Last active
February 14, 2018 13:23
-
-
Save MDBox/02a45550fe9fee761870ec667a41279d to your computer and use it in GitHub Desktop.
Convert caffe realtimePose model to Intel Inference Engine
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#Note you can do all this from SageMaker if you wish. | |
# Convert Caffe Model | |
- First install https://software.intel.com/en-us/inference-engine-devguide | |
- Follow the instructions to install caffe (requires building from source) | |
- Download Original Models from https://github.com/CMU-Perceptual-Computing-Lab/openpose/blob/master/models/getModels.sh | |
For best performance use the MPI_faster version and set the resolution to 184x184. | |
pose_deploy_linevec_faster_4_stages.prototxt | |
input: "image" | |
input_dim: 1 | |
input_dim: 3 | |
input_dim: 184 # Edit this Line to set resolution | |
input_dim: 184 # Edit this Line to set resolution | |
Run this Command to convert the caffe model. | |
LD_LIBRARY_PATH=./ ./ModelOptimizer -o {path to output} -w {path to caffe weight} -d {path to caffe deploy} -i -b 1 -p FP32 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment