Compute Library
 19.05
Importing data from existing models

Table of Contents

Extract data from pre-trained caffe model

One can find caffe pre-trained models on caffe's official github repository.

The caffe_data_extractor.py provided in the scripts folder is an example script that shows how to extract the parameter values from a trained model.

Note
complex networks might require altering the script to properly work.

How to use the script

Install caffe following caffe's document. Make sure the pycaffe has been added into the PYTHONPATH.

Download the pre-trained caffe model.

Run the caffe_data_extractor.py script by 

    python caffe_data_extractor.py -m <caffe model> -n <caffe netlist>

For example, to extract the data from pre-trained caffe Alex model to binary file: 

    python caffe_data_extractor.py -m /path/to/bvlc_alexnet.caffemodel -n /path/to/caffe/models/bvlc_alexnet/deploy.prototxt

The script has been tested under Python2.7.

What is the expected output from the script

If the script runs successfully, it prints the names and shapes of each layer onto the standard output and generates *.npy files containing the weights and biases of each layer.

The arm_compute::utils::load_trained_data shows how one could load the weights and biases into tensor from the .npy file by the help of Accessor.

Extract data from pre-trained tensorflow model

The script tensorflow_data_extractor.py extracts trainable parameters (e.g. values of weights and biases) from a trained tensorflow model. A tensorflow model consists of the following two files:

{model_name}.data-{step}-{global_step}: A binary file containing values of each variable.

{model_name}.meta: A binary file containing a MetaGraph struct which defines the graph structure of the neural network.

Note
Since Tensorflow version 0.11 the binary checkpoint file which contains the values for each parameter has the format of: {model_name}.data-{step}-of-{max_step} instead of: {model_name}.ckpt When dealing with binary files with version >= 0.11, only pass {model_name} to -m option; when dealing with binary files with version < 0.11, pass the whole file name {model_name}.ckpt to -m option.
This script relies on the parameters to be extracted being in the 'trainable_variables' tensor collection. By default all variables are automatically added to this collection unless specified otherwise by the user. Thus should a user alter this default behavior and/or want to extract parameters from other collections, tf.GraphKeys.TRAINABLE_VARIABLES should be replaced accordingly.

How to use the script

Install tensorflow and numpy.

Download the pre-trained tensorflow model.

Run tensorflow_data_extractor.py with 

    python tensorflow_data_extractor -m <path_to_binary_checkpoint_file> -n <path_to_metagraph_file>

For example, to extract the data from pre-trained tensorflow Alex model to binary files: 

    python tensorflow_data_extractor -m /path/to/bvlc_alexnet -n /path/to/bvlc_alexnet.meta

Or for binary checkpoint files before Tensorflow 0.11: 

    python tensorflow_data_extractor -m /path/to/bvlc_alexnet.ckpt -n /path/to/bvlc_alexnet.meta
Note
with versions >= Tensorflow 0.11 only model name is passed to the -m option

The script has been tested with Tensorflow 1.2, 1.3 on Python 2.7.6 and Python 3.4.3.

What is the expected output from the script

If the script runs successfully, it prints the names and shapes of each parameter onto the standard output and generates .npy files containing the weights and biases of each layer.

The arm_compute::utils::load_trained_data shows how one could load the weights and biases into tensor from the .npy file by the help of Accessor.

Extract data from pre-trained frozen tensorflow model

The script tf_frozen_model_extractor.py extracts trainable parameters (e.g. values of weights and biases) from a frozen trained Tensorflow model.

How to use the script

Install Tensorflow and NumPy.

Download the pre-trained Tensorflow model and freeze the model using the architecture and the checkpoint file.

Run tf_frozen_model_extractor.py with 

    python tf_frozen_model_extractor -m <path_to_frozen_pb_model_file> -d <path_to_store_parameters>

For example, to extract the data from pre-trained Tensorflow model to binary files: 

    python tf_frozen_model_extractor -m /path/to/inceptionv3.pb -d ./data

What is the expected output from the script

If the script runs successfully, it prints the names and shapes of each parameter onto the standard output and generates .npy files containing the weights and biases of each layer.

The arm_compute::utils::load_trained_data shows how one could load the weights and biases into tensor from the .npy file by the help of Accessor.

Validating examples

Using one of the provided scripts will generate files containing the trainable parameters.

You can validate a given graph example on a list of inputs by running: 

LD_LIBRARY_PATH=lib ./<graph_example> --validation-range='<validation_range>' --validation-file='<validation_file>' --validation-path='/path/to/test/images/' --data='/path/to/weights/'

e.g:

LD_LIBRARY_PATH=lib ./bin/graph_alexnet –target=CL –layout=NHWC –type=F32 –threads=4 –validation-range='16666,24998' –validation-file='val.txt' –validation-path='images/' –data='data/'

where: validation file is a plain document containing a list of images along with their expected label value. e.g: 

val_00000001.JPEG 65
val_00000002.JPEG 970
val_00000003.JPEG 230
val_00000004.JPEG 809
val_00000005.JPEG 516

–validation-range is the index range of the images within the validation file you want to check: e.g:

–validation-range='100,200' will validate 100 images starting from 100th one in the validation file.

This can be useful when parallelizing the validation process is needed.