Faster Caffe Training

Faster Caffe Training

Stanford University CS231n Lecture 11 gives useful practical tips for speeding up your training process:

1. Use LMDB (seek times and no image decompression ie. jpeg).
2. Use cuBLAS. The first is the CUDA version of BLAS and will be faster than CPU optimized BLAS.
3. Use cuDNN.
4. Use 32bit floating point precision everywhere (when writing new layers), as they compute faster.