers and researchers.
4. Deep Learning Ready
• Supports autograd (automatic differentiation), neural network modules, optimizers, and pre-trained models.

 

1. What is PyTorch?

PyTorch is an open-source deep learning framework developed by Meta (Facebook) AI. It’s widely used for building neural networks and machine learning models.

Key features:

1. Dynamic Computation Graphs (Define-by-Run)
• Unlike older frameworks (like TensorFlow 1.x), PyTorch builds the computation graph on the fly, which makes debugging and experimentation easier.
2. GPU Acceleration
• PyTorch can perform fast computations using CUDA-enabled NVIDIA GPUs, making training large models much faster.
3. Pythonic and Flexible
• It feels like regular Python code, which is intuitive for developers and researchers.
4. Deep Learning Ready
• Supports autograd (automatic differentiation), neural network modules, optimizers, and pre-trained models.

---

2. What are Tensors in PyTorch?

A tensor is the core data structure in PyTorch. It’s essentially a multi-dimensional array, similar to NumPy arrays, but with GPU support.

Think of tensors as the building blocks for inputs, weights, and outputs in neural networks.

---

Tensor Basics

Feature	Description
0D Tensor	A single number (scalar), e.g., tensor(5)
1D Tensor	A vector, e.g., [1, 2, 3]
2D Tensor	A matrix, e.g., [[1,2],[3,4]]
3D+ Tensor	Higher-dimensional data (like images, batch of images, video frames)

---

PyTorch Tensor Example

import torch

# 1D Tensor
a = torch.tensor([1, 2, 3])
print(a)  # tensor([1, 2, 3])

# 2D Tensor (Matrix)
b = torch.tensor([[1, 2], [3, 4]])
print(b)  # tensor([[1, 2], [3, 4]])

# Move tensor to GPU if available
if torch.cuda.is_available():
    b = b.to('cuda')

---

Important Features of Tensors

1. Operations: You can do math, linear algebra, and broadcasting just like NumPy.
2. Gradients: Tensors can track operations for automatic differentiation (requires_grad=True).
3. GPU Support: Tensors can be stored and computed on CPU or GPU.

---

✅ In short:

• PyTorch = a framework to build deep learning models.
• Tensors = multi-dimensional arrays that PyTorch uses to store and compute data efficiently, often on GPUs.