%PDF-1.4 %âãÏÓ 1 0 obj << /Type /Catalog /Pages 2 0 R >> endobj 2 0 obj << /Type /Pages /Count 2 /Kids [5 0 R 7 0 R] >> endobj 3 0 obj << /Type /Font /Subtype /Type1 /BaseFont /Helvetica >> endobj 4 0 obj << /Type /Font /Subtype /Type1 /BaseFont /Helvetica-Bold >> endobj 5 0 obj << /Type /Page /Parent 2 0 R /MediaBox [0 0 595.28 841.89] /Resources << /Font << /F1 3 0 R /F2 4 0 R >> >> /Contents 6 0 R >> endobj 6 0 obj << /Length 5456 >> stream BT /F2 22 Tf 0.06 0.08 0.12 rg 1 0 0 1 46 789.89 Tm (Why Deep Learning Frameworks Require a GPU?) Tj ET BT /F2 11 Tf 0.72 0.14 0.18 rg 1 0 0 1 46 752.89 Tm (TechRounder PDF Edition) Tj ET BT /F1 9.5 Tf 0.36 0.39 0.46 rg 1 0 0 1 46 736.89 Tm (Live article: https://www.techrounder.com/technology/why-deep-learning-frameworks-require-a-gpu/) Tj ET q 0.82 0.85 0.9 RG 1 w 46 718.39 m 549.28 718.39 l S Q BT /F1 10 Tf 0.24 0.27 0.32 rg 1 0 0 1 46 706.39 Tm (By Vipin PG | Published January 20, 2022 | Updated January 4, 2026 | Format: Explainer | 3 min read) Tj ET BT /F2 13 Tf 0.72 0.14 0.18 rg 1 0 0 1 46 683.39 Tm (In brief) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 663.39 Tm (Deep Learning is a subfield of machine learning that attempts to mimic the human brain. Like the) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 648.39 Tm (human brain has neurons that transmit information and learn things, deep learning has a similar) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 633.39 Tm (structure and learns through an iterative process.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 608.39 Tm (Deep Learning is a subfield of machine learning that attempts to mimic the human brain. Like the) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 593.39 Tm (human brain has neurons that transmit information and learn things, deep learning has a similar) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 578.39 Tm (structure and learns through an iterative process.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 556.39 Tm (Deep learning is a neural network with three or more layers. These neural networks seek to imitate) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 541.39 Tm (the activity of the human brain by enabling it to learn from enormous quantities of data, although they) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 526.39 Tm (fall far short of replicating it. While a single-layer neural network may produce approximate) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 511.39 Tm (predictions, more hidden layers can improve and tune for accuracy.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 489.39 Tm (A deep learning model will help you perform the categorization tasks from text, sound, pictures, etc.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 474.39 Tm (Using labeled data and multilayer neural network topologies, deep learning models can achieve) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 459.39 Tm (maximum accuracy.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 437.39 Tm (Deep learning has numerous applications in transportation, mobile, television, communication devices,) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 422.39 Tm (health, medicine, etc. As a result, it has gotten a lot of press recently, and for a good cause.) Tj ET BT /F2 15 Tf 0.08 0.1 0.14 rg 1 0 0 1 46 394.39 Tm (Deep Learning is an Iterative Process) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 370.39 Tm (Deep learning implementation goes through an iterative cycle. To understand the analogy of deep) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 355.39 Tm (learning as an iterative process, consider a baby learning to walk. The learning process for the baby) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 340.39 Tm (takes place as standing, walking, falling, standing, walking, balancing, falling, and so on.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 318.39 Tm (Like this, a deep learning model makes a random prediction at first. It calculates the loss. Then, those) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 303.39 Tm (losses are backpropagated throughout the network such that weights and biases are adjusted to give) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 288.39 Tm (better results on the next run.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 266.39 Tm (In this way, it keeps on making mistakes but learns from them every time. One other thing that) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 251.39 Tm (distinguishes deep learning from classical machine learning is that deep learning learns features) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 236.39 Tm (from the inputs themselves. In classical machine learning techniques, feature extraction is a very) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 221.39 Tm (important step that is absent in deep learning.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 199.39 Tm (This is also why deep learning requires a large amount of data to extract features and learn from) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 184.39 Tm (them to make better predictions. Overall, each step is visited time and again, and deep learning is thus) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 169.39 Tm (an iterative process.