Hi @attaullahnoor9163, our source codes are completely free of cost. You can download the Jupyter notebook and source code by following the instructions given in the below link. To download the source codes, you’ll need to complete a small quiz-don’t worry, all the answers can be found in the video! Happy Learning. Link - www.elastropy.com/more/unlock-free-source-codes Thanks,

@@elastropy sir please check your email

@@elastropy sir i completed the quiz and send you email but i didnt get any codes yet

Hi @attaullahnoor9163, We generally provide source codes only for videos that have a corresponding quiz available and completed by the user. Source codes for other videos are not sent directly. If a quiz is not yet available, you may request it, and we will upload the quiz as soon as possible. Once you’ve completed the quiz, the source code will be provided. *However, since this is your first trial, we are making an exception and just emailed you the source code. We hope this helps you, and we encourage you to participate in more of our quizzes.* But anyways, Thank you for your email and for sharing your quiz score. Congratulations on achieving a score of 80/100 in Video 58! That’s a great achievement, and it’s wonderful to hear about your progress. Happy Learning.

​@@elastropy Thank you so much for your assistance and for providing the requested source code. I truly appreciate the effort and exception you made to help me. Your support has been invaluable, and I am grateful for the resources you have shared.

Hi @MayankSingh-le1uz, I have already explained these in the following tutorials, please refer to them, Implementation from scratch with DeepXDE- kzbin.info/www/bejne/bXbWYo2NnrKkZrs and kzbin.info/www/bejne/nYnZYoZ9pLFjasU Implementation from scratch without DeepXDE - kzbin.info/www/bejne/sJSQhoeFlteXnq8

Hi @mcdev3.0, You're very welcome! I'm glad you found the tutorial helpful.

@elastropy It's a real pleasure to follow your knowledge-rich videos. I myself have been using PINNs recently in my higher studies in mathematics and Numerical Simulation.

Hi @realracing3specter295! It’s best to create a separate environment to avoid conflicts. First, create an environment (if you’re not familiar with environments please refer to the link given below ). Then, install DeepXDE in that environment. Once installed, run "pip list" to see the installed packages and their versions. Compare this list with the packages required for DeepXDE and manually install only the missing ones to avoid any version conflicts. I assumed most users were already familiar with environments, which is why I didn’t explicitly mention this earlier. Let me know if you need more help! 👍 Anaconda - Working with Python Environments - elastropy.notion.site/Anaconda-Working-with-Python-Environments-b9d0673d1dd84001a566c71fb3518d89

@elastropy thanks, i am aware how to create separate environment, as its easy in anaconda ui... the problem is my limited storage, as anaconda tends to take over more storage as we create multiple environments. My question was more towards the use of multiple versions of packages like numpy, keras, etc... which is possible in the same environment or not? Also, do i need to separately install tensorflow in the new environment again?

Hi @realracing3specter295! Unfortunately, using multiple versions of packages like numpy, keras, etc., in the same environment is not straightforward, as Python environments typically allow only one version of a package at a time. Regarding TensorFlow, yes, you’ll need to install it separately in the new environment. Each environment works independently, so even if TensorFlow is installed globally or in another environment, it won’t automatically carry over.

@elastropy okay. Thanks

Hi @HarmonyNursing, Thank you so much for your kind words! 🙏 We’re so glad you found the tutorial helpful. Your support keeps us motivated to create more valuable content. Happy learning!

Hi @THEKHURRAM100, our source codes are completely free of cost. You can download the Jupyter notebook and source code by following the instructions given in the below link. To download the source codes, you’ll need to complete a small quiz-don’t worry, all the answers can be found in the video! Happy Learning. Link - www.elastropy.com/more/unlock-free-source-codes

Hi @rafay_ahmed794! PINNs are fascinating to work with! The first four videos in this playlist (can be found at the end of the playlist) explain the basics, including how PINNs work, loss function settings, and other key details. After discussing these in detail, I started tutorial series on DeepXDE. Let me know if you have any questions as you explore! 🙌 You can find all of it in this playlist: kzbin.info/aero/PLM7DTyYjZGuLmg3f6j40fEF18jyQmYsC2

@elastropy thank you

@@elastropy Also if I had to add initial conditions in the code, where should I put them?

