title
TensorFlow 2.0 Tutorial For Beginners | TensorFlow Demo | Deep Learning & TensorFlow | Simplilearn
description
🔥 Purdue Post Graduate Program In AI And Machine Learning: https://www.simplilearn.com/pgp-ai-machine-learning-certification-training-course?utm_campaign=AI-QPDsEtUK_D4&utm_medium=Descriptionff&utm_source=youtube
🔥Professional Certificate Course In AI And Machine Learning by IIT Kanpur (India Only): https://www.simplilearn.com/iitk-professional-certificate-course-ai-machine-learning?utm_campaign=23AugustTubebuddyExpPCPAIandML&utm_medium=DescriptionFF&utm_source=youtube
🔥AI & Machine Learning Bootcamp(US Only): https://www.simplilearn.com/ai-machine-learning-bootcamp?utm_campaign=AI-QPDsEtUK_D4&utm_medium=Descriptionff&utm_source=youtube
TensorFlow is one of the most commonly used frameworks for deep learning. This TensorFlow 2.0 Tutorial covers everything from basics to advanced. You will learn what TensorFlow is, and the different applications of TensorFlow. You will understand tensors and how a computational graph works. You will get an idea about TensorFlow's architecture and perform a hands-on demo on LSTMs using the air quality dataset.
00:00:00 Deep Learning Frameworks
00:01:32 What Is TensorFlow?
00:01:53 Features of TensorFlow
00:03:44 TensorFlow Applications
00:06:18 How TensorFlow Works?
00:07:44 TensorFlow 1.0 vs 2.0
00:15:25 TensorFlow 2.0 Architecture
00:21:07 TensorFlow Demo
Dataset Link - https://drive.google.com/drive/folders/1UtOikFc9Wz0SA_W-g50gTfa8tXPWfLpU
âś…Subscribe to our Channel to learn more about the top Technologies: https://bit.ly/2VT4WtH
⏩ Check out the Artificial Intelligence training videos: https://bit.ly/2Li4Rur
#TensorFlowTutorial #TensorFlowTutorialForBeginners #DeepLearningWithTensorFlow #TensorFlowExplained #ArtificialIntelligence #DeepLearning #Simplilearn
To learn more about this topic, visit: https://www.simplilearn.com/tutorials/deep-learning-tutorial/tensorflow-2?utm_campaign=WhatisTensorflow2.0&utm_medium=Description&utm_source=youtube
Simplilearn’s Artificial Intelligence course provides training in the skills required for a career in AI. You will master TensorFlow, Machine Learning and other AI concepts, plus the programming languages needed to design intelligent agents, deep learning algorithms & advanced artificial neural networks that use predictive analytics to solve real-time decision-making problems without explicit programming.
The current and future demand for AI engineers is staggering. The New York Times reports a candidate shortage for certified AI Engineers, with fewer than 10,000 qualified people in the world to fill these jobs, which according to Paysa earn an average salary of $172,000 per year in the U.S. (or Rs.17 lakhs to Rs. 25 lakhs in India) for engineers with the required skills.
🔥 Enroll for FREE Artificial Intelligence Course & Get your Completion Certificate: https://www.simplilearn.com/learn-ai-basics-skillup?utm_campaign=AI-QPDsEtUK_D4&utm_medium=Description&utm_source=youtube
➡️ About Post Graduate Program In AI And Machine Learning
This AI ML course is designed to enhance your career in AI and ML by demystifying concepts like machine learning, deep learning, NLP, computer vision, reinforcement learning, and more. You'll also have access to 4 live sessions, led by industry experts, covering the latest advancements in AI such as generative modeling, ChatGPT, OpenAI, and chatbots.
âś… Key Features
- Post Graduate Program certificate and Alumni Association membership
- Exclusive hackathons and Ask me Anything sessions by IBM
- 3 Capstones and 25+ Projects with industry data sets from Twitter, Uber, Mercedes Benz, and many more
- Master Classes delivered by Purdue faculty and IBM experts
- Simplilearn's JobAssist helps you get noticed by top hiring companies
- Gain access to 4 live online sessions on latest AI trends such as ChatGPT, generative AI, explainable AI, and more
- Learn about the applications of ChatGPT, OpenAI, Dall-E, Midjourney & other prominent tools
âś… Skills Covered
- ChatGPT
- Generative AI
- Explainable AI
- Generative Modeling
- Statistics
- Python
- Supervised Learning
- Unsupervised Learning
- NLP
- Neural Networks
- Computer Vision
- And Many More…
👉 Learn More At: 🔥 Purdue Post Graduate Program In AI And Machine Learning: https://www.simplilearn.com/pgp-ai-machine-learning-certification-training-course?utm_campaign=AI-QPDsEtUK_D4&utm_medium=Description&utm_source=youtube
🔥🔥 Interested in Attending Live Classes? Call Us: IN - 18002127688 / US - +18445327688
detail
{'title': 'TensorFlow 2.0 Tutorial For Beginners | TensorFlow Demo | Deep Learning & TensorFlow | Simplilearn', 'heatmap': [{'end': 1197.238, 'start': 1142.744, 'weight': 1}, {'end': 1560.143, 'start': 1393.175, 'weight': 0.831}, {'end': 1923.098, 'start': 1869.872, 'weight': 0.861}, {'end': 2390.967, 'start': 2337.435, 'weight': 0.889}, {'end': 3017.551, 'start': 2958.636, 'weight': 0.85}], 'summary': 'Tutorial on tensorflow 2.0 for beginners covers deep learning frameworks with a focus on tensorflow, its evolution to 2.0, distribution and deployment options, basics, math functions, keras sequential model, data preprocessing, analysis, outlier treatment, lstm and ml model implementation, and optimizing the machine learning process with spark, achieving a square mean error of 0.0088 and test score of 3.16.', 'chapters': [{'end': 359.2, 'segs': [{'end': 128.497, 'src': 'embed', 'start': 102.856, 'weight': 0, 'content': [{'end': 109.302, 'text': 'It is based on Python programming language and performs numerical computations using Dataflow graphs to build models.', 'start': 102.856, 'duration': 6.446}, {'end': 113.526, 'text': "So let's take a look at some of the features of TensorFlow.", 'start': 109.703, 'duration': 3.823}, {'end': 117.21, 'text': 'It works efficiently with multi-dimensional arrays.', 'start': 113.766, 'duration': 3.444}, {'end': 121.233, 'text': "If you've ever played with any of the simpler packages of neural networks,", 'start': 117.93, 'duration': 3.303}, {'end': 128.497, 'text': "you're going to find that you have to pretty much flatten them and make sure your stuff is set in a flat model.", 'start': 122.094, 'duration': 6.403}], 'summary': 'Tensorflow uses dataflow graphs for numerical computations in python, efficiently handling multi-dimensional arrays.', 'duration': 25.641, 'max_score': 102.856, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D4102856.jpg'}, {'end': 199.141, 'src': 'embed', 'start': 167.423, 'weight': 2, 'content': [{'end': 171.686, 'text': 'This is one of the things that TensorFlow addresses and does a very good job on.', 'start': 167.423, 'duration': 4.263}, {'end': 175.108, 'text': 'It supports fast debugging and model building.', 'start': 172.466, 'duration': 2.642}, {'end': 177.429, 'text': 'This is why I love TensorFlow.', 'start': 175.748, 'duration': 1.681}, {'end': 181.611, 'text': 'I can go in there and I can build a model with different layers.', 'start': 178.309, 'duration': 3.302}, {'end': 183.772, 'text': 'Each layer might have different properties.', 'start': 181.891, 'duration': 1.881}, {'end': 191.216, 'text': 'They have like the convolutional neural network which you can then sit on top of a regular neural network with reverse propagation.', 'start': 184.733, 'duration': 6.483}, {'end': 199.141, 'text': "There's a lot of tools in here and a lot of options and each layer that it goes through can utilize those different options and stack differently.", 'start': 191.697, 'duration': 7.444}], 'summary': 'Tensorflow excels in fast debugging, model building, and layer customization.', 'duration': 31.718, 'max_score': 167.423, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D4167423.jpg'}, {'end': 271.205, 'src': 'embed', 'start': 230.683, 'weight': 3, 'content': [{'end': 232.124, 'text': "We're getting into the data science.", 'start': 230.683, 'duration': 1.441}, {'end': 235.647, 'text': "I like to use data science as probably a better term because it's the programming side.", 'start': 232.344, 'duration': 3.303}, {'end': 238.689, 'text': "And it's really the sky is the limit.", 'start': 237.228, 'duration': 1.461}, {'end': 245.332, 'text': 'We can look at face detection, language translation, fraud detection, video detection.', 'start': 239.49, 'duration': 5.842}, {'end': 249.975, 'text': 'There are so many different things out there that TensorFlow can be used for.', 'start': 246.073, 'duration': 3.902}, {'end': 257.778, 'text': 'When you think of neural networks, because TensorFlow is a neural network, think of complicated chaotic data.', 'start': 250.735, 'duration': 7.043}, {'end': 263.441, 'text': "This is very different than if you have a set of numbers like you're looking at the stock market.", 'start': 258.099, 'duration': 5.342}, {'end': 271.205, 'text': "You can use this on the stock market, but if you're doing something where the numbers are very clear and not so chaotic as you have in a picture,", 'start': 264.022, 'duration': 7.183}], 'summary': "Data science offers limitless possibilities, from face detection to stock market analysis using tensorflow's neural network.", 'duration': 40.522, 'max_score': 230.683, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D4230683.jpg'}, {'end': 323.641, 'src': 'embed', 'start': 290.297, 'weight': 5, 'content': [{'end': 292.338, 'text': 'So we talk about tensors and tensor flow.', 'start': 290.297, 'duration': 2.041}, {'end': 296.001, 'text': 'Tensor flow is derived from its core component known as a tensor.', 'start': 292.519, 'duration': 3.482}, {'end': 301.184, 'text': 'A tensor is a vector or a matrix of n dimensions that represent all types of data.', 'start': 296.181, 'duration': 5.003}, {'end': 305.487, 'text': 'And you can see here we have the scalar, which is just a single number.', 'start': 301.424, 'duration': 4.063}, {'end': 307.328, 'text': 'You have your vector, which is two numbers.', 'start': 305.647, 'duration': 1.681}, {'end': 309.27, 'text': 'It might be a number in a direction.', 'start': 308.029, 'duration': 1.241}, {'end': 313.093, 'text': 'You have a simple matrix and then we get into the tensor.', 'start': 309.85, 'duration': 3.243}, {'end': 317.376, 'text': 'I mentioned how a picture is a very complicated tensor,', 'start': 313.773, 'duration': 3.603}, {'end': 323.641, 'text': 'because it has your x y coordinates and then each one of those pixels has three to four channels for your different colors.', 'start': 317.376, 'duration': 6.265}], 'summary': 'Tensor flow is derived from tensors, representing data in n dimensions, including scalar, vector, matrix, and complex images.', 'duration': 33.344, 'max_score': 290.297, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D4290297.jpg'}], 'start': 7.449, 'title': 'Tensorflow in deep learning', 'summary': 'Covers deep learning frameworks, with a focus on tensorflow, emphasizing its popularity, efficiency with multi-dimensional arrays, and scalability of computation across machines and large data sets. it also explores model building, visualization with tensorboard, and applications in data analytics such as face detection, language translation, and fraud detection.', 'chapters': [{'end': 166.823, 'start': 7.449, 'title': 'Tensorflow 2.0 tutorial', 'summary': 'Covers deep learning frameworks, with a focus on tensorflow, including its features, applications, and architecture, emphasizing its popularity, efficiency with multi-dimensional arrays, and scalability of computation across machines and large data sets.', 'duration': 159.374, 'highlights': ['TensorFlow is a popular open source library released in 2015 by Google Brain Team for building machine learning and deep learning models, based on Python programming language and performs numerical computations using Dataflow graphs to build models.', 'TensorFlow works efficiently with multi-dimensional arrays, providing scalability of computation across machines and large data sets, making it suitable for handling complex data structures like images with multiple channels, and ensuring consistent results across different machines.', "The chapter discusses various deep learning frameworks including TensorFlow, Pytorch, Cafe, Theano, DL4J, and Chainer, highlighting TensorFlow's integration with Keras and its position as the most robust and top-of-the-line technology in the field of neural networks."]}, {'end': 359.2, 'start': 167.423, 'title': 'Tensorflow: model building and applications', 'summary': 'Explores the capabilities of tensorflow, including model building, visualization with tensorboard, and applications in data analytics such as face detection, language translation, and fraud detection.', 'duration': 191.777, 'highlights': ['TensorFlow supports fast debugging and model building, with a large community and provides TensorBoard for visualization. TensorFlow facilitates fast debugging and model building, with the provision of TensorBoard for visualization, aiding collaboration and showcasing the model to clients and shareholders.', 'TensorFlow has a wide range of applications including data analytics, face detection, language translation, and fraud detection. TensorFlow is extensively applicable in data analytics, with diverse uses such as face detection, language translation, and fraud detection, showcasing its versatility.', 'TensorFlow is ideal for handling complicated and chaotic data patterns, suitable for neural networks. TensorFlow is well-suited for managing intricate and chaotic data patterns, particularly for neural networks, distinguishing it from linear regression models.', 'TensorFlow is based on its core component, a tensor, which represents data in the form of vectors or matrices of n dimensions. TensorFlow is derived from its core component, the tensor, representing data in the form of vectors or matrices of n dimensions, enabling versatile data representation and manipulation.']