Skip to content

Commit

Permalink
[docs] add multi-gpu usage note to interfaces
Browse files Browse the repository at this point in the history
  • Loading branch information
shelhamer committed Aug 8, 2015
1 parent 2541f0f commit 87a69ea
Showing 1 changed file with 7 additions and 0 deletions.
7 changes: 7 additions & 0 deletions docs/tutorial/interfaces.md
Original file line number Diff line number Diff line change
Expand Up @@ -50,6 +50,13 @@ For a full example of fine-tuning, see examples/finetuning_on_flickr_style, but
# query the first device
caffe device_query -gpu 0

**Parallelism**: the `-gpu` flag to the `caffe` tool can take a comma separated list of IDs to run on multiple GPUs. A solver and net will be instantiated for each GPU so the batch size is effectively multiplied by the number of GPUs. To reproduce single GPU training, reduce the batch size in the network definition accordingly.

# train on GPUs 0 & 1 (doubling the batch size)
caffe train -solver examples/mnist/lenet_solver.prototxt -gpu 0,1
# train on all GPUs (multiplying batch size by number of devices)
caffe train -solver examples/mnist/lenet_solver.prototxt -gpu all

## Python

The Python interface -- pycaffe -- is the `caffe` module and its scripts in caffe/python. `import caffe` to load models, do forward and backward, handle IO, visualize networks, and even instrument model solving. All model data, derivatives, and parameters are exposed for reading and writing.
Expand Down

0 comments on commit 87a69ea

Please sign in to comment.