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 147.39 Tm (The above statements suggest that deep learning requires more data, a longer training time, and) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 132.39 Tm (correspondingly a larger computation power. This is the reason why deep learning frameworks need) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 117.39 Tm (a GPU.) Tj ET BT /F2 15 Tf 0.08 0.1 0.14 rg 1 0 0 1 46 89.39 Tm (What Are GPUs?) Tj ET q 0.86 0.88 0.92 RG 1 w 46 42 m 549.28 42 l S Q BT /F1 8.4 Tf 0.42 0.45 0.5 rg 1 0 0 1 46 30 Tm (TechRounder | Page 1 of 2) Tj ET BT /F1 7.2 Tf 0.42 0.45 0.5 rg 1 0 0 1 46 19 Tm (https://www.techrounder.com/pdf/blog/why-deep-learning-frameworks-require-a-gpu.pdf) Tj ET endstream endobj 7 0 obj << /Type /Page /Parent 2 0 R /MediaBox [0 0 595.28 841.89] /Resources << /Font << /F1 3 0 R /F2 4 0 R >> >> /Contents 8 0 R >> endobj 8 0 obj << /Length 4349 >> stream BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 789.89 Tm (A graphics processing unit \(GPU\) is a specialized processor created to speed up the rendering of) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 774.89 Tm (visuals. In addition, GPUs can handle a large amount of data at once, making them ideal for machine) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 759.89 Tm (learning, video editing, and gaming. As a result, GPUs are an essential component of contemporary) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 744.89 Tm (computing.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 722.89 Tm (Computational science and AI are being transformed by GPU computing and high-performance) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 707.89 Tm (networking. For example, GPU developments have played a significant role in the current growth of) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 692.89 Tm (deep learning.) Tj ET BT /F2 15 Tf 0.08 0.1 0.14 rg 1 0 0 1 46 664.89 Tm (Benefits of Using GPUs) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 640.89 Tm (There are various frameworks like Tensorflow, Pytorch, etc., that can work on deep learning) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 625.89 Tm (algorithms. However, while working with these frameworks, GPUs can perform significantly faster) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 610.89 Tm (than the same performance. This is because while a CPU can do only a handful of operations at once,) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 595.89 Tm (the multiple GPU cores can perform thousands of operations at once.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 573.89 Tm (A task requiring a couple of hours to train on a CPU may only require 10-20 minutes to train on a GPU.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 558.89 Tm (GPUs save a lot of computation time, so they are very popular for deep learning applications.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 536.89 Tm (Deep learning algorithms are not interpretable. With multiple layers, large quantities of neurons, and) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 521.89 Tm (thousands of parameters to learn, it is hard to imagine how information gets propagated between) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 506.89 Tm (layers on the network. So, they learn things in a way that is beyond human imagination. Therefore,) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 491.89 Tm (deep learning frameworks require GPUs for optimal training and learning processes.) Tj ET BT /F2 15 Tf 0.08 0.1 0.14 rg 1 0 0 1 46 463.89 Tm (Hidden Costs) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 439.89 Tm (On the other hand, the cost of GPUs is indeed very high. It is expensive to set up your GPU server. But,) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 424.89 Tm (various companies provide cloud GPU use for deep learning. Such companies include mainly Amazon) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 409.89 Tm (AWS, Microsoft Azure, and Google GCP. There are plenty more. Also, a number of services like Google) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 394.89 Tm (Colab and Kaggle are free of cost.) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 372.89 Tm (Although they have their limitations, they are helpful when working in deep learning applications free) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 357.89 Tm (of cost. GPUs optimize the iterative training processes of deep learning applications, so deep learning) Tj ET BT /F1 11 Tf 0.14 0.16 0.2 rg 1 0 0 1 46 342.89 Tm (frameworks require a GPU.) Tj ET BT /F2 13 Tf 0.08 0.1 0.14 rg 1 0 0 1 46 314.89 Tm (References) Tj ET BT /F1 10 Tf 0.18 0.2 0.24 rg 1 0 0 1 46 294.89 Tm (1. techradar.com - news / computing-components -) Tj ET BT /F1 10 Tf 0.18 0.2 0.24 rg 1 0 0 1 46 281.39 Tm (https://www.techradar.com/news/computing-components/graphics-cards/how-gpus-are-made-1000923) Tj ET BT /F1 10 Tf 0.18 0.2 0.24 rg 1 0 0 1 46 263.89 Tm (2. developer.nvidia.com - deep-learning-frameworks - https://developer.nvidia.com/deep-learning-frameworks) Tj ET q 0.86 0.88 0.92 RG 1 w 46 42 m 549.28 42 l S Q BT /F1 8.4 Tf 0.42 0.45 0.5 rg 1 0 0 1 46 30 Tm (TechRounder | Page 2 of 2) Tj ET BT /F1 7.2 Tf 0.42 0.45 0.5 rg 1 0 0 1 46 19 Tm (https://www.techrounder.com/pdf/blog/why-deep-learning-frameworks-require-a-gpu.pdf) Tj ET endstream endobj xref 0 9 0000000000 65535 f 0000000015 00000 n 0000000064 00000 n 0000000127 00000 n 0000000197 00000 n 0000000272 00000 n 0000000414 00000 n 0000005921 00000 n 0000006063 00000 n trailer << /Size 9 /Root 1 0 R >> startxref 10463 %%EOF