}], 'duration': 351.751, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D47449.jpg', 'highlights': ['TensorFlow is a popular open source library for building machine learning and deep learning models, based on Python programming language and performs numerical computations using Dataflow graphs to build models.', 'TensorFlow works efficiently with multi-dimensional arrays, providing scalability of computation across machines and large data sets, making it suitable for handling complex data structures like images with multiple channels, and ensuring consistent results across different machines.', 'TensorFlow supports fast debugging and model building, with a large community and provides TensorBoard for visualization.', 'TensorFlow has a wide range of applications including data analytics, face detection, language translation, and fraud detection.', 'TensorFlow is ideal for handling complicated and chaotic data patterns, suitable for neural networks.', 'TensorFlow is based on its core component, a tensor, which represents data in the form of vectors or matrices of n dimensions.']}, {'end': 1035.967, 'segs': [{'end': 501.787, 'src': 'embed', 'start': 458.914, 'weight': 0, 'content': [{'end': 461.196, 'text': 'So you can test out these different models to see how they work.', 'start': 458.914, 'duration': 2.282}, {'end': 463.959, 'text': 'Now, TensorFlow has gone through two major stages.', 'start': 461.276, 'duration': 2.683}, {'end': 470.621, 'text': 'We had the original TensorFlow release of 1.0, and then they came out with the 2.0 version.', 'start': 464.82, 'duration': 5.801}, {'end': 475.622, 'text': 'And the 2.0 addressed so many things out there that the 1.0 really needed.', 'start': 470.921, 'duration': 4.701}, {'end': 480.363, 'text': 'So, when we start talking about TensorFlow 1.0 versus 2.0,.', 'start': 475.642, 'duration': 4.721}, {'end': 486.884, 'text': "I guess you would need to know this for a legacy programming job if you're pulling apart somebody else's code.", 'start': 480.363, 'duration': 6.521}, {'end': 491.805, 'text': 'The first thing is that TensorFlow 2.0 supports eager execution by default.', 'start': 487.124, 'duration': 4.681}, {'end': 494.285, 'text': 'It allows you to build your models and run them instantly.', 'start': 491.925, 'duration': 2.36}, {'end': 501.787, 'text': 'And you can see here from TensorFlow 1 to TensorFlow 2, we have almost double the code to do the same thing.', 'start': 494.505, 'duration': 7.282}], 'summary': 'Tensorflow 2.0 addressed many issues in 1.0, supports eager execution by default, and requires almost double the code for the same task.', 'duration': 42.873, 'max_score': 458.914, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D4458914.jpg'}, {'end': 783.432, 'src': 'embed', 'start': 751.022, 'weight': 3, 'content': [{'end': 753.905, 'text': 'So that TensorFlow runs it as a single graph.', 'start': 751.022, 'duration': 2.883}, {'end': 759.512, 'text': 'Autograph feature of tf function helps to write graph code using natural Python syntax.', 'start': 754.125, 'duration': 5.387}, {'end': 763.483, 'text': 'Now we just threw in a new word in you, graph.', 'start': 761.062, 'duration': 2.421}, {'end': 765.824, 'text': 'Graph is not a picture of a person.', 'start': 763.603, 'duration': 2.221}, {'end': 768.565, 'text': "You'll hear graph X and some other things.", 'start': 766.564, 'duration': 2.001}, {'end': 773.167, 'text': 'Graph is what are all those lines that are connecting different objects.', 'start': 769.266, 'duration': 3.901}, {'end': 783.432, 'text': 'So if you remember from before where we had the different layers going through sequentially, each one of those white lined arrows would be a graph X.', 'start': 773.688, 'duration': 9.744}], 'summary': 'Autograph feature of tf function simplifies graph code in tensorflow.', 'duration': 32.41, 'max_score': 751.022, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D4751022.jpg'}, {'end': 899.927, 'src': 'embed', 'start': 872.097, 'weight': 4, 'content': [{'end': 876.098, 'text': 'and then it comes together into another layer, which is another neural network.', 'start': 872.097, 'duration': 4.001}, {'end': 883.021, 'text': 'So you can build these really complicated models and at the low level you can put in your own APIs, you can move that stuff around.', 'start': 876.678, 'duration': 6.343}, {'end': 887.442, 'text': 'And most recently we have the TF code can run on multiple platforms.', 'start': 883.521, 'duration': 3.921}, {'end': 897.386, 'text': "And so you have your CPU, which is basically like on the computer I'm running on, I have 8 cores and 16 dedicated threads.", 'start': 888.263, 'duration': 9.123}, {'end': 899.927, 'text': 'I hear they now have one out there that has over 100 cores.', 'start': 897.927, 'duration': 2}], 'summary': 'Neural networks can be built with multiple platforms; cpu can have over 100 cores.', 'duration': 27.83, 'max_score': 872.097, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D4872097.jpg'}], 'start': 359.2, 'title': 'Tensorflow 2.0 evolution and architecture', 'summary': 'Delves into the basics of tensorflow, its evolution from 1.0 to 2.0, highlighting the improvements in tensorflow 2.0 and its support for eager execution by default. it also covers the significant changes in tensorflow 2.0, including model abstractions, api cleanup, tf function, autograph feature, and its architecture for running code on multiple platforms.', 'chapters': [{'end': 567.45, 'start': 359.2, 'title': 'Understanding tensorflow and its evolution', 'summary': 'Explains the basics of tensorflow, its computational flow, and the transition from tensorflow 1.0 to 2.0, highlighting the improvements in tensorflow 2.0 and its support for eager execution by default, resulting in a more concise code structure.', 'duration': 208.25, 'highlights': ['TensorFlow 2.0 supports eager execution by default, resulting in almost double the code reduction compared to TensorFlow 1.0.', 'The transition from TensorFlow 1.0 to 2.0 has addressed numerous deficiencies and complexities, making the code structure more concise and user-friendly.', 'Keras, the high-level API in TensorFlow 2.0, provides various model building APIs, including sequential, functional, and subclassing.']}, {'end': 1035.967, 'start': 567.45, 'title': 'Tensorflow 2.0: new features and architecture', 'summary': 'Covers the significant changes in tensorflow 2.0, including the shift from tensorflow 1.0, the three main model abstractions - sequential, functional, and subclassing, the removal of legacy features in the api cleanup, the introduction of tf function and autograph feature, the hierarchy and architecture of tensorflow 2.0, and the ability to run tf code on multiple platforms.', 'duration': 468.517, 'highlights': ['The significant changes in TensorFlow 2.0, including the shift from TensorFlow 1.0 and the new model abstractions - sequential, functional, and subclassing. TensorFlow 2.0 introduces major changes from TensorFlow 1.0, such as new model abstractions including sequential, functional, and subclassing, providing developers with more flexibility and power in building their own models.', 'The removal of legacy features in the API cleanup, such as tf.gans, tf.app, tf.contrib, tf.flags, and tf.logging. In TensorFlow 2.0, legacy features like tf.gans, tf.app, tf.contrib, tf.flags, and tf.logging have been removed, simplifying the APIs and streamlining the development process.', 'The introduction of TF function and autograph feature, allowing for JIT compilation and writing graph code using natural Python syntax. TensorFlow 2.0 introduces TF function and autograph feature, enabling developers to mark Python functions for JIT compilation and write graph code using natural Python syntax, enhancing the efficiency and flexibility of TensorFlow programming.', "The hierarchy and architecture of TensorFlow 2.0, featuring high-level object-oriented APIs, TF Keras, estimators, TF layers, TF losses, and TF metrics, as well as low-level TF API for extensive control and custom model building. TensorFlow 2.0 offers a hierarchy with high-level object-oriented APIs like TF Keras, estimators, and low-level TF API, providing extensive control for custom model building and allowing for reusable libraries for model construction, showcasing the versatility and power of TensorFlow's architecture.", 'The ability to run TF code on multiple platforms, including CPU, GPU, and TPU, enhancing the versatility and accessibility of TensorFlow models. TensorFlow 2.0 allows the execution of TF code on various platforms, including CPU, GPU, and TPU, expanding the reach and applicability of TensorFlow models across diverse hardware setups and environments.']}], 'duration': 676.767, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D4359200.jpg', 'highlights': ['TensorFlow 2.0 supports eager execution by default, resulting in almost double the code reduction compared to TensorFlow 1.0.', 'The transition from TensorFlow 1.0 to 2.0 has addressed numerous deficiencies and complexities, making the code structure more concise and user-friendly.', 'The significant changes in TensorFlow 2.0, including the shift from TensorFlow 1.0 and the new model abstractions - sequential, functional, and subclassing.', 'The introduction of TF function and autograph feature, allowing for JIT compilation and writing graph code using natural Python syntax.', 'The ability to run TF code on multiple platforms, including CPU, GPU, and TPU, enhancing the versatility and accessibility of TensorFlow models.']}, {'end': 1358.885, 'segs': [{'end': 1099.389, 'src': 'embed', 'start': 1054.578, 'weight': 0, 'content': [{'end': 1062.827, 'text': 'Now they have TensorFlow Lite, so you can actually run a TensorFlow on an Android or an iOS or Raspberry Pi a little breakout board there.', 'start': 1054.578, 'duration': 8.249}, {'end': 1070.916, 'text': "In fact, they just came out with a new one that has a built-in, it's just a little mini TPU with a camera on it so it can pre-process a video.", 'start': 1062.847, 'duration': 8.069}, {'end': 1073.458, 'text': 'So you can load your TensorFlow model onto that.', 'start': 1071.056, 'duration': 2.402}, {'end': 1078.561, 'text': 'talking about an affordable way to beta test a new product.', 'start': 1074.079, 'duration': 4.482}, {'end': 1082.302, 'text': 'You have the TensorFlow.js, which is for browser and node server.', 'start': 1078.941, 'duration': 3.361}, {'end': 1087.764, 'text': "So you can get that out on the browser for some simple computations that don't require a lot of heavy lifting,", 'start': 1082.642, 'duration': 5.122}, {'end': 1090.025, 'text': 'but you want to distribute to a lot of endpoints.', 'start': 1087.764, 'duration': 2.261}, {'end': 1092.346, 'text': 'And now they also have other language bindings.', 'start': 1090.425, 'duration': 1.921}, {'end': 1099.389, 'text': 'So you can now create your TensorFlow backend, save it and have it accessed from C, Java Go, C, Sharp,', 'start': 1092.426, 'duration': 6.963}], 'summary': 'Tensorflow lite enables running models on android, ios, or raspberry pi, with new tpu-equipped board for video processing. also supports tensorflow.js for browser and node server, along with language bindings for c, java, go, and c#.', 'duration': 44.811, 'max_score': 1054.578, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41054578.jpg'}, {'end': 1199.18, 'src': 'heatmap', 'start': 1120.272, 'weight': 3, 'content': [{'end': 1125.314, 'text': "So very basic things you need to know and understand when you're working with the TensorFlow setup.", 'start': 1120.272, 'duration': 5.042}, {'end': 1127.936, 'text': 'So constants in TensorFlow.', 'start': 1125.795, 'duration': 2.141}, {'end': 1131.497, 'text': 'In TensorFlow, constants are created using the function constant.', 'start': 1128.216, 'duration': 3.281}, {'end': 1135.359, 'text': "In other words, they're going to stay static the whole time, whatever you're working with.", 'start': 1131.817, 'duration': 3.542}, {'end': 1139.062, 'text': 'the syntax for constant value.', 'start': 1135.999, 'duration': 3.063}, {'end': 1141.163, 'text': 'dtype 9 shape equals none.', 'start': 1139.062, 'duration': 2.101}, {'end': 1142.744, 'text': 'name constant verify.', 'start': 1141.163, 'duration': 1.581}, {'end': 1143.965, 'text': 'shape equals false.', 'start': 1142.744, 'duration': 1.221}, {'end': 1149.09, 'text': "that's kind of the syntax you're looking at and we'll explore this with our hands on a little more in depth.", 'start': 1143.965, 'duration': 5.125}, {'end': 1151.431, 'text': 'and you can see here we do z equals tf dot.', 'start': 1149.09, 'duration': 2.341}, {'end': 1153.493, 'text': 'constant 5.2.', 'start': 1151.431, 'duration': 2.062}, {'end': 1154.974, 'text': 'name equals x.', 'start': 1153.493, 'duration': 1.481}, {'end': 1156.295, 'text': 'dtype is a float.', 'start': 1154.974, 'duration': 1.321}, {'end': 1157.837, 'text': "that means that we're never going to change that 5.2.", 'start': 1156.295, 'duration': 1.542}, {'end': 1159.598, 'text': "it's going to be a constant value.", 'start': 1157.837, 'duration': 1.761}, {'end': 1162.301, 'text': 'And then we have our variables in TensorFlow.', 'start': 1159.798, 'duration': 2.503}, {'end': 1166.085, 'text': 'Variables in TensorFlow are in memory buffers that store tensors.', 'start': 1162.741, 'duration': 3.344}, {'end': 1170.43, 'text': 'And so we can declare a two by three tensor populated by ones.', 'start': 1166.906, 'duration': 3.524}, {'end': 1172.613, 'text': 'You could also do constants this way, by the way.', 'start': 1170.831, 'duration': 1.782}, {'end': 1176.117, 'text': 'So you can create an array of ones for your constants.', 'start': 1172.673, 'duration': 3.444}, {'end': 1179.401, 'text': "I'm not sure why you do that, but you might need that for some reason.", 'start': 1176.197, 'duration': 3.204}, {'end': 1189.51, 'text': 'In here we have v equals, tf.variables, and then in TensorFlow you have tf.ones and you have the shape, which is 2, 3,', 'start': 1180.321, 'duration': 9.189}, {'end': 1194.195, 'text': "which is then going to create a nice 2 by 3 array that's filled with ones.", 'start': 1189.51, 'duration': 4.685}, {'end': 1197.238, 'text': 'And then, of course, you can go in there in their variables, so you can change them.', 