CS231n
Lecture 8: Deep Learning Software
CPU < GPU
CPU: Fewer cores, but each core is much faster and much more capable; great at sequential tasks
GPU: More cores, but each core is much slower and “dumber”; great for parallel tasks
Deep Learning Frameworks
computational graph的概念適合深度學習
TensorFlow
靜態圖
源於Theano
- 定義tf.placeholder()
tf.get_variable()
- 可以手動計算、更新參數,但是最好使用提供的接口
- 使用tf.Session()
- 使用封裝好的層tf.layers
keras是tensorflow的高層封裝
tensorboard
PyTorch
動態圖
源於Torch
類似tensorboard的Visdom
Caffe
比較
Compare | TensorFlow | PyTorch | Caffe |
---|---|---|---|
Pros | safe, wide use | best | production |
Cons | not perfect | new,can be rough |