/learn_notes

learn notes

Primary LanguageJupyter NotebookMIT LicenseMIT

learn_notes

Personal learn notes, interested topic. random update.

  • Python Version: 3.7

CNN

  1. LeNet structure
  2. kaggle [submission](CNN/kaggle submission/notebook78c718f51d.ipynb): mnist
  3. transfer learning - Xception
  4. grad-cam for Conv Layer (matplotlib==3.0.3)

[image](CNN/transfer learning/image_predict_digest.ipynb) and text both able to view; text used image resize to rescale grads to txt len

NLP

  • textcnn: classification for short description of products

conv1d, text-to-vec preprocess comparison.

CV

  • classification for website pictures.
  • InceptionResNetV2 with grad-cam to review model quality

TODO

Python tricks

  1. tkinter
  2. flask [demo](python tricks/coding_check_flask/cnn_predict.py)
  3. Send Email [yagmail based](python tricks/send_email.py)

TODO:

  • pygame