'start': 1194.355, 'duration': 2.883}, {'end': 1199.18, 'text': "It's a tensor, so you have full control over that.", 'start': 1197.298, 'duration': 1.882}], 'summary': 'Introduction to tensorflow constants and variables, including syntax and examples.', 'duration': 78.908, 'max_score': 1120.272, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41120272.jpg'}, {'end': 1189.51, 'src': 'embed', 'start': 1162.741, 'weight': 4, 'content': [{'end': 1166.085, 'text': 'Variables in TensorFlow are in memory buffers that store tensors.', 'start': 1162.741, 'duration': 3.344}, {'end': 1170.43, 'text': 'And so we can declare a two by three tensor populated by ones.', 'start': 1166.906, 'duration': 3.524}, {'end': 1172.613, 'text': 'You could also do constants this way, by the way.', 'start': 1170.831, 'duration': 1.782}, {'end': 1176.117, 'text': 'So you can create an array of ones for your constants.', 'start': 1172.673, 'duration': 3.444}, {'end': 1179.401, 'text': "I'm not sure why you do that, but you might need that for some reason.", 'start': 1176.197, 'duration': 3.204}, {'end': 1189.51, 'text': 'In here we have v equals, tf.variables, and then in TensorFlow you have tf.ones and you have the shape, which is 2, 3,', 'start': 1180.321, 'duration': 9.189}], 'summary': 'Tensorflow uses in-memory buffers for variables and constants, like creating a 2x3 tensor populated by ones.', 'duration': 26.769, 'max_score': 1162.741, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41162741.jpg'}, {'end': 1237.828, 'src': 'embed', 'start': 1203.601, 'weight': 5, 'content': [{'end': 1209.722, 'text': 'A session in TensorFlow is used to run a computational graph to evaluate the nodes.', 'start': 1203.601, 'duration': 6.121}, {'end': 1215.583, 'text': "And remember when we're talking a graph or GraphX, we're talking about all that information.", 'start': 1210.542, 'duration': 5.041}, {'end': 1220.244, 'text': 'then goes through all those arrows and whatever computations they have that take it to the next node.', 'start': 1215.583, 'duration': 4.661}, {'end': 1223.885, 'text': 'And you can see down here where we have import TensorFlow as tf.', 'start': 1220.584, 'duration': 3.301}, {'end': 1237.828, 'text': "If we do x equals a tf.constant of 10, we do y equals a tf constant of 20.0, and then you can do z equals tf dot variable and it's a tf dot.", 'start': 1224.365, 'duration': 13.463}], 'summary': 'Using tensorflow to run computational graphs with nodes and variables.', 'duration': 34.227, 'max_score': 1203.601, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41203601.jpg'}, {'end': 1319.665, 'src': 'embed', 'start': 1290.53, 'weight': 6, 'content': [{'end': 1293.311, 'text': "PyCharm, whatever it is you're going to use in here.", 'start': 1290.53, 'duration': 2.781}, {'end': 1298.314, 'text': 'You know, you have your Spyder and your Qt console for different programming environments.', 'start': 1293.331, 'duration': 4.983}, {'end': 1304.837, 'text': "The thing to note, it's kind of hard to see, but I have my main Pi 3.6.", 'start': 1298.494, 'duration': 6.343}, {'end': 1309.62, 'text': 'Right now, when I was writing this, TensorFlow works in Python version 3.6.', 'start': 1304.837, 'duration': 4.783}, {'end': 1315.983, 'text': "If you have Python version 3.7 or 3.8, you're probably going to get some errors in there.", 'start': 1309.62, 'duration': 6.363}, {'end': 1319.665, 'text': "It might be that they've already updated it and I don't know it and I have an older version.", 'start': 1316.543, 'duration': 3.122}], 'summary': 'Python 3.6 is required for tensorflow to work without errors, other versions may cause issues.', 'duration': 29.135, 'max_score': 1290.53, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41290530.jpg'}], 'start': 1035.967, 'title': 'Tensorflow distribution and basics', 'summary': "Explores tensorflow's distribution and deployment options including tensorflow serving cloud, tensorflow lite, and tensorflow.js, providing affordable ways to beta test products and distribute computations. it also delves into tensorflow basics covering constants, variables, and sessions with syntax examples and emphasizes python 3.6 compatibility.", 'chapters': [{'end': 1099.389, 'start': 1035.967, 'title': 'Tensorflow distribution and deployment options', 'summary': "Discusses tensorflow's distribution and deployment options, including tensorflow serving cloud, tensorflow lite for android, ios, and raspberry pi, tensorflow.js for browser and node server, and other language bindings, offering affordable ways to beta test new products and distribute computations across multiple endpoints.", 'duration': 63.422, 'highlights': ['The availability of TensorFlow Lite enables running TensorFlow on Android, iOS, or Raspberry Pi, allowing for affordable beta testing of new products.', 'TensorFlow.js is designed for browser and node server, providing a solution for simple computations that can be distributed across multiple endpoints.', 'TensorFlow also offers language bindings for C, Java, Go, and C#, expanding its accessibility and integration capabilities.']}, {'end': 1358.885, 'start': 1099.389, 'title': 'Tensorflow basics: constants, variables, and sessions', 'summary': 'Discusses the basics of tensorflow, covering constants, variables, and sessions, with specific syntax examples and explanations of their functionalities, and emphasizes the importance of using python version 3.6 for tensorflow compatibility.', 'duration': 259.496, 'highlights': ["TensorFlow constants are created using the function constant, remaining static throughout, with an example of syntax and value assignment provided, such as z = tf.constant(5.2, name='x', dtype=float). Explains the creation and syntax of TensorFlow constants and provides a specific example (z = tf.constant(5.2, name='x', dtype=float)).", 'Variables in TensorFlow are memory buffers that store tensors, allowing for the creation of arrays and manipulation of their values, demonstrated by the declaration v = tf.variables(tf.ones([2, 3])). Details the purpose and usage of variables in TensorFlow, with an example of creating a 2x3 array filled with ones (v = tf.variables(tf.ones([2, 3]))).', 'Sessions in TensorFlow are used to run computational graphs and evaluate nodes, exemplified by the initialization and running of a session to perform a simple mathematical operation. Describes the role and functionality of sessions in TensorFlow, showcasing the initialization and execution of a session for performing computations.', 'Emphasizes the importance of using Python version 3.6 for TensorFlow compatibility, mentioning potential errors with versions 3.7 or 3.8, and providing guidance on setting up the appropriate environment and installing TensorFlow. Stresses the significance of Python version 3.6 for TensorFlow compatibility, warning about potential errors with other versions and offering instructions for environment setup and TensorFlow installation.']}], 'duration': 322.918, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41035967.jpg', 'highlights': ['TensorFlow Lite enables affordable beta testing on Android, iOS, or Raspberry Pi.', 'TensorFlow.js provides a solution for simple computations across multiple endpoints.', 'TensorFlow offers language bindings for C, Java, Go, and C# for accessibility and integration capabilities.', 'TensorFlow constants remain static and are created using the function constant.', 'Variables in TensorFlow store tensors and allow manipulation of their values.', 'Sessions in TensorFlow are used to run computational graphs and evaluate nodes.', 'Emphasizes Python 3.6 compatibility for TensorFlow, warning about potential errors with other versions.']}, {'end': 2013.182, 'segs': [{'end': 1560.143, 'src': 'heatmap', 'start': 1359.846, 'weight': 0, 'content': [{'end': 1360.986, 'text': "Let's go ahead and import.", 'start': 1359.846, 'duration': 1.14}, {'end': 1362.947, 'text': 'Import TensorFlow as tf.', 'start': 1361.267, 'duration': 1.68}, {'end': 1365.708, 'text': "At this point we'll go ahead and just run it real quick.", 'start': 1363.727, 'duration': 1.981}, {'end': 1366.988, 'text': 'No errors.', 'start': 1366.388, 'duration': 0.6}, {'end': 1368.589, 'text': 'Yay! No errors.', 'start': 1367.328, 'duration': 1.261}, {'end': 1373.77, 'text': 'I do that whenever I do my imports, because I, unbearably,', 'start': 1370.749, 'duration': 3.021}, {'end': 1381.692, 'text': 'will have opened up a new environment and forgotten to install TensorFlow into that environment or something along those lines.', 'start': 1373.77, 'duration': 7.922}, {'end': 1383.193, 'text': "So it's always good to double check.", 'start': 1381.752, 'duration': 1.441}, {'end': 1389.556, 'text': "And if we're going to double check that, it's also good to know what version we're working with.", 'start': 1384.482, 'duration': 5.074}, {'end': 1392.184, 'text': 'And we can do that simply by.', 'start': 1390.198, 'duration': 1.986}, {'end': 1404.996, 'text': 'using the version command in TensorFlow, which you should know is probably intuitively the tf.', 'start': 1393.175, 'duration': 11.821}, {'end': 1412.353, 'text': ',version, It always confuses me because sometimes you do tf.version for one thing, you do tf.', 'start': 1404.996, 'duration': 7.357}, {'end': 1413.934, 'text': ',version, for another thing.', 'start': 1412.353, 'duration': 1.581}, {'end': 1418.539, 'text': 'This is a double underscore in TensorFlow for pulling your version out.', 'start': 1414.615, 'duration': 3.924}, {'end': 1421.06, 'text': "And it's good to know what you're working with.", 'start': 1419.259, 'duration': 1.801}, {'end': 1424.902, 'text': "We're going to be working in TensorFlow version 2.1.", 'start': 1421.14, 'duration': 3.762}, {'end': 1429.805, 'text': '0 And I did tell you that we were going to dig a little deeper into our constants.', 'start': 1424.902, 'duration': 4.903}, {'end': 1431.766, 'text': 'And you can do an array of constants.', 'start': 1429.885, 'duration': 1.881}, {'end': 1436.228, 'text': "And we'll just create this nice array, a equals tf.constant.", 'start': 1432.286, 'duration': 3.942}, {'end': 1440.97, 'text': "And we're just going to put the array right in there, 4, 3, 6, 1.", 'start': 1436.908, 'duration': 4.062}, {'end': 1441.631, 'text': 'We can run this.', 'start': 1440.971, 'duration': 0.66}, {'end': 1443.412, 'text': 'And now that is what a is equal to.', 'start': 1441.671, 'duration': 1.741}, {'end': 1449.735, 'text': "And if we want to just double check that, remember we're in Jupyter Notebook, where I can just put the letter A,", 'start': 1443.809, 'duration': 5.926}, {'end': 1451.336, 'text': "and it knows that that's going to be print.", 'start': 1449.735, 'duration': 1.601}, {'end': 1454.019, 'text': "Otherwise you'd surround it in print.", 'start': 1452.317, 'duration': 1.702}, {'end': 1456.361, 'text': "And you can see it's a TF tensor.", 'start': 1454.599, 'duration': 1.762}, {'end': 1459.424, 'text': 'It has the shape, the type, and the array on here.', 'start': 1456.781, 'duration': 2.643}, {'end': 1460.565, 'text': "It's a two by two array.", 'start': 1459.444, 'duration': 1.121}, {'end': 1464.196, 'text': 'And just like we can create a constant, we can go ahead and create a variable.', 'start': 1461.053, 'duration': 3.143}, {'end': 1466.798, 'text': 'And this is also going to be a two by two array.', 'start': 1464.776, 'duration': 2.022}, {'end': 1470.421, 'text': "And if we go ahead and print the V out, we'll run that.", 'start': 1466.818, 'duration': 3.603}, {'end': 1474.144, 'text': "And sure enough, there's our TF variable in here.", 'start': 1471.482, 'duration': 2.662}, {'end': 1479.489, 'text': "Then we can also, let's just go back up here and add this in here.", 'start': 1476.066, 'duration': 3.423}, {'end': 1483.813, 'text': "I could create another tensor and we'll make it a constant this time.", 'start': 1480.55, 'duration': 3.263}, {'end': 1487.008, 'text': "And we're going to put that in over here.", 'start': 1485.626, 'duration': 1.382}, {'end': 1489.531, 'text': "We'll have b tf constant.", 'start': 1487.589, 'duration': 1.942}, {'end': 1494.097, 'text': "And if we go and print out v and b, we're going to run that.", 'start': 1490.332, 'duration': 3.765}, {'end': 1497.602, 'text': 'And this is an interesting thing that always happens in here.', 'start': 1494.738, 'duration': 2.864}, {'end': 1503.787, 'text': "You'll see right here, when I print them both out, What happens is it only prints the last one unless you use print commands.", 'start': 1498.123, 'duration': 5.664}, {'end': 1506.227, 'text': 'So important to remember that in Jupyter Notebooks.', 'start': 1504.407, 'duration': 1.82}, {'end': 1511.128, 'text': 'We can easily fix that by go ahead and print and surround V with brackets.', 'start': 1506.687, 'duration': 4.441}, {'end': 1517.369, 'text': 'And now we can see with the two different variables we have, we have the 3, 1, 5, 2, which is a variable.', 'start': 1511.388, 'duration': 5.981}, {'end': 1520.39, 'text': 'And this is just a constant.', 'start': 1517.949, 'duration': 2.441}, {'end': 1524.051, 'text': 'So it comes up as a TF tensor, shape 2, kind of 2.', 'start': 1520.43, 'duration': 3.621}, {'end': 1529.772, 'text': "And that's interesting to note that this label is a TF dot tensor, and this is a TF variable.", 'start': 1524.051, 'duration': 5.721}, {'end': 1535.616, 'text': "So that's how it's looking in the back end when you're talking about the difference between a variable and a constant.", 'start': 1530.732, 'duration': 4.884}, {'end': 1543.863, 'text': 'The other thing I want you to notice is that in variable we capitalize the V and with the constant we have a lowercase c.', 'start': 1536.697, 'duration': 7.166}, {'end': 1549.067, 'text': "Little things like that can lose you when you're programming and you're trying to find out hey, why doesn't this work?", 'start': 1543.863, 'duration': 5.204}, {'end': 1551.729, 'text': 'So those are a couple of things to note in here.', 'start': 1550.028, 'duration': 1.701}, {'end': 1560.143, 'text': 'And just