title
6.1: Introduction to TensorFlow.js - Intelligence and Learning

description
TensorFlow.js is a WebGL accelerated, browser based JavaScript library for training and deploying ML models. 🎥 Next Video: https://youtu.be/D-XzAeVvMkg 🔗 TensorFlow.js: https://js.tensorflow.org/ 🔗 What is a tensor?: https://js.tensorflow.org/tutorials/core-concepts.html 🔗 ITP's ML5.js library: https://ml5js.org 🔗 Keras: https://keras.io/ 🎥 My Neural Networks series: https://www.youtube.com/playlist?list=PLRqwX-V7Uu6aCibgK1PTWWu9by6XFdCfh 🎥 Machine Learning in JavaScript (TensorFlow Dev Summit 2018): https://youtu.be/YB-kfeNIPCE 🚂 The Coding Train website: http://thecodingtrain.com/ 💖 Support this channel on Patreon: https://patreon.com/codingtrain 🛒 To buy Coding Train merchandise: https://www.designbyhumans.com/shop/codingtrain/ 📚 Book recommendations: https://www.amazon.com/shop/thecodingtrain 💻 https://github.com/CodingTrain/website 🎥 For More Intelligence and Learning videos: https://www.youtube.com/playlist?list=PLRqwX-V7Uu6YJ3XfHhT2Mm4Y5I99nrIKX 🎥 For More Coding Challenges: https://www.youtube.com/playlist?list=PLRqwX-V7Uu6ZiZxtDDRCi6uhfTH4FilpH 🔗 https://p5js.org/ 🔗 https://processing.org 📄 Code of Conduct: https://github.com/CodingTrain/Code-of-Conduct

detail
{'title': '6.1: Introduction to TensorFlow.js - Intelligence and Learning', 'heatmap': [{'end': 206.918, 'start': 148.136, 'weight': 0.706}, {'end': 873.354, 'start': 826.144, 'weight': 0.736}, {'end': 979.711, 'start': 946.195, 'weight': 0.753}], 'summary': 'Introduces tensorflow.js as the foundation for future content, covering its relationship to machine learning and highlighting its adoption by the larger tensorflow project. it also discusses tensorflow in python, its relevance to data scientists, and the integration of tensorflow.js with keras and the ml5 library.', 'chapters': [{'end': 232.292, 'segs': [{'end': 97.581, 'src': 'embed', 'start': 25.552, 'weight': 0, 'content': [{'end': 27.333, 'text': 'comprehensive, but a few demonstrations.', 'start': 25.552, 'duration': 1.781}, {'end': 31.194, 'text': 'I built a little neural network library in JavaScript and went through some matrix math stuff.', 'start': 27.373, 'duration': 3.821}, {'end': 33.355, 'text': "So I'm ready finally at the time.", 'start': 31.534, 'duration': 1.821}, {'end': 40.677, 'text': "I've been saying this all along that I'm making this neural network library and kind of looking at how things work and trying to make some creative examples.", 'start': 33.635, 'duration': 7.042}, {'end': 51.681, 'text': 'But later, eventually, someday, I will use a more robust, thoughtful, well-designed framework as the guts,', 'start': 41.037, 'duration': 10.644}, {'end': 55.805, 'text': 'as the foundation from which I will build all of my examples and projects.', 'start': 51.681, 'duration': 4.124}, {'end': 58.107, 'text': "I'll use somebody else's machine learning code.", 'start': 56.165, 'duration': 1.942}, {'end': 62.751, 'text': "And so today is the day that I'm going to start talking about doing that.", 'start': 58.547, 'duration': 4.204}, {'end': 68.956, 'text': 'And the foundation that I will be using for almost all of these videos is a project called TensorFlow.js.', 'start': 63.151, 'duration': 5.805}, {'end': 77.359, 'text': "Let's discuss for a moment where did TensorFlow.js come from.", 'start': 71.792, 'duration': 5.567}, {'end': 81.564, 'text': 'So you might have heard of something called TensorFlow.', 'start': 78.22, 'duration': 3.344}, {'end': 87.972, 'text': 'Where to write this down? Ten-ser-flow.', 'start': 83.807, 'duration': 4.165}, {'end': 94.197, 'text': 'Now, first of all, you might even be asking yourself huh, why is it even called TensorFlow?', 'start': 89.011, 'duration': 5.186}, {'end': 97.581, 'text': 'What is this thing called a tensor?', 'start': 94.638, 'duration': 2.943}], 'summary': 'The speaker discusses building a neural network library in javascript, and plans to use tensorflow.js in future projects.', 'duration': 72.029, 'max_score': 25.552, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD025552.jpg'}, {'end': 180.544, 'src': 'embed', 'start': 148.136, 'weight': 3, 'content': [{'end': 157.665, 'text': 'And a two-dimensional matrix might have a grid of numbers, like 3, 4, 1, 5.', 'start': 148.136, 'duration': 9.529}, {'end': 169.675, 'text': 'So a tensor is a structure, a data structure essentially, that really can store any n-dimensional version of these types of things.', 'start': 157.665, 'duration': 12.01}, {'end': 180.544, 