Introduction before you test out.

• These are Tensorflow small projects for practice. These projects-codes are for those who want to test some projects without wanting to tinker and for those who want to compare the results.
• These projects are tested on an AMD Ryzen3 5425U laptop, 16gb of ram and Desktop Nvidia 3060 12gb GPU running in Arcolinux (linux-kernal 6.1) OS and CPU governor performance. So, time results may vary from tested resources to resources (CPU, GPU, OS and power state.).
• If laptops are in power save mode, this takes 2x slower than performance mode.
• There are some projects which need to download datasets from Kaggle or Tensorflow Dataset and import them on your own. Lists of datasets that need to download are :

  1. flower dataset
  2. sign minst
  3. BBC dataset
  4. disater dataset
  5. sarasm dataset

• For those who want to make models. This is best practice to use Jupyter notebook at first. Because you can adjust and check for preparing data, input-output shape, and models. Stand-alone python files are for a direct run.
• But my Jupyter notebooks don't contain many useful comments. I just test them out before making python stand-alone files.
• In some folders, there are readme files for showing results, difficulties, opinions, further study and future improvement.
• In some models, I train with functional API.
• There is a small regression dataset. This can be compared with the Time series dataset. The difference between normal regression is making windows and horizons (past and future).