like any other array in math, we can do like a concatenate or concatenate the different values here.', 'start': 1552.461, 'duration': 7.682}], 'summary': 'Transcript covers tensorflow import, version check, constant and variable creation, and tensor manipulation.', 'duration': 76.382, 'max_score': 1359.846, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41359846.jpg'}, {'end': 1595.009, 'src': 'embed', 'start': 1524.051, 'weight': 3, 'content': [{'end': 1529.772, 'text': "And that's interesting to note that this label is a TF dot tensor, and this is a TF variable.", 'start': 1524.051, 'duration': 5.721}, {'end': 1535.616, 'text': "So that's how it's looking in the back end when you're talking about the difference between a variable and a constant.", 'start': 1530.732, 'duration': 4.884}, {'end': 1543.863, 'text': 'The other thing I want you to notice is that in variable we capitalize the V and with the constant we have a lowercase c.', 'start': 1536.697, 'duration': 7.166}, {'end': 1549.067, 'text': "Little things like that can lose you when you're programming and you're trying to find out hey, why doesn't this work?", 'start': 1543.863, 'duration': 5.204}, {'end': 1551.729, 'text': 'So those are a couple of things to note in here.', 'start': 1550.028, 'duration': 1.701}, {'end': 1560.143, 'text': 'And just like any other array in math, we can do like a concatenate or concatenate the different values here.', 'start': 1552.461, 'duration': 7.682}, {'end': 1570.285, 'text': "And you can see we can take ab concatenated, you just do a tf.concat values and there's our ab axes on one.", 'start': 1560.903, 'duration': 9.382}, {'end': 1574.986, 'text': "Hopefully you're familiar with axes and how that works when you're dealing with matrices.", 'start': 1570.305, 'duration': 4.681}, {'end': 1580.625, 'text': "And if we go ahead and print this out, you'll see right here we end up with a tensor.", 'start': 1575.683, 'duration': 4.942}, {'end': 1583.545, 'text': "So let's put it in as a constant, not as a variable.", 'start': 1580.925, 'duration': 2.62}, {'end': 1588.267, 'text': 'And you have your array, 4, 3, 7, 8, and 6, 1, 4, 5.', 'start': 1584.326, 'duration': 3.941}, {'end': 1589.887, 'text': "It's concatenated the two together.", 'start': 1588.267, 'duration': 1.62}, {'end': 1593.128, 'text': 'And again, I want to highlight a couple things on this.', 'start': 1589.907, 'duration': 3.221}, {'end': 1595.009, 'text': 'Our axis equals 1.', 'start': 1593.428, 'duration': 1.581}], 'summary': 'Explanation of tf variable vs. constant, array concatenation, and axis in tensorflow.', 'duration': 70.958, 'max_score': 1524.051, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41524051.jpg'}, {'end': 1717.892, 'src': 'embed', 'start': 1691.167, 'weight': 6, 'content': [{'end': 1694.669, 'text': 'because conceptually your mind is just too many dimensions sometimes.', 'start': 1691.167, 'duration': 3.502}, {'end': 1698.193, 'text': 'The second thing I want you to notice is it says a numpy array.', 'start': 1695.629, 'duration': 2.564}, {'end': 1704.422, 'text': 'So TensorFlow is utilizing numpy as part of their format as far as Python is concerned.', 'start': 1698.774, 'duration': 5.648}, {'end': 1709.389, 'text': 'And so you can treat this output like a numpy array because it is just that.', 'start': 1705.083, 'duration': 4.306}, {'end': 1710.431, 'text': "It's going to be a numpy array.", 'start': 1709.409, 'duration': 1.022}, {'end': 1717.892, 'text': 'Another thing that comes up more than you would think is filling one of these with zeros or ones.', 'start': 1710.891, 'duration': 7.001}], 'summary': 'Tensorflow utilizes numpy array as part of its format in python for handling data.', 'duration': 26.725, 'max_score': 1691.167, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41691167.jpg'}, {'end': 1762.405, 'src': 'embed', 'start': 1733.135, 'weight': 7, 'content': [{'end': 1737.636, 'text': 'you can see, I have a nice array with shape 3, 4 of zeros.', 'start': 1733.135, 'duration': 4.501}, {'end': 1742.918, 'text': 'One of the things I want to highlight here is integer 32.', 'start': 1739.076, 'duration': 3.842}, {'end': 1753.682, 'text': 'If I go to the TensorFlow data types, I want you to notice how we have float 16, float 32, float 64, complex.', 'start': 1742.918, 'duration': 10.764}, {'end': 1756.763, 'text': "If we scroll down, you'll see the integer down here of 32.", 'start': 1753.742, 'duration': 3.021}, {'end': 1762.405, 'text': 'The reason for this is that we want to control how many bits are used in the precision.', 'start': 1756.763, 'duration': 5.642}], 'summary': 'Tensorflow data types include float 16, float 32, float 64, and complex, with the ability to control precision using 32-bit integers.', 'duration': 29.27, 'max_score': 1733.135, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41733135.jpg'}, {'end': 1848.901, 'src': 'embed', 'start': 1823.788, 'weight': 8, 'content': [{'end': 1829.612, 'text': 'So be very careful about starting a neural network for one of your rows or something like that with ones and zeros.', 'start': 1823.788, 'duration': 5.824}, {'end': 1833.794, 'text': 'On the other hand, I use this for masking.', 'start': 1830.492, 'duration': 3.302}, {'end': 1836.135, 'text': 'You can do a lot of work with masking.', 'start': 1834.254, 'duration': 1.881}, {'end': 1844.579, 'text': 'You can also have, it might be that one tensor row is masked, you know, zero is false, one is true, or whatever you want to do it.', 'start': 1836.175, 'duration': 8.404}, {'end': 1848.901, 'text': 'And so in that case, you do want to use the zeros and ones.', 'start': 1845.299, 'duration': 3.602}], 'summary': 'Neural networks use masking with ones and zeros to handle tensor rows effectively.', 'duration': 25.113, 'max_score': 1823.788, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41823788.jpg'}, {'end': 1923.098, 'src': 'heatmap', 'start': 1869.872, 'weight': 0.861, 'content': [{'end': 1873.795, 'text': 'So if we take our, remember this is shaped 3 comma 4.', 'start': 1869.872, 'duration': 3.923}, {'end': 1877.578, 'text': 'Maybe we want to swap that to 4 comma 3.', 'start': 1873.795, 'duration': 3.783}, {'end': 1881.642, 'text': 'And if we print this out, you will see, let me just go ahead and do that.', 'start': 1877.578, 'duration': 4.064}, {'end': 1883.403, 'text': 'Control V.', 'start': 1882.042, 'duration': 1.361}, {'end': 1884.164, 'text': 'Let me run that.', 'start': 1883.403, 'duration': 0.761}, {'end': 1887.726, 'text': "And you'll see that the order of these is now switched.", 'start': 1884.944, 'duration': 2.782}, {'end': 1890.248, 'text': 'Instead of 4 across, now we have 3 across and 4 down.', 'start': 1887.786, 'duration': 2.462}, {'end': 1895.632, 'text': "And just for fun, let's go back up here where we did the ones.", 'start': 1892.631, 'duration': 3.001}, {'end': 1901.533, 'text': "And I'm going to change the ones to tf.random uniform.", 'start': 1896.472, 'duration': 5.061}, {'end': 1903.413, 'text': "And we'll go ahead and just take off.", 'start': 1902.353, 'duration': 1.06}, {'end': 1904.574, 'text': "Well, we'll go ahead and leave that.", 'start': 1903.653, 'duration': 0.921}, {'end': 1905.594, 'text': "We'll go ahead and run this.", 'start': 1904.694, 'duration': 0.9}, {'end': 1908.174, 'text': "And you'll see now we have 0.0441.", 'start': 1906.634, 'duration': 1.54}, {'end': 1913.696, 'text': 'And this way, you can actually see how the reshape looks a lot different, 0.041, 0.15, 0.71.', 'start': 1908.174, 'duration': 5.522}, {'end': 1921.656, 'text': 'And then instead of having this one, it rolls down here to the 0.14.', 'start': 1913.696, 'duration': 7.96}, {'end': 1923.098, 'text': 'And this is what I was talking about.', 'start': 1921.656, 'duration': 1.442}], 'summary': 'Demonstrating reshaping and printing of data with tensorflow, showing 4x3 and 3x4 arrays, and using tf.random_uniform to generate different values.', 'duration': 53.226, 'max_score': 1869.872, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41869872.jpg'}, {'end': 1994.414, 'src': 'embed', 'start': 1965.326, 'weight': 9, 'content': [{'end': 1975.488, 'text': "or whatever you want to do There's a lot of or then convert it to an integer with one way to round it off, kind of a cheap and dirty trick.", 'start': 1965.326, 'duration': 10.162}, {'end': 1981.23, 'text': "So we can take this and we can take the same tensor and we'll go ahead and create a as an integer.", 'start': 1975.488, 'duration': 5.742}, {'end': 1994.414, 'text': "and so we're going to take this tensor, we're going to tf dot, cast it and if we print, And then we're going to go ahead and print our tensor.", 'start': 1981.23, 'duration': 13.184}], 'summary': 'Convert tensor to integer using casting and print', 'duration': 29.088, 'max_score': 1965.326, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41965326.jpg'}], 'start': 1359.846, 'title': 'Tensorflow basics and data manipulation', 'summary': 'Covers the basics of tensorflow version 2.1.0, including data manipulation, types, reshaping tensors, and practical demonstrations, emphasizing visual inspection and efficient computing control.', 'chapters': [{'end': 1671.509, 'start': 1359.846, 'title': 'Tensorflow basics', 'summary': 'Covers the basics of tensorflow, including importing the library, checking the version, creating constants and variables, concatenating arrays, and understanding axes, with a focus on tensorflow version 2.1.0 and practical demonstrations.', 'duration': 311.663, 'highlights': ["The chapter emphasizes the importance of importing TensorFlow correctly to avoid errors, as demonstrated by the speaker's habit of running the import command to ensure no errors are encountered. Emphasizing the significance of importing TensorFlow correctly to avoid errors", "The speaker demonstrates the process of checking the TensorFlow version, highlighting the use of 'tf.version' command to determine the working version, which is specifically noted as TensorFlow version 2.1.0. Demonstrating the process of checking the TensorFlow version using 'tf.version' command", 'The chapter provides a practical example of creating an array of constants using TensorFlow, showcasing the creation of a 2x2 array with the values 4, 3, 6, 1. Practical example of creating an array of constants with specific values', 'The speaker demonstrates the creation of a variable in TensorFlow, showcasing a 2x2 array and the process of printing the variable to confirm its creation. Demonstrating the creation of a variable in TensorFlow and the confirmation of its creation', 'The chapter explains the difference between constant and variable in TensorFlow, emphasizing the naming conventions and the distinction between TF tensor and TF variable, providing clarity on the programming concepts. Explaining the difference between constant and variable in TensorFlow, emphasizing naming conventions', 'The practical demonstration includes the process of concatenating arrays in TensorFlow, with a specific focus on axis 1, showcasing the resulting concatenated array with the provided values. Practical demonstration of concatenating arrays in TensorFlow with a focus on axis 1', 'The chapter further demonstrates the process of concatenating arrays, this time focusing on axis 0 to showcase the resulting concatenated array by rows, providing a comprehensive understanding of axis manipulation in TensorFlow. Demonstrating the process of concatenating arrays with a focus on axis 0']}, {'end': 2013.182, 'start': 1671.515, 'title': 'Tensorflow basics and data manipulation', 'summary': 'Covers the basics of tensorflow, including data manipulation, data types, and reshaping tensors, emphasizing the importance of visual inspection, the use of numpy arrays, and the control of data precision for efficient computing.', 'duration': 341.667, 'highlights': ['The importance of visual inspection when working with output data and the utilization of numpy arrays within TensorFlow for efficient data manipulation. The speaker emphasizes the importance of visually inspecting output data to ensure correct formatting and discusses the utilization of numpy arrays within TensorFlow for efficient data manipulation.', 'Control of data precision by specifying data types such as float 16, float 32, float 64, and integer 32 for efficient computation and platform compatibility. The speaker discusses the control of data precision by specifying data types such as float 16, float 32, float 64, and integer 32 within TensorFlow, emphasizing the importance for efficient computation and platform compatibility.', 'The use of zeros and ones in TensorFlow for tasks such as masking, initialization, and reshaping, cautioning against using them for neural network initialization due to potential bias. The speaker explains the use of zeros and ones in TensorFlow for tasks such as masking, initialization, and reshaping, cautioning against using them for neural network initialization due to potential bias.', 'The process of recasting data within TensorFlow, including converting float values to integers for specific project requirements, such as rounding off to a dollar amount. The speaker explains the process of recasting data within TensorFlow, including converting float values to integers for specific project requirements, such as rounding off to a dollar amount.']