资源算法OpenFacePytorch

OpenFacePytorch

2019-09-17 | |  106 |   0 |   0

OpenFace for Pytorch

I made a dirty code to use OpenFace in PyTorch. I converted 'nn4.small2.v1.t7' to a .hdf5 file using 'torch-hdf5'. Then I read layer informations from .hdf5 file, which can be displayed as follows:

nn.Sequential {
  [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> (7) -> (8) -> (9) -> (10) -> (11) -> (12) -> (13) -> (14) -> (15) -> (16) -> (17) -> (18) -> (19) -> (20) -> (21) -> (22) -> (23) -> (24) -> (25) -> (26) -> output]
  (1): nn.SpatialConvolutionMM(3 -> 64, 7x7, 2,2, 3,3)
  (2): nn.SpatialBatchNormalization (4D) (64)
  (3): nn.ReLU
  (4): nn.SpatialMaxPooling(3x3, 2,2, 1,1)
  (5): nn.SpatialCrossMapLRN
...

Then I manually coded layers in PyTorch (see loadOpenFace.py) with some tentative layers code which may be supported by PyTorch officially laters (SpatialCrossMapLRN_temp.py, adopted from PyTorch's nn.legacy). The final model is 'openface.pth' (which may need to be renamed to 'openface_nn4_small2_v1.pth'), which can be loaded by codes in loadOpenFace.py.

Please see main section of loadOpenFace.py for how-to-use. Simply,

net = prepareOpenFace(useCuda=True, gpuDevice=0, useMultiGPU=False).eval()
feature = net(input_tensor)    # input_tensor should be (batch_size, 3, 96, 96)


上一篇:pytorch-zssr

下一篇:pytorch_compact_bilinear_pooling

用户评价
全部评价

热门资源

  • Keras-ResNeXt

    Keras ResNeXt Implementation of ResNeXt models...

  • seetafaceJNI

    项目介绍 基于中科院seetaface2进行封装的JAVA...

  • spark-corenlp

    This package wraps Stanford CoreNLP annotators ...

  • capsnet-with-caps...

    CapsNet with capsule-wise convolution Project ...

  • inferno-boilerplate

    This is a very basic boilerplate example for pe...