Google Colab notebooks containing codes I wrote while learning Tensorflow by deeplearning.ai on Coursera.
🔗 Link to the course: [ link ]
📄 Certificate of completion [ link ]
100.0%
All notebooks are saved in this Google Drive link.
Alternatively, you can view the notebooks here on GitHub or in Colab, from the table below.
📋 Description | 🔗 GitHub | 🔗 Colab | |
---|---|---|---|
Week #1 | Intro to Tensorflow | [ link ] | [ link ] |
Exercise 1 (House Prices) | [ link ] | [ link ] | |
Week #2 | Fashion MNIST | [ link ] | [ link ] |
Callbacks | [ link ] | [ link ] | |
Exercise 2 (MNIST handwritten digits) | [ link ] | [ link ] | |
Week #3 | Convolutional Neural Networks (most execises solved) | [ link ] | [ link ] |
Convolutions sidebar Used some numpy to replace some manual multiplication with np.dot |
[ link ] | [ link ] | |
Exercise 3 (convnets on MNIST dataset) | [ link ] | [ link ] | |
Week #4 | Horse-or-Human without validation (needs more work) | [ link ] | [ link ] |
Horse-or-Human with validation | [ link ] | [ link ] | |
Compacted images (error freed) | [ link ] | [ link ] | |
Exercise 4 (Happy or Sad) | [ link ] | [ link ] |
🔗 Link to the course: [ link ]
📄 Certificate of completion [ link ]
100.0%
All notebooks are saved in this Google Drive link.
Alternatively, you can view the notebooks here on GitHub or in Colab, from the table below.
📋 Description | 🔗 GitHub | 🔗 Colab | |
---|---|---|---|
Week #1 | Exploring larger datasets | [ link ] | [ link ] |
Week 1 assignment | [ link ] | [ link ] | |
Horses or Humans (with augmentation) | [ link ] | [ link ] | |
Week #2 | Week 2 Assignment (couldn't get the desired accuracy due to possibly corrupt data, as mentioned in epoch 1) |
[ link ] | [ link ] |
Week #3 | Week 3 Assignment (tried with a different last layer, hence accuracy slightly lower than desired) |
[ link ] | [ link ] |
Week #4 | Week 4 Assignment | [ link ] | [ link ] |
(Will be reformatted once course is complete)
- NLP Practise: [ link-github ] [ link-colab ]
- Word Embeddings: [ link-github ] [ link-colab]