wenfahu

layer {
  name: "train-data"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TRAIN
  }
  transform_param {
    mirror: true

wenfahu

def triplet_loss(anchor, positive, negative, alpha):
    """Calculate the triplet loss according to the FaceNet paper
    
    Args:
      anchor: the embeddings for the anchor images.
      positive: the embeddings for the positive images.
      negative: the embeddings for the negative images.
  
    Returns:
      the triplet loss according to the FaceNet paper as a float tensor.

wenfahu

def triplet_loss_soft(anchor, positive, negative, m=1):
    """Calculate the triplet loss according to the FaceNet paper
    
    Args:
      anchor: the embeddings for the anchor images.
      positive: the embeddings for the positive images.
      negative: the embeddings for the negative images.
  
    Returns:
      the triplet loss according to the FaceNet paper as a float tensor.

wenfahu

from collections import defaultdict
import numpy as np
import pdb
from IPython import embed 


sky = defaultdict(list)

def find_tuple(pts, tgts):
    idx = []

wenfahu

Overall process

This toolset provides channel level pruning of inception-renet v2 model( the details of inception resnet v2 model, please refer to .. _Inception-ResnetV2: https://arxiv.org/abs/1602.07261

1. python inf.py [meta] [ckpt] [output_mask] --threshold [threshold] : get the indices (mask) for the convolutional channel weights under the threshold.
2. python freeze_graph.py [model_dir] [output_file]: freeze the model weights
3. simplify the tensorflow graph using .. GTT: https://github.com/tensorflow/tensorflow/blob/master/tensorflow/tools/graph_transforms/README.md#optimizing-for-deployment
4. python conv_travser.py [graph_path] [mask_path] [output_graph] [output_mask]: prune the model based on the dependency of inception resnet v2, the output_graph is the pruned model and the output_mask is used for further training.

wenfahu

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import tensorflow as tf
import numpy as np
import pickle
from tensorflow.core.framework import attr_value_pb2
from tensorflow.core.framework import graph_pb2
from tensorflow.core.framework import node_def_pb2

wenfahu

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import tensorflow as tf
import numpy as np
import pickle
from tensorflow.core.framework import attr_value_pb2
from tensorflow.core.framework import graph_pb2
from tensorflow.core.framework import node_def_pb2

wenfahu

wenfahu

wenfahu