Hi @AshishSingh-ui1tt, Thank you so much for your kind words! 😊 It means a lot to hear that you found the video helpful. 🙌✨

Hi @p.zh.6132, Thank you for your comment! Solving y'' + λ*y = 0 with BCs: y(0) = y(2π)=0 (a Sturm-Liouville eigenvalue problem) using PINNs is indeed a complex problem, especially due to the presence of multiple eigenvalues. To address this, additional constraints such as orthogonality conditions are typically required to isolate specific eigenvalues or ensure the solution satisfies the physical properties of the problem. These constraints can be incorporated into the loss function alongside the governing equation and boundary conditions.

Hi @799usman, thank you for your suggestion! Generally, in PINNs, we generate our own data at collocation points to solve the equations. However, there are cases like multi-fidelity learning or inverse problems where external datasets are needed, often for applying boundary conditions or for imposing special conditions. Typically, these datasets are integrated into the loss function. Feel free to elaborate on your specific requirements or examples of datasets you're working with. While we're currently focusing on fundamentals as we develop this series, we plan to cover these topics, including integrating external datasets, in future tutorials. Stay tuned!

Hi @799usman, The dataset shown in the video tutorial is primarily used for validating the PINNs solution to the Burgers' equation. If you only have x and t, you can still use the PINNs model with your dataset. However, ensure that the x and t values in your dataset fall within the domains of x and t used during the training of the PINNs model. Outside these boundaries, the model's predictions are likely to be inaccurate or unreliable.

Hi @MunirAlTamami, thank you for your comment! I used normal Python along with a few general Python libraries, as mentioned in the video. Regarding your second question, no, I did not use MATLAB or create an M file for this project. Everything was implemented in Python.

@ thanks

Hi @johndiamantopoulos587, thank you for the kind words! 😊 In PINNs, the number of layers and neurons depends on the complexity of the PDE. More layers capture complex interactions, while more neurons refine the solution. After a point, adding more layers or neurons doesn’t improve the solution or residuals and just increases training time. If the solution or residuals aren’t changing much, it’s a good sign to stick with those numbers.

@elastropy thank you very much for you answer!

Hi @rakesh-do1hw, thank you for the kind words! 😊 I’m glad you enjoyed the video. Our source codes are completely free of cost. You can download the Jupyter notebook and source code by following the instructions given in the below link. To download the source codes, you’ll need to complete a small quiz-don’t worry, all the answers can be found in the video! Happy Learning. Link - www.elastropy.com/more/unlock-free-source-codes

Hi @WilliamBranhamsermons, thank you for the suggestion! 😊 I’ve already covered initial value problems in two videos-one on Newton’s law of cooling and another on the Burgers equation. You can find both in the playlist here: kzbin.info/aero/PLM7DTyYjZGuLmg3f6j40fEF18jyQmYsC2

Hi @WilliamBranhamsermons, thank you! 😊 We’re glad you find these tutorials valuable. It’s true that resources on this topic are rare, which is why we’re committed to sharing them. A detailed list of our upcoming tutorials can be found in the below link-hope you enjoy them too! Link - elastropy.notion.site/Exploring-the-Potential-of-DeepXDE-A-Tentative-Curriculum-for-the-Ultimate-Tutorial-Series-13f8b834612a80aa8b07cf6a24b9796d

Hi @anelengcongo166! 😊 Thank you for your interest. Our source codes are completely free of cost. To download the source code, you’ll need to complete a small quiz-don’t worry, all the answers can be found in the video! Instructions are provided in this link: www.elastropy.com/more/unlock-free-source-codes. For incorporating measured data points in PINNs, we typically add a data loss term to the overall loss function, combining it with the PDE residual and boundary condition residuals. This approach ensures the model adheres to the governing physics while aligning with the true measured values.