}], 'duration': 653.336, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D41359846.jpg', 'highlights': ['Emphasizing the significance of importing TensorFlow correctly to avoid errors', "Demonstrating the process of checking the TensorFlow version using 'tf.version' command", 'Practical example of creating an array of constants with specific values', 'Explaining the difference between constant and variable in TensorFlow, emphasizing naming conventions', 'Practical demonstration of concatenating arrays in TensorFlow with a focus on axis 1', 'Demonstrating the process of concatenating arrays with a focus on axis 0', 'The speaker emphasizes the importance of visually inspecting output data to ensure correct formatting and discusses the utilization of numpy arrays within TensorFlow for efficient data manipulation', 'The speaker discusses the control of data precision by specifying data types such as float 16, float 32, float 64, and integer 32 within TensorFlow, emphasizing the importance for efficient computation and platform compatibility', 'The speaker explains the use of zeros and ones in TensorFlow for tasks such as masking, initialization, and reshaping, cautioning against using them for neural network initialization due to potential bias', 'The speaker explains the process of recasting data within TensorFlow, including converting float values to integers for specific project requirements, such as rounding off to a dollar amount']}, {'end': 2742.006, 'segs': [{'end': 2204.248, 'src': 'embed', 'start': 2167.646, 'weight': 0, 'content': [{'end': 2176.823, 'text': "And that's where we get the 36 and 30.", 'start': 2167.646, 'duration': 9.177}, {'end': 2182.209, 'text': "Now I know we're covering a lot really quickly as far as the basic functionality.", 'start': 2176.823, 'duration': 5.386}, {'end': 2185.892, 'text': 'So the matrix, or your matrix multiplier,', 'start': 2183.41, 'duration': 2.482}, {'end': 2194.862, 'text': 'is a very commonly used backend tool as far as computing different models or linear regression stuff like that.', 'start': 2185.892, 'duration': 8.97}, {'end': 2202.787, 'text': 'One of the things is to note is that just like in NumPy, you have all of your different math.', 'start': 2195.642, 'duration': 7.145}, {'end': 2204.248, 'text': 'So we have our TF math.', 'start': 2202.947, 'duration': 1.301}], 'summary': 'Covering basic functionality including matrix multiplication, commonly used for computing models and linear regression.', 'duration': 36.602, 'max_score': 2167.646, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42167646.jpg'}, {'end': 2390.967, 'src': 'heatmap', 'start': 2337.435, 'weight': 0.889, 'content': [{'end': 2343.359, 'text': 'where we have the number of rows equals rows, the number of columns equals columns, and D type is a 32.', 'start': 2337.435, 'duration': 5.924}, {'end': 2355.769, 'text': 'And then if we go ahead and just print out our identity, you can see we have a nice identity column with our ones going across here.', 'start': 2343.359, 'duration': 12.41}, {'end': 2361.566, 'text': "Now. clearly we're not going to go through every math module available,", 'start': 2356.403, 'duration': 5.163}, {'end': 2366.628, 'text': 'but we do want to start looking at this as a prediction model and seeing how it functions.', 'start': 2361.566, 'duration': 5.062}, {'end': 2374.833, 'text': "So we're going to move on to more of a direct setup where you can actually see the full TensorFlow in use.", 'start': 2367.269, 'duration': 7.564}, {'end': 2379.115, 'text': "For that, let's go back and create a new setup.", 'start': 2375.373, 'duration': 3.742}, {'end': 2385.363, 'text': "And we'll go in here, new Python 3 module.", 'start': 2381.921, 'duration': 3.442}, {'end': 2387.004, 'text': 'There we go.', 'start': 2386.344, 'duration': 0.66}, {'end': 2390.967, 'text': 'Bring this out so it takes up the whole window because I like to do that.', 'start': 2388.325, 'duration': 2.642}], 'summary': 'Introduction to tensorflow with 32-bit data type and identity column demonstration for prediction model.', 'duration': 53.532, 'max_score': 2337.435, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42337435.jpg'}, {'end': 2467.124, 'src': 'embed', 'start': 2436.239, 'weight': 3, 'content': [{'end': 2439.56, 'text': "We'll talk about that if it comes up on this particular setup.", 'start': 2436.239, 'duration': 3.321}, {'end': 2440.861, 'text': 'And of course date time.', 'start': 2439.88, 'duration': 0.981}, {'end': 2444.083, 'text': 'Pandas, again, is your data frame.', 'start': 2441.801, 'duration': 2.282}, {'end': 2445.785, 'text': 'Think rows and columns.', 'start': 2444.624, 'duration': 1.161}, {'end': 2447.006, 'text': 'We import it as PD.', 'start': 2445.905, 'duration': 1.101}, {'end': 2453.392, 'text': 'NumPy is your numbers array, which, of course, TensorFlow is integrated heavily with.', 'start': 2447.747, 'duration': 5.645}, {'end': 2461.86, 'text': 'Seaborn for our graphics, and the Seaborn as SNS is going to be set on top of our Matplot library, which we import as MPL.', 'start': 2454.213, 'duration': 7.647}, {'end': 2467.124, 'text': "and then of course, we're going to import our matplotlib pipe plot as plt.", 'start': 2462.481, 'duration': 4.643}], 'summary': 'Using pandas for data frames, numpy for arrays, seaborn for graphics in this setup.', 'duration': 30.885, 'max_score': 2436.239, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42436239.jpg'}, {'end': 2594.507, 'src': 'embed', 'start': 2565.076, 'weight': 1, 'content': [{'end': 2566.897, 'text': 'And then we have our cross models.', 'start': 2565.076, 'duration': 1.821}, {'end': 2568.177, 'text': "We're going to import sequential.", 'start': 2566.917, 'duration': 1.26}, {'end': 2573.778, 'text': 'Now, if you remember from our slide, there was three different options.', 'start': 2568.797, 'duration': 4.981}, {'end': 2577.419, 'text': 'Let me just flip back over there so we can have a quick recall on that.', 'start': 2573.938, 'duration': 3.481}, {'end': 2583.901, 'text': 'And so in cross, we have sequential, functional, and subclassing.', 'start': 2577.759, 'duration': 6.142}, {'end': 2587.502, 'text': 'So remember those three different setups in here we talked about earlier.', 'start': 2584.081, 'duration': 3.421}, {'end': 2594.507, 'text': "And if you remember from here, we have sequential, where it's going one tensor flow layer at a time.", 'start': 2587.842, 'duration': 6.665}], 'summary': 'Discussing three different options for cross models: sequential, functional, and subclassing.', 'duration': 29.431, 'max_score': 2565.076, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42565076.jpg'}, {'end': 2685.932, 'src': 'embed', 'start': 2652.758, 'weight': 2, 'content': [{'end': 2655.362, 'text': 'So from Keras we have layers import dents.', 'start': 2652.758, 'duration': 2.604}, {'end': 2660.067, 'text': 'from Keras Layers Import LSTM.', 'start': 2656.122, 'duration': 3.945}, {'end': 2666.233, 'text': 'When we talk about these layers, Keras has so many built-in layers, you can do your own layers.', 'start': 2660.507, 'duration': 5.726}, {'end': 2670.297, 'text': 'The dense layer is your standard neural network.', 'start': 2666.874, 'duration': 3.423}, {'end': 2674.582, 'text': 'By default, it uses ReLU for its activation.', 'start': 2671.038, 'duration': 3.544}, {'end': 2680.027, 'text': 'And then the LSTM is the long short term memory layer.', 'start': 2675.803, 'duration': 4.224}, {'end': 2685.932, 'text': "Since we're going to be looking probably at sequential data, we want to go ahead and do the LSTM.", 'start': 2680.307, 'duration': 5.625}], 'summary': 'Keras offers various layers including dense and lstm for neural networks, with default relu activation.', 'duration': 33.174, 'max_score': 2652.758, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42652758.jpg'}], 'start': 2014.002, 'title': 'Tensorflow math functions and keras sequential model', 'summary': 'Covers tensorflow math functions including float to integer conversion, reshaping, transposing tensors, matrix and bitwise multiplication, and creating an identity matrix. it also introduces importing keras and the sequential model in tensorflow, explaining the options available for building neural networks and pre-built layers like dense and lstm.', 'chapters': [{'end': 2542.403, 'start': 2014.002, 'title': 'Tensorflow math functions and setups', 'summary': 'Covers the usage of math functions in tensorflow, including the conversion of float to integer, reshaping, transposing tensors, matrix multiplication, bitwise multiplication, and creating an identity matrix. it also introduces the setup for using tensorflow with other libraries and modules.', 'duration': 528.401, 'highlights': ['The chapter covers the usage of math functions in TensorFlow, including the conversion of float to integer, reshaping, transposing tensors, matrix multiplication, bitwise multiplication, and creating an identity matrix. This encompasses the fundamental math operations in TensorFlow, such as conversion, reshaping, transposing, and matrix operations, providing a comprehensive understanding of the mathematical capabilities of TensorFlow.', 'The setup for using TensorFlow with other libraries and modules is introduced, including imports of NumPy, Pandas, Seaborn, and Matplotlib. The chapter provides insight into the integration of TensorFlow with other essential libraries and modules, demonstrating the initial setup required for utilizing TensorFlow in conjunction with these tools.', 'The usage of warnings and filtering them out is demonstrated, emphasizing the importance of handling warnings in a project setup. It highlights the practical aspect of handling warnings in a project setup, showcasing the significance of effectively managing warnings to ensure a smooth development process.']}, {'end': 2742.006, 'start': 2542.563, 'title': 'Importing keras and sequential model in tensorflow', 'summary': 'Discusses the process of importing keras and the sequential model in tensorflow, explaining the different options available in keras for building neural networks, such as sequential, functional, and subclassing, along with the pre-built layers like dense and lstm.', 'duration': 199.443, 'highlights': ['Keras sits on TensorFlow, so importing Keras and the sequential model effectively imports TensorFlow underneath it. Importing Keras and the sequential model means importing TensorFlow underneath it, highlighting the relationship between Keras and TensorFlow.', 'Keras offers three different options for building neural networks: sequential, functional, and subclassing. The chapter explains the three different options in Keras for building neural networks, providing an overview of the available choices.', 'The dense layer is a standard neural network layer that uses ReLU for its activation, while LSTM is the long short term memory layer suitable for handling sequential data. Explanation and differentiation of the dense layer as a standard neural network layer using ReLU activation and the LSTM layer designed for sequential data.', "Keras provides pre-built layers, including dense and LSTM, as well as the capability to create custom layers, making it popular for its comprehensive functionality. The discussion highlights Keras' pre-built layers, such as dense and LSTM, and its capacity for creating custom layers, contributing to its popularity."]}], 'duration': 728.004, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42014002.jpg', 'highlights': ['The chapter covers the usage of math functions in TensorFlow, including the conversion of float to integer, reshaping, transposing tensors, matrix multiplication, bitwise multiplication, and creating an identity matrix.', 'Keras offers three different options for building neural networks: sequential, functional, and subclassing.', 'Keras provides pre-built layers, including dense and LSTM, as well as the capability to create custom layers, making it popular for its comprehensive functionality.', 'The setup for using TensorFlow with other libraries and modules is introduced, including imports of NumPy, Pandas, Seaborn, and Matplotlib.']}, {'end': 3403.164, 'segs': [{'end': 2769.353, 'src': 'embed', 'start': 2742.066, 'weight': 0, 'content': [{'end': 2745.129, 'text': "It's open source, and you have all these tools right at your fingertips.", 'start': 2742.066, 'duration': 3.063}, {'end': 2752.155, 'text': "So from Keras, we're just going to import a couple layers, the dense layer and the long short-term memory layer.", 'start': 2745.629, 'duration': 6.526}, {'end': 2762.966, 'text': 'And then, of course, from sklearn, our scikit, we want to go ahead and do our min-max scalar, standard scalar for pre-editing our data,', 'start': 2753.036, 'duration': 9.93}, {'end': 2766.649, 'text': 'and then metrics, just so we can take a look at the errors and compute those.', 'start': 2762.966, 'duration': 3.683}, {'end': 2768.251, 'text': 'Let me go ahead and run this.', 'start': 2767.27, 'duration': 0.981}, {'end': 2769.353, 'text': 'That just loads it up.', 'start': 2768.472, 'duration': 0.881}], 'summary': 'Utilizing open-source tools like keras and scikit-learn for data pre-processing and error computation.', 'duration': 27.287, 'max_score': 2742.066, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42742066.jpg'}, {'end': 2812.6, 'src': 'embed', 'start': 2785.174, 'weight': 1, 'content': [{'end': 2789.615, 'text': 'And you can see here we have a number of columns, a number of rows.', 'start': 2785.174, 'duration': 4.441}, {'end': 2792.235, 'text': 'It actually goes down to like 8, 000.', 'start': 2789.695, 'duration': 2.54}, {'end': 2798.257, 'text': 'The first thing we want to notice is that the first row is kind of just a random number put in going down.', 'start': 2792.236, 'duration': 6.021}, {'end': 2800.558, 'text': "Probably not something we're going to work with.", 'start': 2798.817, 'duration': 1.741}, {'end': 2803.938, 'text': 'The second row is Bandung.', 'start': 2801.238, 'duration': 2.7}, {'end': 2806.679, 'text': "I'm guessing that's a reference for the profile.", 'start': 2804.518, 'duration': 2.161}, {'end': 2812.6, 'text': "If we scroll to the bottom, which I'm not going to do because it takes forever to get back up, They're all the same.", 'start': 2807.359, 'duration': 5.241}], 'summary': 'Data table with 8000 rows, last row contains repetitive data.', 'duration': 27.426, 'max_score': 2785.174, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42785174.jpg'}, {'end': 2910.458, 'src': 'embed', 'start': 2874.588, 'weight': 2, 'content': [{'end': 2878.511, 'text': "And so if we go ahead and run this, we'll print out the head of our data.", 'start': 2874.588, 'duration': 3.923}, {'end': 2881.533, 'text': 'And again, this looks very similar to what we were just