kotlin
good way to inherit class which mandates multiple constructors
https://medium.com/@rygel/stop-repeating-yourself-and-create-custom-views-on-android-with-kotlin-f5b3fc581c0e
deep learning
best path(greedy) vs. beam width search ctc decoding
While training CRNN for text prediction, I found that best path decoding predicts more properly and clearly compared to beam width. Beam width decoding results tended to be excessively messy.
deep learning
Paper Review: “An Empirical Exploration of Recurrent Network Architectures”
paper link Key Points set forget bias to 1 when training LSTM layers to get GRU comparable results in language models, lstm is better than gru
linux
how to install cudnn(from tar file)
from nvidia docs 2.3.1. Installing from a Tar File Navigate to your <cudnnpath> directory containing the cuDNN Tar file. Unzip the cuDNN package. $ tar -xzvf cudnn-9.0-linux-x64-v7.tgz Copy the following files into the CUDA Toolkit directory, and change the file permissions. $ sudo cp cuda/include/cudnn.h /usr/local/cuda/include $ sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64 Read more…
tensorflow
setting gpu fraction in tensorflow keras
put this code at the beginning before building/loading the model.