Tensorflow
Download as PPTX, PDF8 likes1,630 views
TensorFlow is an open-source software library developed by Google for numerical computation and machine learning, particularly useful for deep learning applications. It operates using data flow graphs where nodes represent mathematical operations and edges represent data (tensors) that flow between them. The framework enables flexible deployment across CPUs and GPUs and is extensively utilized by various companies for AI and ML solutions.
Tensorflow
- 2. TensorFlow • Purpose: Machine Learning & deep Learning research • Open Source (By Google) • Numerical Computation using data flow graphs
- 3. Agenda • Artificial Intelligence • Machine learning • Deep Learning • Why would you use Deep Learning? • TensorFlow o TensorFlow Story o Tensorflow Model o Data flow graphs o Architecture o Portable & scalable o Programming model o Companies using TensorFlow
- 5. AI & ML & DL
- 6. ArtificialIntelligence • AI or artificial intelligence : is the simulation of human intelligence processes by machines, especially computer systems. • These processes include learning (the information and rules for using the information). • reasoning (using the rules to reach approximate or definite conclusions).
- 7. Machine learning • Machine learning is a type of artificial intelligence (AI) that provides computers with the ability to learn without being explicitly programmed. • Machine learning focuses on the development of computer programs that can change when exposed to new data.
- 8. Deep Learning • Deep Learning is a new area of Machine Learning research, which has been introduced with the objective of moving Machine Learning closer to one of its original goals • that is concerned with emulating the learning approach that human beings use to gain certain types of knowledge.
- 9. Why would you use Deep Learning?
- 10. Gaming
- 16. Self- driving
- 17. Android Camera Demo • Uses a Google Inception model to classify camera frames in real-time, displaying the top results in an overlay on the camera image.
- 19. TensorFlow • TensorFlow™ is an open source software library for numerical computation and machine learning. • Especially useful for Deep learning • For research and production • Developed by Google Brain Team
- 20. TensorFlow Story • TensorFlow was originally developed by researchers and engineers working on the Google Brain Team within Google's Machine Intelligence research organization for the purposes of conducting machine learning and deep neural networks research. • Most Most popular machine learning library. o Over 32,000 downlowds
- 21. Tensorflow Model • Big idea: Express a numeric computation as a graph. • Nodes = Mathematical Operations o Nodes in the graph represent mathematical operations, • Data Edges = Multi-Dimensional Arrays (Tensor) o communicated between them • Data (Tensors) flow through the graph
- 22. Data flow graphs • Computation is defined as directed a cyclic graph (GDA) o Optimize an objective function. • Graph is defined in high-level Language (python, C++,Go) • Graph is compiled and optimized • Graph is executed (parts or fully) On level devices (CPU, GPU).
- 23. Architecture • Core in C++ • Front end code python and C++. • The flexible architecture allows you to deploy computation to one or more CPUs or GPUs
- 24. Portable & scalable • deploy computation in a desktop, server, or mobile device with a single API
- 25. Programming model 1.Build a graph a. Graph contains parameter specifications, model architecture, optimization process, … b. Somewhere between 5 and 5000 lines 2.Initialize a session 3.Fetch and feed data with Session.run a. Compilation, optimization, etc. happens at this step — you probably won’t notice
- 26. Programming model (Build a graph) • Variables are 0-ary stateful nodes which output their current value. • Placeholders are 0-ary nodes whose value is fed in at execution time. • Mathematical operations: o MatMul: Multiply two matrix values. o Add: Add elementwise (with broadcasting). o ReLU: Activate with elementwise rectified linear function.
- 27. Programming model (Build a graph) • import tensorflow as tf b = tf.Variable(tf.zeros((100,))) W=tf.Variable(tf.random_uniform((784, 100), -1, 1)) x = tf.placeholder(tf.float32, (None, 784)) h_i = tf.nn.relu(tf.matmul(x, W) + b)
- 28. Programming model (Initialize & Run session) • a session: a binding to a particular execution context sess.run(fetches, feeds) Fetches: List of graph nodes. Return the outputs of these nodes. Feeds: Dictionary mapping from graph nodes to concrete values. Specifies the value of each graph node given in the dictionary.
- 29. Programming model (Build a graph) import numpy as np import tensorflow as tf b = tf.Variable(tf.zeros((100,))) W = tf.Variable(tf.random_uniform((784, 100), -1, 1)) x = tf.placeholder(tf.float32, (None, 784)) h_i = tf.nn.relu(tf.matmul(x, W) + b) -----Initial and Run Session------- sess = tf.Session() sess.run(tf.initialize_all_variables()) sess.run(h_i, {x: np.random.random(64, 784)})