In this session, we will use the Kaggle flowers dataset- this is the same one you had used in the session on transfer learning ( and hence you can use the same notebook on Nimblebox, you can download the following instructions.
You can download the notebook used in this session below. The notebook is also present in the Nimblebox.
Important note: For most of the code in the notebook, you can use a CPU on Nimblebox and switch to the CPU later while training the final model (towards the end of the notebook). You will go through the script ‘resnet.py’ used in the architecture in the new few segments.
NOTE: The code in all of the notebooks provided for this session is updated occasionally. Therefore it may vary slightly from the one shown in the videos.
Let’s get started. First, we’ll look at some of the tools in our kit – python libraries.
Let’s now examine how the data is saved.
Please keep running the code in the notebook along with the videos. Rohit mentioned that the ‘path’ variable points to the main folder which contains two directories named ‘daisy’ and ‘rose’ respectively. Change this data path to your storage directory where flower dataset is present.
Coming Up
Next, we will examine the data a little. We strongly recommend that you run the python code along with the lectures.
Additional Reading
- Here is a paper on detecting cracks on the road using image processing.
Report an error