Wokring Environment
conda create --name tf2 python=3.7
conda activate tf2
conda install pip numpy scipy pandas matplotlib ipython jupyter seaborn pillow h5py
pip install tensorflow-gpu==2.0 gdown
Introduction to TensorFlow
- simple house price prediction
- fashion mnist with simple neural network
- basic image convolution and pooling
- neural network with convoluions
- binary classification: horse or human
- all images are synthesized with computer graphics.
- ImageDataGenerator for managing dataset and feeding them into training pipeline in a systematic way.
- horse or human with validtation
- add the validation dataset in the training process.
- happy or sad
Convolutional Neural Networks in TensorFlow
- Cats or Dogs - Filtered
- Cats or Dogs - Full
- Cats or Dogs - Augmented
- Horse or Human -Augmentation
- Horse or Human - InceptionV3
- Transfer Learning
- Transfer Learning: Human or Horse
- Rock Paper Scissor
- Sign Lanugage Classifier
Natural Language Processing in TensorFlow
- BBC News
- IMDB Review
- BBC excercise
- IMDB Subworkds 8k with Single Layer LSTM
- IMDB Subworkds 8k with Multi Layer LSTM
- IMDB Subworkds 8k with 1D Convolutional Layer
- Sarcasm Bidirectional LSTM
- Sarcasm 1D Convolutional Layer
- IMDB Review GRU
- Exploring Overfitting
- Shapespeare
Sequences, Time Series and Prediction