'text': "And because the building blocks of any machine learning algorithm are matrices of numbers, this idea of tensor flow, let's flow with the tensors.", 'start': 170.035, 'duration': 10.509}], 'summary': 'Tensors can store n-dimensional data for machine learning, like 3d, 4d, and 5d arrays.', 'duration': 32.408, 'max_score': 148.136, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0148136.jpg'}, {'end': 206.918, 'src': 'heatmap', 'start': 148.136, 'weight': 0.706, 'content': [{'end': 157.665, 'text': 'And a two-dimensional matrix might have a grid of numbers, like 3, 4, 1, 5.', 'start': 148.136, 'duration': 9.529}, {'end': 169.675, 'text': 'So a tensor is a structure, a data structure essentially, that really can store any n-dimensional version of these types of things.', 'start': 157.665, 'duration': 12.01}, {'end': 180.544, 'text': "And because the building blocks of any machine learning algorithm are matrices of numbers, this idea of tensor flow, let's flow with the tensors.", 'start': 170.035, 'duration': 10.509}, {'end': 185.307, 'text': 'Insert animation of me flowing down the river of tensors.', 'start': 182.105, 'duration': 3.202}, {'end': 188.288, 'text': 'Will that happen in post-production? I seriously doubt it.', 'start': 186.107, 'duration': 2.181}, {'end': 191.51, 'text': 'This is where the name TensorFlow comes from.', 'start': 189.969, 'duration': 1.541}, {'end': 198.634, 'text': "So TensorFlow is Google's open source machine learning library.", 'start': 191.61, 'duration': 7.024}, {'end': 206.918, 'text': "It is written, you might be surprised to hear this because you might think, ah, TensorFlow, it's Python, right? Well, yes, kind of.", 'start': 198.654, 'duration': 8.264}], 'summary': "Tensorflow is google's open source machine learning library for n-dimensional matrices.", 'duration': 58.782, 'max_score': 148.136, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0148136.jpg'}, {'end': 244.371, 'src': 'embed', 'start': 208.297, 'weight': 4, 'content': [{'end': 210.879, 'text': "I'm sure there are people watching this who know more about me.", 'start': 208.297, 'duration': 2.582}, {'end': 216.082, 'text': "So if you're watching the recorded version of this, check the video's description for all of the corrections.", 'start': 211.239, 'duration': 4.843}, {'end': 217.282, 'text': "I'll go look for it.", 'start': 216.582, 'duration': 0.7}, {'end': 219.404, 'text': "I'll try to make any corrections at the end.", 'start': 217.322, 'duration': 2.082}, {'end': 223.466, 'text': 'But the TensorFlow is actually a library written in C++.', 'start': 219.424, 'duration': 4.042}, {'end': 232.292, 'text': 'It is a low-level C++ library with a lot of functionality for doing machine learning.', 'start': 224.507, 'duration': 7.785}, {'end': 244.371, 'text': "Now, the reason why you might have thought to yourself, oh, Isn't it Python? Well, there simply is a sort of bindings for Python, so to speak.", 'start': 233.532, 'duration': 10.839}], 'summary': 'Tensorflow is a c++ library with python bindings for machine learning.', 'duration': 36.074, 'max_score': 208.297, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0208297.jpg'}], 'start': 0.389, 'title': 'Neural networks and tensorflow.js', 'summary': 'Covers completion of a tutorial on neural networks and plans for using a more robust framework, as well as introduces tensorflow.js as the foundation for future videos, explaining tensors and its relationship to machine learning, and highlighting tensorflow as an open source library.', 'chapters': [{'end': 62.751, 'start': 0.389, 'title': 'Neural network library and framework', 'summary': 'Discusses the completion of a tutorial on neural networks and basic machine learning, along with the plan to use a more robust framework for future projects.', 'duration': 62.362, 'highlights': ['The speaker completed a set of tutorials on neural networks and basic machine learning, along with building a neural network library in JavaScript.', 'The speaker plans to use a more robust, thoughtful, well-designed framework as the foundation for future examples and projects.', "The speaker will eventually use somebody else's machine learning code for future projects."]