looking at.', 'start': 2878.731, 'duration': 2.802}, {'end': 2886.177, 'text': 'Being in Jupyter, I can take this and go the other way.', 'start': 2882.774, 'duration': 3.403}, {'end': 2890.1, 'text': 'make it real small, so you can see all the columns going across and we can get a full view of it.', 'start': 2886.177, 'duration': 3.923}, {'end': 2893.763, 'text': 'Or we can bring it back up in size.', 'start': 2892.122, 'duration': 1.641}, {'end': 2895.425, 'text': "That's pretty small on there.", 'start': 2894.484, 'duration': 0.941}, {'end': 2899.748, 'text': "Overshot But you can see it's the same data we were just looking at.", 'start': 2895.965, 'duration': 3.783}, {'end': 2910.458, 'text': "We're looking at the number, we're looking at the profile, which is the Bandung, the date, we have a timestamp, our 03 count CO and so forth on here.", 'start': 2900.209, 'duration': 10.249}], 'summary': 'Demonstration of resizing and displaying data in jupyter, with a focus on columns and data visualization.', 'duration': 35.87, 'max_score': 2874.588, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42874588.jpg'}, {'end': 3017.551, 'src': 'heatmap', 'start': 2958.636, 'weight': 0.85, 'content': [{'end': 2963.459, 'text': "When we go ahead and do this, that's all of our information that we want to go ahead and create.", 'start': 2958.636, 'duration': 4.823}, {'end': 2966.902, 'text': 'And then it goes for I in range, DF shape 0.', 'start': 2964.06, 'duration': 2.842}, {'end': 2972.505, 'text': "So we're going to go through the whole setup and we're going to list tab append, DF location I.", 'start': 2966.902, 'duration': 5.603}, {'end': 2976.989, 'text': 'And here is our date going in there.', 'start': 2973.806, 'duration': 3.183}, {'end': 2979.791, 'text': 'And then return the numpy array list tab.', 'start': 2977.529, 'duration': 2.262}, {'end': 2981.893, 'text': 'D types date time 64.', 'start': 2979.991, 'duration': 1.902}, {'end': 2982.494, 'text': "That's all we're doing.", 'start': 2981.893, 'duration': 0.601}, {'end': 2984.355, 'text': "We're just switching this to a date time stamp.", 'start': 2982.514, 'duration': 1.841}, {'end': 2988.979, 'text': 'And if we go ahead and do df date time equals combined date time.', 'start': 2984.555, 'duration': 4.424}, {'end': 2991.942, 'text': 'And then I always like to print.', 'start': 2989.62, 'duration': 2.322}, {'end': 2995.064, 'text': "We'll do df head just so we can see what that looks like.", 'start': 2992.102, 'duration': 2.962}, {'end': 2999.746, 'text': 'And so when we come out of this, we now have our setup on here.', 'start': 2996.005, 'duration': 3.741}, {'end': 3001.627, 'text': "And of course, it's added it on to the far right.", 'start': 2999.806, 'duration': 1.821}, {'end': 3002.587, 'text': "Here's our date time.", 'start': 3001.687, 'duration': 0.9}, {'end': 3004.707, 'text': "You can see the format's changed.", 'start': 3003.227, 'duration': 1.48}, {'end': 3008.368, 'text': "So we've added in the date time column.", 'start': 3005.588, 'duration': 2.78}, {'end': 3010.169, 'text': "And we've brought the date over.", 'start': 3008.829, 'duration': 1.34}, {'end': 3012.249, 'text': "And we've taken this format here.", 'start': 3010.569, 'duration': 1.68}, {'end': 3017.551, 'text': "And it's an actual variable with a 000 on here.", 'start': 3012.269, 'duration': 5.282}], 'summary': 'Creating a new date time column in dataframe using python, resulting in a format change and additional variable with 000.', 'duration': 58.915, 'max_score': 2958.636, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42958636.jpg'}, {'end': 3048.014, 'src': 'embed', 'start': 3018.111, 'weight': 3, 'content': [{'end': 3019.011, 'text': "Well, that doesn't look good.", 'start': 3018.111, 'duration': 0.9}, {'end': 3022.152, 'text': 'So we need to also include the time part of this.', 'start': 3019.371, 'duration': 2.781}, {'end': 3024.693, 'text': 'And we want to convert it into hourly data.', 'start': 3022.172, 'duration': 2.521}, {'end': 3027.215, 'text': "So let's go ahead and do that.", 'start': 3025.914, 'duration': 1.301}, {'end': 3035.684, 'text': "To do that, to finish combining our date time, let's go ahead and create a little script here to combine the time in there.", 'start': 3027.916, 'duration': 7.768}, {'end': 3043.113, 'text': "Same thing we just did, we're just creating a numpy array, returning a numpy array and forcing this into a date time format.", 'start': 3036.445, 'duration': 6.668}, {'end': 3048.014, 'text': 'And we can actually spend hours just going through these conversions.', 'start': 3043.493, 'duration': 4.521}], 'summary': 'Converting data to hourly format using numpy array and date time format.', 'duration': 29.903, 'max_score': 3018.111, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43018111.jpg'}, {'end': 3216.914, 'src': 'embed', 'start': 3188.96, 'weight': 5, 'content': [{'end': 3194.722, 'text': 'And this gives us our, you know, describe, gives us our basic data analytics, information we might be looking for,', 'start': 3188.96, 'duration': 5.762}, {'end': 3200.324, 'text': 'like what is the mean standard deviation, minimum amount, maximum amount?', 'start': 3194.722, 'duration': 5.602}, {'end': 3205.686, 'text': 'We have our first quarter, second quarter, and third quarter numbers also in there.', 'start': 3200.504, 'duration': 5.182}, {'end': 3210.892, 'text': 'So you can get a quick look at a glance describing the data or descriptive analysis.', 'start': 3206.506, 'duration': 4.386}, {'end': 3216.914, 'text': "And even though we have our quantile information in here, we're going to dig a little deeper into that.", 'start': 3211.573, 'duration': 5.341}], 'summary': 'Data analytics summary includes mean, standard deviation, min, max, and quarterly numbers for deeper analysis.', 'duration': 27.954, 'max_score': 3188.96, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43188960.jpg'}], 'start': 2742.066, 'title': 'Data preprocessing and data analysis', 'summary': 'Covers the usage of open-source tools like keras and scikit-learn for data preprocessing, including importing layers, standard scaling, and computing metrics, and demonstrates the examination of a dataset for preprocessing. additionally, it discusses the process of reading air quality data into a pandas data frame, combining date and time columns, reorganizing the data into hourly format, and conducting data analytics using pandas to calculate mean, standard deviation, minimum, and maximum values as part of data preprocessing. the dataset features approximately 8,000 rows and columns, and involves identifying irrelevant rows and columns, recognizing sequential and timestamp data.', 'chapters': [{'end': 2826.903, 'start': 2742.066, 'title': 'Data preprocessing and tool usage', 'summary': 'Covers the usage of open-source tools like keras and scikit-learn for data preprocessing, including importing layers, standard scaling, and computing metrics, and demonstrates the examination of a dataset for preprocessing, featuring a file with approximately 8,000 rows and columns, identifying irrelevant rows and columns, and recognizing sequential and timestamp data.', 'duration': 84.837, 'highlights': ['Demonstration of data preprocessing using open-source tools like Keras and scikit-learn, including importing layers, standard scaling, and computing metrics. Usage of Keras and scikit-learn for data preprocessing.', 'Examination of a dataset with approximately 8,000 rows and columns, identifying irrelevant rows and columns, and recognizing sequential and timestamp data. Dataset with around 8,000 rows and columns, identification of irrelevant rows and columns, recognition of sequential and timestamp data.']}, {'end': 3403.164, 'start': 2827.763, 'title': 'Data analysis and pandas data frame', 'summary': 'Discusses the process of reading air quality data into a pandas data frame, combining date and time columns, reorganizing the data into hourly format, and conducting data analytics using pandas to calculate mean, standard deviation, minimum, and maximum values as part of data preprocessing.', 'duration': 575.401, 'highlights': ['The process of reading air quality data into a pandas data frame is discussed, including the use of pd.readcsv to load the data and printing the head of the data frame. Loading data into a pandas data frame, printing the head of the data frame.', 'The combination of date and time columns is demonstrated using a simple script, resulting in the creation of a new date time column in the data frame. Combining date and time columns, creating a new date time column.', 'The reorganization of the data into hourly format is explained, involving the creation of a script to combine the time and the grouping of data by date time to obtain mean values. Reorganizing data into hourly format, grouping data by date time to calculate mean values.', 'The process of conducting data analytics using Pandas to calculate mean, standard deviation, minimum, and maximum values is detailed, including the use of describe and quantile functions. Calculating mean, standard deviation, minimum, and maximum values using Pandas describe and quantile functions.']}], 'duration': 661.098, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D42742066.jpg', 'highlights': ['Usage of Keras and scikit-learn for data preprocessing.', 'Dataset with around 8,000 rows and columns, identification of irrelevant rows and columns, recognition of sequential and timestamp data.', 'Loading data into a pandas data frame, printing the head of the data frame.', 'Combining date and time columns, creating a new date time column.', 'Reorganizing data into hourly format, grouping data by date time to calculate mean values.', 'Calculating mean, standard deviation, minimum, and maximum values using Pandas describe and quantile functions.']}, {'end': 3991.417, 'segs': [{'end': 3444.362, 'src': 'embed', 'start': 3403.164, 'weight': 0, 'content': [{'end': 3409.409, 'text': "and the reason we've computed these numbers is if you take the minimum value and it's less than your minimum IQR,", 'start': 3403.164, 'duration': 6.245}, {'end': 3412.988, 'text': "That means something's going wrong there.", 'start': 3411.447, 'duration': 1.541}, {'end': 3415.649, 'text': "Usually, in this case, it's going to show us an outlier.", 'start': 3413.568, 'duration': 2.081}, {'end': 3418.411, 'text': 'So we want to go ahead and find the minimum value.', 'start': 3416.41, 'duration': 2.001}, {'end': 3422.033, 'text': "If it's less than the minimum IQR, it's an outlier.", 'start': 3418.871, 'duration': 3.162}, {'end': 3427.656, 'text': 'And if the max value is greater than the max IQR, we have an outlier.', 'start': 3422.113, 'duration': 5.543}, {'end': 3429.016, 'text': "And that's all this is doing.", 'start': 3428.016, 'duration': 1}, {'end': 3430.777, 'text': 'Low outlier is found.', 'start': 3429.557, 'duration': 1.22}, {'end': 3433.299, 'text': 'Minimum value, high outlier is found.', 'start': 3431.298, 'duration': 2.001}, {'end': 3434.899, 'text': 'really important.', 'start': 3434.279, 'duration': 0.62}, {'end': 3437.18, 'text': 'actually, outliers are almost everything in data.', 'start': 3434.899, 'duration': 2.281}, {'end': 3443.742, 'text': 'sometimes, sometimes you do this project just to find the outliers because you want to know crime detection.', 'start': 3437.18, 'duration': 6.562}, {'end': 3444.362, 'text': 'what are we looking for?', 'start': 3443.742, 'duration': 0.62}], 'summary': 'Identifying outliers using minimum and maximum values to detect anomalies in data.', 'duration': 41.198, 'max_score': 3403.164, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43403164.jpg'}, {'end': 3571.906, 'src': 'embed', 'start': 3541.711, 'weight': 2, 'content': [{'end': 3544.933, 'text': 'We found out that humidity probably has some weird values in it.', 'start': 3541.711, 'duration': 3.222}, {'end': 3546.694, 'text': 'We have our outliers.', 'start': 3545.593, 'duration': 1.101}, {'end': 3549.635, 'text': "That's all this is.", 'start': 3548.555, 'duration': 1.08}, {'end': 3560.34, 'text': 'And so when we go ahead and finish this and we take a look at our outliers and we run this code here, we have a low outlier 2.04.', 'start': 3550.295, 'duration': 10.045}, {'end': 3561.56, 'text': 'We have a high outlier 99.06.', 'start': 3560.34, 'duration': 1.22}, {'end': 3562.721, 'text': 'Outliers have been interpolated.', 'start': 3561.56, 'duration': 1.161}, {'end': 3571.906, 'text': "That means we've given them a new value.", 'start': 3570.245, 'duration': 1.661}], 'summary': 'Identified outliers in humidity: low outlier 2.04, high outlier 99.06, and interpolated new values.', 'duration': 30.195, 'max_score': 3541.711, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43541711.jpg'}, {'end': 3686.767, 'src': 'embed', 'start': 3661.486, 'weight': 3, 'content': [{'end': 3668.751, 'text': "It's interesting because it's like how do you even know to use logarithmic on the temp value? That's domain specific.", 'start': 3661.486, 'duration': 7.265}, {'end': 3672.273, 'text': "We're talking about being an expert in air care.", 'start': 3669.531, 'duration': 2.742}, {'end': 3673.674, 'text': "I'm not an expert in air care.", 'start': 3672.553, 'duration': 1.121}, {'end': 3676.363, 'text': "You know, it's not what I go look at.", 'start': 3674.282, 'duration': 2.081}, {'end': 3677.584, 'text': "I don't look at air care data.", 'start': 3676.403, 'duration': 1.181}, {'end': 3679.924, 'text': "In fact, this is probably the first air care data setup I've looked at.", 'start': 3677.604, 'duration': 2.32}, {'end': 3686.767, 'text': 'But the experts come in there and they come to you and say, hey, in data science, this is exponentially variable on here.', 'start': 3680.125, 'duration': 6.642}], 'summary': 'Experts highlight the domain-specific use of logarithmic and exponential variables in air care data for data science.', 'duration': 25.281, 'max_score': 3661.486, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43661486.jpg'}, {'end': 3798.033, 'src': 'embed', 'start': 