We do research into deep learning strategies. This means we are more or less unconcerned with performance in the comercial sence. However, when our models start taking weeks to run, then we need to invest time in better engenering. This repo is a journal of our journy through that process.
- 2020/05/05: What is the performance difference in R vs Python?
- 2020/05/09: What is the performance difference when using a CPU vs a GPU vs 2 GPUs
Tensor Flow is super fickle when it comes to setup. The code used in this paper uses the v1 guidence found here