}, {'end': 232.292, 'start': 63.151, 'title': 'Introduction to tensorflow.js', 'summary': 'Introduces tensorflow.js as the foundation for the following videos, explaining the concept of tensors and its relationship to machine learning, and highlighting that tensorflow is an open source machine learning library written in c++.', 'duration': 169.141, 'highlights': ['TensorFlow.js is the foundation for the upcoming videos, serving as the basis for the discussion on tensors and their significance in machine learning.', 'A tensor is a data structure that can store any n-dimensional version of scalars, vectors, and matrices, which are the building blocks of machine learning algorithms.', 'TensorFlow, despite its association with Python, is actually a low-level C++ library with extensive functionality for machine learning.']}], 'duration': 231.903, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0389.jpg', 'highlights': ['TensorFlow.js is the foundation for upcoming videos, discussing tensors and their significance in machine learning.', 'The speaker plans to use a more robust, thoughtful, well-designed framework as the foundation for future examples and projects.', 'The speaker completed tutorials on neural networks and basic machine learning, along with building a neural network library in JavaScript.', 'A tensor is a data structure that can store any n-dimensional version of scalars, vectors, and matrices, the building blocks of machine learning algorithms.', 'TensorFlow, despite its association with Python, is actually a low-level C++ library with extensive functionality for machine learning.', "The speaker will eventually use somebody else's machine learning code for future projects."]}, {'end': 759.474, 'segs': [{'end': 275.122, 'src': 'embed', 'start': 233.532, 'weight': 1, 'content': [{'end': 244.371, 'text': "Now, the reason why you might have thought to yourself, oh, Isn't it Python? Well, there simply is a sort of bindings for Python, so to speak.", 'start': 233.532, 'duration': 10.839}, {'end': 246.752, 'text': 'Python, a wrapper, so to speak, for Python.', 'start': 244.811, 'duration': 1.941}, {'end': 254.195, 'text': "So Python being a programming language that's primarily used, not primarily used, but is very popular in the world of data science.", 'start': 246.772, 'duration': 7.423}, {'end': 261.338, 'text': "It makes sense if you're a data scientist and working with data and you want to do some stuff with machine learning,", 'start': 254.715, 'duration': 6.623}, {'end': 265.059, 'text': "that you would and you're already in Python you'd want to be able to access something like TensorFlow.", 'start': 261.338, 'duration': 3.721}, {'end': 275.122, 'text': "Most all and every example that you would see working with TensorFlow is you're just kind of operating the low-level TensorFlow stuff from Python.", 'start': 266.88, 'duration': 8.242}], 'summary': 'Python is popular in data science for tensorflow, a machine learning platform.', 'duration': 41.59, 'max_score': 233.532, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0233532.jpg'}, {'end': 340.991, 'src': 'embed', 'start': 317.435, 'weight': 3, 'content': [{'end': 324.754, 'text': 'but that aside, This project, TensorFlow has been around for quite a while.', 'start': 317.435, 'duration': 7.319}, {'end': 327.339, 'text': "Let me go look and find out how long it's been around, and then I'll come back.", 'start': 324.814, 'duration': 2.525}, {'end': 330.305, 'text': "OK, so I'm back.", 'start': 328.902, 'duration': 1.403}, {'end': 332.668, 'text': 'I had to look that up.', 'start': 331.468, 'duration': 1.2}, {'end': 334.409, 'text': 'TensorFlow was open sourced in 2015.', 'start': 332.748, 'duration': 1.661}, {'end': 340.991, 'text': 'So TensorFlow actually is a project, according to Wikipedia, started in 2011.', 'start': 334.409, 'duration': 6.582}], 'summary': 'Tensorflow, an open source project, started in 2011 and was officially released in 2015.', 'duration': 23.556, 'max_score': 317.435, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0317435.jpg'}, {'end': 709.191, 'src': 'embed', 'start': 662.249, 'weight': 0, 'content': [{'end': 665.711, 'text': 'There happens to be something in the world of JavaScript called WebGL.', 'start': 662.249, 'duration': 3.462}, {'end': 678.369, 'text': "WebGL is the browser's interface to OpenGL, for doing operations on the graphics card, for drawing and making graphic stuff happen in the browser.", 'start': 667.032, 'duration': 11.337}, {'end': 684.933, 'text': 'So if the math operations of TensorFlow in C++ can run on the GPU,', 'start': 678.809, 'duration': 6.124}, {'end': 692.398, 'text': "why can't the math operations inside of this thing called deeplearn.js run via the GPU via WebGL?", 'start': 684.933, 'duration': 7.465}, {'end': 698.363, 'text': "So that's really the magic in my mind of what was