arrow-left

All pages
gitbookPowered by GitBook
1 of 1

Loading...

Week 3 - Introduction to TensorFlow,Part II

hashtag
This week you will...

  • see how to improve the basic neural network for computer vision you learned last week.

  • learn about what happens if your data is more complex; if the images are larger, or if the features are not always in the same place, and how to handle such issues.

hashtag
Learning Resources

  • (StatQuest Video, 6 min)

  • (StatQuest, 7 min)

  • (StatQuest, 7 min)

hashtag
Until next week you should...

  • Watch the following videos:

    • Theory:

      • (StatQuest with Josh Starmer, 15 min)

Callback Functions in TensorFlowarrow-up-right (DigitalSreeni, 10 min)

C4W1L04 Paddingarrow-up-right (DeepLearningAI, 9:50 min)

  • C4W2L10 Data Augmentationarrow-up-right (DeepLearningAI, 9:30 min)

  • Practice:

    • TensorFlow Tutorial 4 - Convolutional Neural Networks with Sequential and Functional APIarrow-up-right (Aladdin Persson, first 10 min)

      • Note: After the first 10 min, the functional API is covered, which you likely will not need for your projects. Of course, you are free to optionally also watch the rest of the video.

    • (Aladdin Persson, first 7 min)

      • Note: The second method presented in the video (minutes 8 through 14) is deprecated and should no longer be used. The methods presented in the remainder of the video are not relevant to this week's assignment but may be interesting if you're doing a computer-vision project.

    • (Kody Simpson, 23 min)

      • Note: Don't be confused by the alternative approach to do data augmentation that is shown first in the video. The presentation of the method relevant for this week's assignment starts at around timestamp 9:30.

  • Set up a repository for your project following the instructions given herearrow-up-right

  • Conduct a literature review according to the instructions given herearrow-up-right

  • Document your findings on the literature review according to the instructions of above

  • Complete the assignment in the following notebook:

    • Assignment Notebookarrow-up-right

  • Optional:

    • To get a better theoretical understanding of Convolutional Neural Networks, the playlistarrow-up-right from DeepLearningAI from which some of the videos for this week were taken is generally a good source. In particular, you may find the videos (11 min) and C4W1L11 Why Convolutionsarrow-up-right (9:40 min) interesting. Naturally, you are free to watch more videos.

  • file-pdf
    656KB
    250508_Introduction-to-TensorFlow-Part-II.pdf
    PDF
    arrow-up-right-from-squareOpen
    Cross Validationarrow-up-right
    Bias and Variance (Overfitting)arrow-up-right
    Model Evaluation (Confusion Matrix)arrow-up-right
    Neural Networks Part 8: Image Classification with Convolutional Neural Networks (CNNs) arrow-up-right
    TensorFlow Tutorial 18 - Custom Dataset for Imagesarrow-up-right
    Data Augmentation - Deep Learning with Tensorflow | Ep. 19arrow-up-right
    C4W1L06 Convolutions Over Volumesenvelope