3757.479, 'weight': 4, 'content': [{'end': 3758.26, 'text': 'And this is kind of nice.', 'start': 3757.479, 'duration': 0.781}, {'end': 3759.901, 'text': 'Plot a figure, figure size.', 'start': 3758.38, 'duration': 1.521}, {'end': 3762.023, 'text': "Here's our PLT from matplotlib.", 'start': 3759.942, 'duration': 2.081}, {'end': 3766.007, 'text': "And then we'll just do a distribution underscore df.", 'start': 3762.964, 'duration': 3.043}, {'end': 3767.268, 'text': "There's our data frames.", 'start': 3766.227, 'duration': 1.041}, {'end': 3771.692, 'text': 'This is nice because it just integrates the histogram right into pandas.', 'start': 3768.068, 'duration': 3.624}, {'end': 3772.893, 'text': 'Love pandas.', 'start': 3772.072, 'duration': 0.821}, {'end': 3780.501, 'text': 'And this is a chart you would send back to your data analysis and say, hey, is this what you wanted? This is how the data is converting on here.', 'start': 3773.554, 'duration': 6.947}, {'end': 3785.786, 'text': "As a data scientist, the first thing I note is we've gone from a 10, 20, 30 scale to 2.5, 3.0, 3.5 scale.", 'start': 3780.821, 'duration': 4.965}, {'end': 3798.033, 'text': 'And the data itself has kind of been adjusted a little bit based on some kind of a skew on there.', 'start': 3790.39, 'duration': 7.643}], 'summary': 'Integrating histogram into pandas for data analysis, scaling from 10, 20, 30 to 2.5, 3.0, 3.5.', 'duration': 40.554, 'max_score': 3757.479, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43757479.jpg'}, {'end': 3823.951, 'src': 'embed', 'start': 3798.513, 'weight': 5, 'content': [{'end': 3804.135, 'text': "So let's jump into, we're getting a little closer to actually doing our Kerasan here.", 'start': 3798.513, 'duration': 5.622}, {'end': 3807.817, 'text': "We'll go ahead and split our data up.", 'start': 3805.675, 'duration': 2.142}, {'end': 3811.2, 'text': 'And this, of course, is any good data scientist.', 'start': 3808.638, 'duration': 2.562}, {'end': 3814.062, 'text': 'You want to have a training set and a test set.', 'start': 3811.88, 'duration': 2.182}, {'end': 3816.925, 'text': "And we'll go ahead and do the train size.", 'start': 3814.583, 'duration': 2.342}, {'end': 3820.388, 'text': "We're going to use 0.75% of the data.", 'start': 3818.006, 'duration': 2.382}, {'end': 3821.669, 'text': "Make sure it's an integer.", 'start': 3820.748, 'duration': 0.921}, {'end': 3823.951, 'text': "You don't want to take a slice as a float value.", 'start': 3821.689, 'duration': 2.262}], 'summary': 'Split data into training and test sets, using 75% for training.', 'duration': 25.438, 'max_score': 3798.513, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43798513.jpg'}, {'end': 3896.902, 'src': 'embed', 'start': 3868.239, 'weight': 6, 'content': [{'end': 3873.842, 'text': "So we're looking at just that column with this train size and the test with the train and test data set here.", 'start': 3868.239, 'duration': 5.603}, {'end': 3879.129, 'text': "And let's go ahead and convert an array of values into a data set matrix.", 'start': 3874.226, 'duration': 4.903}, {'end': 3882.272, 'text': "We're going to create a little setup in here.", 'start': 3879.43, 'duration': 2.842}, {'end': 3883.272, 'text': 'We create our data set.', 'start': 3882.292, 'duration': 0.98}, {'end': 3884.533, 'text': 'Our data set is going to come in.', 'start': 3883.372, 'duration': 1.161}, {'end': 3885.754, 'text': "We're going to do a look back of 1.", 'start': 3884.553, 'duration': 1.201}, {'end': 3888.676, 'text': "So we're going to look back one piece of data going backward.", 'start': 3885.754, 'duration': 2.922}, {'end': 3896.902, 'text': 'And we have our data x and our data y for ion range, length of data set, look back minus 1.', 'start': 3888.696, 'duration': 8.206}], 'summary': 'Data set matrix created with a look back of 1 piece of data.', 'duration': 28.663, 'max_score': 3868.239, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43868239.jpg'}], 'start': 3403.164, 'title': 'Data analysis and outlier treatment in data', 'summary': 'Covers the identification and treatment of outliers in data, and the use of data interpolation techniques to address outlier values, demonstrated by interpolating humidity outlier values 2.04 and 99.06 based on a linear regression model. it also explains the process of data analysis, including creating histograms and setting up a kerasan model to predict future data points based on previous values using a look back of one piece of data.', 'chapters': [{'end': 3732.042, 'start': 3403.164, 'title': 'Data outliers and data preparation', 'summary': 'Discusses the identification and treatment of outliers in data, emphasizing their significance and impact on data analysis, as well as the use of data interpolation techniques to address outlier values, exemplified by the humidity outlier values 2.04 and 99.06 being interpolated based on a linear regression model.', 'duration': 328.878, 'highlights': ['The chapter emphasizes the significance of outliers in data analysis, highlighting their impact on data interpretation and the importance of detecting and addressing them to ensure accurate results. ', 'The process of identifying outliers involves comparing the minimum and maximum values with the minimum and maximum Interquartile Range (IQR), where values falling outside these ranges are considered as outliers. ', 'The transcript showcases the use of data interpolation techniques to address outlier values, with the example of humidity outlier values 2.04 and 99.06 being interpolated based on a linear regression model. ', 'The importance of domain expertise is highlighted, as experts from the air care domain recommend using a logarithmic scale on temperature values, demonstrating the domain-specific knowledge required for effective data analysis. ']}, {'end': 3991.417, 'start': 3732.042, 'title': 'Data analysis and kerasan setup', 'summary': 'Explains the process of data analysis, including creating histograms and splitting the data into training and test sets, followed by setting up a kerasan model to predict future data points based on previous values, with a focus on using a look back of one piece of data.', 'duration': 259.375, 'highlights': ['Creating Histograms with Pandas The chapter demonstrates the integration of histograms into pandas using matplotlib, providing a visual representation of data distribution, facilitating analysis and decision-making.', 'Splitting Data into Training and Test Sets The process involves splitting the data into a training set (75% of the data) and a test set, ensuring a proper distribution for model training and evaluation.', 'Setting Up Kerasan Model with Look Back of 1 The chapter outlines the creation of a data set matrix with a look back of 1, aiming to predict the next data point based on the preceding value, illustrating the predictive nature of the Kerasan model.']}], 'duration': 588.253, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43403164.jpg', 'highlights': ['The chapter emphasizes the significance of outliers in data analysis and the importance of detecting and addressing them for accurate results.', 'The process of identifying outliers involves comparing the minimum and maximum values with the minimum and maximum Interquartile Range (IQR).', 'The chapter showcases the use of data interpolation techniques to address outlier values, with the example of humidity outlier values 2.04 and 99.06 being interpolated based on a linear regression model.', 'The importance of domain expertise is highlighted, as experts from the air care domain recommend using a logarithmic scale on temperature values, demonstrating the domain-specific knowledge required for effective data analysis.', 'Creating Histograms with Pandas provides a visual representation of data distribution, facilitating analysis and decision-making.', 'Splitting Data into Training and Test Sets ensures a proper distribution for model training and evaluation.', 'Setting Up Kerasan Model with Look Back of 1 illustrates the predictive nature of the Kerasan model.']}, {'end': 4504.108, 'segs': [{'end': 4087.738, 'src': 'embed', 'start': 4060.539, 'weight': 0, 'content': [{'end': 4064.144, 'text': 'We added the one dimension, and then we have our train X shape one.', 'start': 4060.539, 'duration': 3.605}, {'end': 4069.11, 'text': 'And this could have, depends again on how far back in the long, short-term memory you want to go.', 'start': 4064.845, 'duration': 4.265}, {'end': 4072.872, 'text': 'That is what that piece of code is for and that reshape is.', 'start': 4069.831, 'duration': 3.041}, {'end': 4079.235, 'text': 'And you can see the new shape is now 129911 versus the 12991.', 'start': 4073.312, 'duration': 5.923}, {'end': 4083.716, 'text': 'And then the other part of the shape, 43211.', 'start': 4079.235, 'duration': 4.481}, {'end': 4087.738, 'text': 'Again, this is our X in and, of course, our test X.', 'start': 4083.716, 'duration': 4.022}], 'summary': 'Reshaped data to 129911 and 43211 dimensions for training and testing.', 'duration': 27.199, 'max_score': 4060.539, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44060539.jpg'}, {'end': 4136.563, 'src': 'embed', 'start': 4109.809, 'weight': 1, 'content': [{'end': 4113.612, 'text': "and if we're going to do that, the first thing we need is we're going to need to go ahead and create a model.", 'start': 4109.809, 'duration': 3.803}, {'end': 4119.555, 'text': "and we'll do this sequential model, and if you remember, sequential means it just goes in order.", 'start': 4113.612, 'duration': 5.943}, {'end': 4121.336, 'text': 'that means we have, if you have, two layers.', 'start': 4119.555, 'duration': 1.781}, {'end': 4126.258, 'text': 'The layers go from layer 1 to layer 2 or layer 0 to layer 1..', 'start': 4121.836, 'duration': 4.422}, {'end': 4127.779, 'text': 'This is different than functional.', 'start': 4126.258, 'duration': 1.521}, {'end': 4134.282, 'text': 'Functional allows you to split the data and run two completely separate models and then bring them back together.', 'start': 4128.359, 'duration': 5.923}, {'end': 4136.563, 'text': "We're doing just sequential on here.", 'start': 4134.942, 'duration': 1.621}], 'summary': 'Creating a sequential model with two layers for data processing.', 'duration': 26.754, 'max_score': 4109.809, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44109809.jpg'}, {'end': 4204.674, 'src': 'embed', 'start': 4173.399, 'weight': 2, 'content': [{'end': 4175.661, 'text': "And then we're going to add the LSTM model.", 'start': 4173.399, 'duration': 2.262}, {'end': 4177.823, 'text': "And then we're going to add a dense model.", 'start': 4175.72, 'duration': 2.103}, {'end': 4183.413, 'text': 'And if you remember, from the very top of our code, we did the imports.', 'start': 4178.403, 'duration': 5.01}, {'end': 4185.357, 'text': 'oops, there we go our cross.', 'start': 4183.413, 'duration': 1.944}, {'end': 4186.278, 'text': 'this is it right here?', 'start': 4185.357, 'duration': 0.921}, {'end': 4190.863, 'text': "here's our importing a dense model and here's our importing an lstm.", 'start': 4186.278, 'duration': 4.585}, {'end': 4195.39, 'text': 'now, just about every tensorflow model uses dents.', 'start': 4190.863, 'duration': 4.527}, {'end': 4204.674, 'text': 'your dense model is your basic forward propagation reverse propagation error, or it does reverse propagation to program the model?', 'start': 4195.39, 'duration': 9.284}], 'summary': 'Adding lstm and dense models for tensorflow with imports.', 'duration': 31.275, 'max_score': 4173.399, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44173399.jpg'}, {'end': 4242.216, 'src': 'embed', 'start': 4215.497, 'weight': 3, 'content': [{'end': 4219.958, 'text': 'The real question that we want to look at right now is where do you find these models?', 'start': 4215.497, 'duration': 4.461}, {'end': 4221.458, 'text': 'What kind of models do you have available?', 'start': 4220.018, 'duration': 1.44}, {'end': 4227.724, 'text': "and so for that let's go to the cross website, which is the cross.io.", 'start': 4221.878, 'duration': 5.846}, {'end': 4242.216, 'text': "if you go under api slash layers and i always have to do a search just search for cross api layers it'll open up and you can see we have Your base layers right here class trainable weights,", 'start': 4227.724, 'duration': 14.492}], 'summary': 'Exploring available models on cross.io api for layers.', 'duration': 26.719, 'max_score': 4215.497, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44215497.jpg'}, {'end': 4402.599, 'src': 'embed', 'start': 4375.547, 'weight': 4, 'content': [{'end': 4378.869, 'text': "We're going to put in what the loss is, which we'll use the mean squared error.", 'start': 4375.547, 'duration': 3.322}, {'end': 4381.71, 'text': "And we'll go and use the atom optimizer.", 'start': 4378.889, 'duration': 2.821}, {'end': 4384.631, 'text': "Clearly, there's a lot of choices on here, depending on what you're doing.", 'start': 4382.13, 'duration': 2.501}, {'end': 4392.934, 'text': "And just like any of these different prediction models, if you've been doing any scikit from Python,", 'start': 4385.031, 'duration': 7.903}, {'end': 4394.855, 'text': "you'll recognize that we have to then fit the model.", 'start': 4392.934, 'duration': 1.921}, {'end': 4399.357, 'text': "So what are we doing in here? We're going to send in our train X, our train Y.", 'start': 4395.835, 'duration': 3.522}, {'end': 4402.599, 'text': "We're going to decide how many epics we're going to run it through.", 'start': 4399.357, 'duration': 3.242}], 'summary': 'Using mean squared error for loss, atom optimizer, and fitting model with train data for prediction.', 'duration': 27.052, 'max_score': 4375.547, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44375547.jpg'}, {'end': 4488.145, 'src': 'embed', 'start': 4455.666, 'weight': 5, 'content': [{'end': 4463.153, 'text': "And this is going to take a few minutes to run because it's going through 500 times through all the data.", 'start': 4455.666, 'duration': 7.487}, {'end': 4465.716, 'text': 