accomplished with this original project called deeplearn.js.", 'start': 692.418, 'duration': 5.945}, {'end': 701.605, 'text': "So let's go look at that website for deeplearn.js for a second.", 'start': 698.543, 'duration': 3.062}, {'end': 709.191, 'text': 'On March 30th, less than a month ago, deeplearn.js became adopted.', 'start': 702.666, 'duration': 6.525}], 'summary': 'Webgl enables gpu acceleration for deeplearn.js, adopted on march 30th.', 'duration': 46.942, 'max_score': 662.249, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0662249.jpg'}], 'start': 233.532, 'title': 'Tensorflow in python and tensorflow.js for data science and machine learning', 'summary': 'Delves into the use of tensorflow in python, emphasizing its relevance to data scientists. it also explores the evolution of tensorflow into tensorflow.js, highlighting its adoption by the larger tensorflow project as of less than a month ago.', 'chapters': [{'end': 275.122, 'start': 233.532, 'title': 'Tensorflow in python for data science', 'summary': 'Discusses the use of tensorflow in python for data science and machine learning, highlighting its popularity and relevance to data scientists, as well as its integration with python for low-level operations.', 'duration': 41.59, 'highlights': ["Python's popularity in data science and machine learning, making it a common choice for working with TensorFlow", 'Integration of TensorFlow with Python for low-level operations, particularly in the context of machine learning and accessing data']}, {'end': 759.474, 'start': 275.222, 'title': 'Tensorflow.js and javascript machine learning', 'summary': 'Discusses the evolution of tensorflow into tensorflow.js, a javascript library for machine learning, its compatibility with webgl for gpu acceleration, and its adoption by the larger tensorflow project as of march 30th, less than a month ago.', 'duration': 484.252, 'highlights': ['TensorFlow open sourced in 2015, started in 2011 TensorFlow was open sourced in 2015 under the Apache license, after being used as a proprietary machine learning library at Google since 2011.', 'Introduction of deeplearn.js adopted by TensorFlow as TensorFlow.js The speculative project deeplearn.js, focused on machine learning in JavaScript, was adopted by the larger TensorFlow project and became TensorFlow.js as of March 30th, less than a month ago.', "Compatibility of deeplearn.js/TensorFlow.js with WebGL for GPU acceleration deeplearn.js and subsequently TensorFlow.js leverage WebGL for GPU acceleration, enabling math operations to run via the GPU, aligning with the capabilities of TensorFlow's math operations in C++."]}], 'duration': 525.942, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0233532.jpg', 'highlights': ['TensorFlow.js adopted deeplearn.js as of March 30th, less than a month ago', "Python's popularity in data science and machine learning", 'Integration of TensorFlow with Python for low-level operations', 'TensorFlow open sourced in 2015, started in 2011', 'Compatibility of deeplearn.js/TensorFlow.js with WebGL for GPU acceleration']}, {'end': 1129.839, 'segs': [{'end': 883.683, 'src': 'heatmap', 'start': 818.362, 'weight': 0, 'content': [{'end': 826.144, 'text': 'in terms of working with machine learning operations, TensorFlow is a low-level library to do the actual matrix,', 'start': 818.362, 'duration': 7.782}, {'end': 833.605, 'text': 'math and Gradient descent learning training algorithms all yourself written into the code Yourself.', 'start': 826.144, 'duration': 7.461}, {'end': 836.566, 'text': "with TensorFlow it's common operations that are implemented for you.", 'start': 833.605, 'duration': 2.961}, {'end': 839.147, 'text': 'But this is really a low-level control of the algorithm itself.', 'start': 836.566, 'duration': 2.581}, {'end': 844.668, 'text': 'you could invent new machine learning models by writing them in TensorFlow yourself.', 'start': 839.147, 'duration': 5.521}, {'end': 851.956, 'text': 'then, In between, that there previously was, there still is, sorry, this project called Keras.', 'start': 844.668, 'duration': 7.288}, {'end': 853.878, 'text': 'Keras Keras? Keras.', 'start': 852.557, 'duration': 1.321}, {'end': 854.678, 'text': "I don't know how to pronounce it.", 'start': 853.898, 'duration': 0.78}, {'end': 857.32, 'text': 'Keras I always think everything is French for some reason.', 'start': 855.119, 'duration': 2.201}, {'end': 859.062, 'text': "I don't know why.", 'start': 857.34, 'duration': 1.722}, {'end': 864.966, 'text': "So, and that's probably not