Tensorflow classifier example

tensor flow classifier example.md

1. Instalar docker

curl -sSL https://get.docker.com/ | sh
sudo usermod -aG docker <tu-usuario>

2. Descargar un dataset y organizar las imagenes en formato jpg por carpetas

mkdir tf_files_train

guarda las carpetas en tf_files_train

2. Levantar una imagen de docker para poder desarrollar en tensorflow

docker run -tid --name tf-latest -v $HOME/tf_files_train:/tf_files_train/ gcr.io/tensorflow/tensorflow:latest-devel bash

2.1 Si ya hiciste este paso antes solo levanta el contenedor que se creo

docker start tf-latest

3. Ahora ingresar al contenedor de docker

docker exec -ti tf-latest bash

4. Ejecutar la fase de entrenamiento. Tomara tiempo dependiendo de la cantidad de clases e imagenes por clase

python /tensorflow/tensorflow/examples/image_retraining/retrain.py \
 --bottleneck_dir=/tf_files/bottlenecks \
 --how_many_training_steps 500 \
 --model_dir=/tf_files/inception \
 --output_graph=/tf_files/retrained_graph.pb \
 --output_labels=/tf_files/retrained_labels.txt \
 --image_dir=/tf_files_train

5. Haremos las pruebas, coloca algunas imagenes de prueba en el directorio local: ~/tf_files_train/

6. descarga label_image.py y colocalo en el directorio local: ~/tf_files_train/

7. Ahora dentro de docker ejecuta

mv /tf_files_train/label_image.py /tf_files/
cd /tf_files

8. Haz la prueba con tus imagenes

python label_image.py /tf_files_train/imagen-de-prueba-1.jpg

Referencias

• https://github.com/llSourcell/tensorflow_image_classifier/
• https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/image_retraining/retrain_test.py
• https://www.tensorflow.org/install/install_linux#cpu-only
• https://www.youtube.com/watch?v=QfNvhPx5Px8

convert.sh

## CONVERT A VIDEO ON IMAGES
for i in *.avi;
  do name=`echo $i | cut -d'.' -f1`;
  echo $name;
  mkdir -p $name;
#  ffmpeg -i "$i" "${name}.mov";
  ffmpeg -i "$i" ${name}/${name}_%04d.jpg
done

label_image.py

import tensorflow as tf
import sys

# change this as you see fit
image_path = sys.argv[1]

# Read in the image_data
image_data = tf.gfile.FastGFile(image_path, 'rb').read()

# Loads label file, strips off carriage return
label_lines = [line.rstrip() for line 
                   in tf.gfile.GFile("/tf_files/retrained_labels.txt")]

# Unpersists graph from file
with tf.gfile.FastGFile("/tf_files/retrained_graph.pb", 'rb') as f:
    graph_def = tf.GraphDef()
    graph_def.ParseFromString(f.read())
    _ = tf.import_graph_def(graph_def, name='')

with tf.Session() as sess:
    # Feed the image_data as input to the graph and get first prediction
    softmax_tensor = sess.graph.get_tensor_by_name('final_result:0')
    
    predictions = sess.run(softmax_tensor, \
             {'DecodeJpeg/contents:0': image_data})
    
    # Sort to show labels of first prediction in order of confidence
    top_k = predictions[0].argsort()[-len(predictions[0]):][::-1]
    
    for node_id in top_k:
        human_string = label_lines[node_id]
        score = predictions[0][node_id]
        print('%s (score = %.5f)' % (human_string, score))

Gracias 👍 👍 👍