'So, if you have a huge data set?', 'start': 4463.654, 'duration': 2.062}, {'end': 4470.28, 'text': "this is the point where you're kind of wondering oh my gosh, is this going to finish tomorrow?", 'start': 4465.716, 'duration': 4.564}, {'end': 4475.525, 'text': "If I'm running this on a single machine and I have a terabyte of data going into it,", 'start': 4471.521, 'duration': 4.004}, {'end': 4488.145, 'text': "If this is my personal computer and I'm running a terabyte of data into this, this is running rather quickly through all 500 iterations.", 'start': 4480.103, 'duration': 8.042}], 'summary': 'Processing 500 iterations on a personal computer with a terabyte of data running rather quickly.', 'duration': 32.479, 'max_score': 4455.666, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44455666.jpg'}], 'start': 3991.417, 'title': 'Lstm and ml model implementation', 'summary': 'Covers reshaping input data for lstm, creating a sequential lstm model, and using dense model in the neural network training process. it also discusses the availability of machine learning models and the implementation process, including the types of layers and settings, such as activation functions and optimization methods, for model compilation and fitting, emphasizing the potential computational time for large datasets.', 'chapters': [{'end': 4214.917, 'start': 3991.417, 'title': 'Lstm neural network training', 'summary': 'Covers reshaping input data for lstm, creating a sequential lstm model, and using dense model in the neural network training process.', 'duration': 223.5, 'highlights': ['Reshaping the input array in the form of sample time step features by adding one more level to the shape, resulting in new shape 129911 versus the original 12991. The input array is reshaped to incorporate sample time step features by adding one dimension, resulting in a new shape of 129911 versus the original 12991.', 'Creating and fitting the LSTM neural network model with sequential and dense layers, and explaining the difference between sequential and functional models. The process involves creating and fitting the LSTM neural network model with sequential and dense layers, and explaining the difference between sequential and functional models.', 'Explanation of the long short-term memory (LSTM) model and the dense model in the context of TensorFlow, with the dense model handling forward and reverse propagation. The explanation covers the long short-term memory (LSTM) model and the dense model in the context of TensorFlow, with the dense model handling forward and reverse propagation.']}, {'end': 4504.108, 'start': 4215.497, 'title': 'Finding and implementing machine learning models', 'summary': 'Discusses the availability of machine learning models and the implementation process, including the types of layers and settings, such as activation functions and optimization methods, for model compilation and fitting, emphasizing the potential computational time for large datasets.', 'duration': 288.611, 'highlights': ['The availability of machine learning models and their types, including base layers, activation layers, core layers, and layers for specific tasks like image categorization, is discussed. The chapter covers the availability of various machine learning models, including base layers, activation layers, and core layers, as well as layers specific to tasks like image categorization.', 'The process of model compilation, including setting the loss function and optimizer, is detailed, highlighting the importance of choosing the appropriate settings based on the specific project requirements. The process of model compilation is explained, emphasizing the setting of the loss function and optimizer, and the importance of choosing appropriate settings based on project requirements.', 'The potential computational time for training large datasets is discussed, with considerations for the processing time and potential delay when dealing with substantial amounts of data. The potential computational time for training large datasets is addressed, considering the processing time and potential delays when dealing with substantial data volumes.']}], 'duration': 512.691, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D43991417.jpg', 'highlights': ['Reshaping input array to incorporate sample time step features by adding one dimension, resulting in new shape of 129911 versus original 12991.', 'Creating and fitting LSTM neural network model with sequential and dense layers, explaining the difference between sequential and functional models.', 'Explanation of LSTM model and dense model in TensorFlow, with dense model handling forward and reverse propagation.', 'Availability of machine learning models and their types, including base layers, activation layers, core layers, and layers for specific tasks like image categorization.', 'Process of model compilation, including setting the loss function and optimizer, emphasizing the importance of choosing appropriate settings based on project requirements.', 'Discussion on potential computational time for training large datasets, considering processing time and potential delays when dealing with substantial data volumes.']}, {'end': 5194.211, 'segs': [{'end': 4546.759, 'src': 'embed', 'start': 4505.182, 'weight': 0, 'content': [{'end': 4508.666, 'text': 'And then we want to start looking at putting it into some other framework,', 'start': 4505.182, 'duration': 3.484}, {'end': 4514.852, 'text': 'like Spark or something that will build the process on there more across multiple processors and multiple computers.', 'start': 4508.666, 'duration': 6.186}, {'end': 4523.3, 'text': "And if we scroll all the way down to the bottom, you're going to see here's our square mean error, 0.0088.", 'start': 4516.134, 'duration': 7.166}, {'end': 4529.725, 'text': "If we scroll way up, you'll see it kind of oscillates between 0.088 and 0.089.", 'start': 4523.3, 'duration': 6.425}, {'end': 4535.15, 'text': "It's right around 250 where you start seeing that oscillation where it's really not going anywhere.", 'start': 4529.725, 'duration': 5.425}, {'end': 4537.992, 'text': "So we really didn't need to go through a full 500 epics.", 'start': 4535.63, 'duration': 2.362}, {'end': 4544.858, 'text': "If you're retraining the stuff over and over again, it's kind of good to know where that error zone is,", 'start': 4539.453, 'duration': 5.405}, {'end': 4546.759, 'text': "so you don't have to do all the extra processing.", 'start': 4544.858, 'duration': 1.901}], 'summary': 'Consider implementing spark for processing to reduce error oscillation and unnecessary processing.', 'duration': 41.577, 'max_score': 4505.182, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44505182.jpg'}, {'end': 4699.482, 'src': 'embed', 'start': 4656.352, 'weight': 2, 'content': [{'end': 4658.353, 'text': 'But the answer really helps to zoom in.', 'start': 4656.352, 'duration': 2.001}, {'end': 4665.258, 'text': 'So we have a train score which is 2.40 of our root mean square error.', 'start': 4659.254, 'duration': 6.004}, {'end': 4670.022, 'text': 'And we have a test score of 3.16 of the root mean square error.', 'start': 4666.199, 'duration': 3.823}, {'end': 4677.713, 'text': "If these were reversed, if our test score is better than our training score, then we've overtrained.", 'start': 4671.289, 'duration': 6.424}, {'end': 4679.133, 'text': "Something's really wrong.", 'start': 4677.973, 'duration': 1.16}, {'end': 4681.855, 'text': "At that point, you've got to go back and figure out what you did wrong.", 'start': 4679.313, 'duration': 2.542}, {'end': 4687.178, 'text': 'Because you should never have a better result on your test data than you do on your training data.', 'start': 4682.695, 'duration': 4.483}, {'end': 4688.718, 'text': "And that's why we put them both through.", 'start': 4687.518, 'duration': 1.2}, {'end': 4691.54, 'text': "That's why we look at the error for both the training and the testing.", 'start': 4688.738, 'duration': 2.802}, {'end': 4697.321, 'text': "When you're going out and quoting, you're publishing this and you're saying hey, how good is my model?", 'start': 4692.44, 'duration': 4.881}, {'end': 4699.482, 'text': "It's the test score that you're showing people.", 'start': 4697.761, 'duration': 1.721}], 'summary': 'Train score: 2.40, test score: 3.16. test score should not be better than training score to avoid overtraining.', 'duration': 43.13, 'max_score': 4656.352, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44656352.jpg'}, {'end': 5000.977, 'src': 'embed', 'start': 4972.679, 'weight': 4, 'content': [{'end': 4976.762, 'text': "You know, when you show up into a meeting and this is the final output and you say, hey, this is what we're looking at.", 'start': 4972.679, 'duration': 4.083}, {'end': 4979.605, 'text': "Here's our original data in blue.", 'start': 4977.903, 'duration': 1.702}, {'end': 4982.647, 'text': "Here's our training prediction.", 'start': 4980.745, 'duration': 1.902}, {'end': 4986.71, 'text': 'You can see that it trains pretty close to what the data is up there.', 'start': 4983.508, 'duration': 3.202}, {'end': 4997.176, 'text': 'I would also probably put a little timestamp and do just right before and right after where we go from train to test prediction.', 'start': 4987.751, 'duration': 9.425}, {'end': 5000.977, 'text': 'And you can see with the test prediction, the data comes in in red.', 'start': 4997.856, 'duration': 3.121}], 'summary': 'Data analysis presentation: original vs. training vs. test prediction with timestamp.', 'duration': 28.298, 'max_score': 4972.679, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44972679.jpg'}], 'start': 4505.182, 'title': 'Optimizing machine learning process with spark and model prediction and evaluation', 'summary': 'Discusses optimizing the machine learning process with spark to achieve a square mean error of 0.0088, identifying processing inefficiencies, and model prediction and evaluation including a test score of 3.16 and a train score of 2.40.', 'chapters': [{'end': 4546.759, 'start': 4505.182, 'title': 'Optimizing machine learning process with spark', 'summary': 'Discusses optimizing the machine learning process with spark, achieving a square mean error of 0.0088, and identifying the point at around 250 epochs where the oscillation occurs, leading to unnecessary processing.', 'duration': 41.577, 'highlights': ['The square mean error achieved is 0.0088, indicating successful optimization of the machine learning process with Spark.', 'The oscillation in the error occurs around 250 epochs, signifying the point where additional processing becomes unnecessary.']}, {'end': 5194.211, 'start': 4547.239, 'title': 'Model prediction and evaluation', 'summary': 'Discusses running model predictions, comparing training and test errors, calculating root mean square error, and plotting the predictions, with a test score of 3.16 and a train score of 2.40.', 'duration': 646.972, 'highlights': ["The test score of 3.16 and the train score of 2.40 are used to compare the errors and determine the model's performance. The test score of 3.16 and the train score of 2.40 are used to compare the errors and determine the model's performance.", 'The chapter emphasizes the importance of comparing training and testing errors to ensure model validity and avoid overtraining. The chapter emphasizes the importance of comparing training and testing errors to ensure model validity and avoid overtraining.', "The process of plotting the predictions is explained to visually represent the model's performance and provide a clear understanding for stakeholders. The process of plotting the predictions is explained to visually represent the model's performance and provide a clear understanding for stakeholders."]}], 'duration': 689.029, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/QPDsEtUK_D4/pics/QPDsEtUK_D44505182.jpg', 'highlights': ['The square mean error achieved is 0.0088, indicating successful optimization of the machine learning process with Spark.', 'The oscillation in the error occurs around 250 epochs, signifying the point where additional processing becomes unnecessary.', "The test score of 3.16 and the train score of 2.40 are used to compare the errors and determine the model's performance.", 'The chapter emphasizes the importance of comparing training and testing errors to ensure model validity and avoid overtraining.', "The process of plotting the predictions is explained to visually represent the model's performance and provide a clear understanding for stakeholders."]}], 'highlights': ['TensorFlow 2.0 supports eager execution by default, resulting in almost double the code reduction compared to TensorFlow 1.0.', 'The transition from TensorFlow 1.0 to 2.0 has addressed numerous deficiencies and complexities, making the code structure more concise and user-friendly.', 'The significant changes in TensorFlow 2.0, including the shift from TensorFlow 1.0 and the new model abstractions - sequential, functional, and subclassing.', 'The introduction of TF function and autograph feature, allowing for JIT compilation and writing graph code using natural Python syntax.', 'The ability to run TF code on multiple platforms, including CPU, GPU, and TPU, enhancing the versatility and accessibility of TensorFlow models.', 'The chapter covers the usage of math functions in TensorFlow, including the conversion of float to integer, reshaping, transposing tensors, matrix multiplication, bitwise multiplication, and creating an identity matrix.', 'Keras offers three different options for building neural networks: sequential, functional, and subclassing.', 'Keras provides pre-built layers, including dense and LSTM, as well as the capability to create custom layers, making it popular for its comprehensive functionality.', 'Usage of Keras and scikit-learn for data preprocessing.', 'The square mean error achieved is 0.0088, indicating successful optimization of the machine learning process with Spark.', "The test score of 3.16 and the train score of 2.40 are used to compare the errors and determine the model's performance."]}