even a French pronunciation of that word, but that aside, sorry.", 'start': 859.502, 'duration': 5.464}, {'end': 866.107, 'text': 'Sorry that you had to watch that.', 'start': 865.206, 'duration': 0.901}, {'end': 873.354, 'text': 'Keras was meant to be a higher level project API built on top of tensorflow,', 'start': 867.148, 'duration': 6.206}, {'end': 883.683, 'text': 'and in fact Keras was actually originally designed to be a higher-level API that could sit on top of a variety of other low-level machine learning frameworks.', 'start': 873.354, 'duration': 10.329}], 'summary': 'Tensorflow is a low-level library for machine learning, while keras is a higher-level api built on top of it.', 'duration': 65.321, 'max_score': 818.362, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0818362.jpg'}, {'end': 931.305, 'src': 'embed', 'start': 903.165, 'weight': 1, 'content': [{'end': 911.009, 'text': 'But Keras, the idea of Keras is that you could kind of write your code, make a machine learning thing, and it could operate on top of any of these.', 'start': 903.165, 'duration': 7.844}, {'end': 917.152, 'text': 'So Keras, however, more recently became part of the TensorFlow project itself.', 'start': 911.369, 'duration': 5.783}, {'end': 923.178, 'text': 'And so Keras is actually, and TensorFlow are linked together.', 'start': 919.094, 'duration': 4.084}, {'end': 928.042, 'text': "So this is a higher level API that's built on top of TensorFlow.", 'start': 923.478, 'duration': 4.564}, {'end': 931.305, 'text': 'And it exists as part of TensorFlow.js.', 'start': 928.142, 'duration': 3.163}], 'summary': 'Keras is now part of tensorflow, providing a higher level api for machine learning.', 'duration': 28.14, 'max_score': 903.165, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0903165.jpg'}, {'end': 979.711, 'src': 'heatmap', 'start': 946.195, 'weight': 0.753, 'content': [{'end': 949.357, 'text': "although I'm going to start with a few that just look at the core stuff,", 'start': 946.195, 'duration': 3.162}, {'end': 952.519, 'text': "because it's kind of important to have a sense of what that is and how that works.", 'start': 949.357, 'duration': 3.162}, {'end': 957.682, 'text': 'but layers, so layers in TensorFlow.js is the equivalent of this thing called Keras now.', 'start': 952.519, 'duration': 5.163}, {'end': 966.417, 'text': "So a project that's being developed here at New York University with some collaborators from ITP and guests and researchers and students,", 'start': 958.8, 'duration': 7.617}, {'end': 967.58, 'text': 'is a project called ML5..', 'start': 966.417, 'duration': 1.163}, {'end': 979.711, 'text': "The five in ML5 is an homage homage that's French right to the five in P5 in the sense that I mean this is flawed for many reasons.", 'start': 969.264, 'duration': 10.447}], 'summary': 'New york university is developing a project called ml5, an homage to p5.', 'duration': 33.516, 'max_score': 946.195, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0946195.jpg'}, {'end': 979.711, 'src': 'embed', 'start': 952.519, 'weight': 3, 'content': [{'end': 957.682, 'text': 'but layers, so layers in TensorFlow.js is the equivalent of this thing called Keras now.', 'start': 952.519, 'duration': 5.163}, {'end': 966.417, 'text': "So a project that's being developed here at New York University with some collaborators from ITP and guests and researchers and students,", 'start': 958.8, 'duration': 7.617}, {'end': 967.58, 'text': 'is a project called ML5..', 'start': 966.417, 'duration': 1.163}, {'end': 979.711, 'text': "The five in ML5 is an homage homage that's French right to the five in P5 in the sense that I mean this is flawed for many reasons.", 'start': 969.264, 'duration': 10.447}], 'summary': 'Nyu is developing a project called ml5, an homage to p5, with collaborators from itp.', 'duration': 27.192, 'max_score': 952.519, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0952519.jpg'}, {'end': 1033.026, 'src': 'embed', 'start': 1007.97, 'weight': 4, 'content': [{'end': 1015.735, 'text': 'to do some common to allow to sound like even abstract the concepts even a bit further.', 'start': 1007.97, 'duration': 7.765}, {'end': 1022.38, 'text': 'and you know, I think one of the goals of ml5 is for it to be a library that a high school class could use,', 'start': 1015.735, 'duration': 6.645}, {'end': 1030.785, 'text': 'a kind of weekend workshop for artists could use these sort of contexts of people wanting to get a basic understanding and try some machine learning stuff out.', 'start': 