Aufgabe | Punkte |
---|---|
7 | 14.5/24 |
8 | 17.5/22 |
9 |
Übung 4 links:
- http://cs231n.github.io/convolutional-networks/
- https://wiseodd.github.io/techblog/2016/07/16/convnet-conv-layer/
- https://towardsdatascience.com/building-convolutional-neural-network-using-numpy-from-scratch-b30aac50e50a
- install all packages
pip -r install requirements.txt
- download data from https://drive.google.com/drive/folders/1MYGsij12XWBKV5UpOSpzXIoJKv6MRNj0
- unzip Images.tar using
tar -xvzf Images.tar
(git bash/mingw/etc.) in classifier root directory
- RUN ON GPU https://stackoverflow.com/questions/45662253/can-i-run-keras-model-on-gpu
- save classification model in keras format - h5 [done]
- create classification report [-]
- create confusion matrix [done]
- save plots:
- accuracy plot [done]
- report as heatmap [-]
- confusion matrix [done]
- https://www.tensorflow.org/install/gpu
pip install tensorflow-gpu
- run
python
sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))