1022.38, 'duration': 8.405}, {'end': 1033.026, 'text': 'So anyway, so this is all the stuff.', 'start': 1031.305, 'duration': 1.721}], 'summary': 'Ml5 aims to be accessible for high school and art workshop use.', 'duration': 25.056, 'max_score': 1007.97, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD01007970.jpg'}], 'start': 759.494, 'title': 'Programming languages and keras-tensorflow integration', 'summary': 'Discusses high-level and low-level programming languages, the relationship between tensorflow, keras, machine learning frameworks, tensorflow.js integration, ml5 library development, and the upcoming release aiming to make machine learning accessible to high school classes and artists.', 'chapters': [{'end': 902.644, 'start': 759.494, 'title': 'Programming language levels', 'summary': 'Discusses the differences between high-level and low-level programming languages, and the relationship between tensorflow, keras, and other machine learning frameworks, with tensorflow being a low-level library for machine learning operations.', 'duration': 143.15, 'highlights': ['TensorFlow as a low-level library for machine learning operations, providing common operations already implemented. TensorFlow is a low-level library for machine learning operations, providing common operations that are already implemented, such as matrix math and gradient descent learning training algorithms.', 'Keras as a higher-level API designed to sit on top of various low-level machine learning frameworks. Keras was designed to be a higher-level API that could sit on top of various low-level machine learning frameworks, such as TensorFlow, Fiano, PyTorch, and others.', 'Differences between high-level and low-level programming languages, with low-level involving manipulation of RAM and central processing unit data. The chapter discusses the differences between high-level and low-level programming languages, with low-level involving the manipulation of RAM and central processing unit data, while high-level involves a more abstract approach.']}, {'end': 1129.839, 'start': 903.165, 'title': 'Keras and tensorflow.js integration', 'summary': 'Discusses the integration of keras and tensorflow.js, the development of ml5 library, and the upcoming release in june, aiming to make machine learning accessible to high school classes and artists.', 'duration': 226.674, 'highlights': ['Keras is now part of the TensorFlow project, serving as a higher level API built on top of TensorFlow, with TensorFlow.js incorporating the core and layers APIs.', 'ML5 is being developed as a JavaScript library on top of TensorFlow.js, targeting accessibility for high school classes and artists, with a planned public release in June.', 'The goal of ML5 is to enable basic understanding and experimentation in machine learning, with the intention to provide examples and features upon its public release in June.', 'The upcoming ML5 library aims to make machine learning more accessible to a wider audience, including high school classes, weekend workshops for artists, and individuals seeking a basic understanding of machine learning concepts.']}], 'duration': 370.345, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/Qt3ZABW5lD0/pics/Qt3ZABW5lD0759494.jpg', 'highlights': ['TensorFlow is a low-level library for machine learning operations, providing common operations already implemented.', 'Keras as a higher-level API designed to sit on top of various low-level machine learning frameworks.', 'Keras is now part of the TensorFlow project, serving as a higher level API built on top of TensorFlow.', 'ML5 is being developed as a JavaScript library on top of TensorFlow.js, targeting accessibility for high school classes and artists.', 'The upcoming ML5 library aims to make machine learning more accessible to a wider audience.']}], 'highlights': ['TensorFlow.js is the foundation for upcoming videos, discussing tensors and their significance in machine learning.', 'TensorFlow.js adopted deeplearn.js as of March 30th, less than a month ago', 'The speaker plans to use a more robust, thoughtful, well-designed framework as the foundation for future examples and projects.', 'A tensor is a data structure that can store any n-dimensional version of scalars, vectors, and matrices, the building blocks of machine learning algorithms.', 'TensorFlow is a low-level library for machine learning operations, providing common operations already implemented.', 'Keras is now part of the TensorFlow project, serving as a higher level API built on top of TensorFlow.']}