如何在Tensorflow中可视化Mobilenet层的学习内核过滤器？

Question

我已经培训了MobilenetV1的精简版，现在我想在评估步骤中可视化所有层的学习过滤器。我使用的Mobilenet来自：

https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet_v1.py

我已经尝试过仅与简单CNN相关的tensorflow调试器工具和其他技术。我找不到对当前情况有用的任何信息。

这是我的评估代码：

with tf.Graph().as_default() as graph:
    tf.logging.set_verbosity(tf.logging.INFO)

    dataset = get_split('validation', dataset_dir)
    images, raw_images, labels = load_batch(dataset, batch_size=batch_size, is_training=False)

    num_batches_per_epoch = dataset.num_samples / batch_size
    num_steps_per_epoch = num_batches_per_epoch


    with slim.arg_scope(mobilenet_arg_scope()):
        logits, end_points = mobilenet(images, num_classes=dataset.num_classes, is_training=False)


    variables_to_restore = slim.get_variables_to_restore()
    saver = tf.train.Saver(variables_to_restore)

    def restore_fn(sess):
        return saver.restore(sess, checkpoint_file)


    probabilities = end_points['Predictions']
    predictions = tf.argmax(probabilities, 1)


    accuracy, accuracy_update = tf.contrib.metrics.streaming_accuracy(predictions, labels)


    metrics_op = tf.group(accuracy_update)



    global_step = get_or_create_global_step()
    global_step_op = tf.assign(global_step,
                               global_step + 1) 

    def eval_step(sess, metrics_op, global_step):

        start_time = time.time()
        _, global_step_count, accuracy_value = sess.run([metrics_op, global_step_op, accuracy])
        time_elapsed = time.time() - start_time


        logging.info('Global Step %s: Streaming Accuracy: %.4f (%.2f sec/step)', global_step_count, accuracy_value,
                     time_elapsed)

        return accuracy_value


    tf.summary.scalar('Validation_Accuracy', accuracy)
    matrix = tf.confusion_matrix(labels, predictions, num_classes=3)
    image_tensor = draw_confusion_matrix(matrix)
    image_summary = tf.summary.image('confusion_matrix', image_tensor)

    my_summary_op = tf.summary.merge_all()

    sv = tf.train.Supervisor(logdir=log_eval, summary_op=None, init_fn=restore_fn)


    with sv.managed_session() as sess:
        for step in xrange(int(num_batches_per_epoch * num_epochs)):

            if step % num_batches_per_epoch == 0:
                logging.info('Epoch: %s/%s', step / num_batches_per_epoch + 1, num_epochs)
                logging.info('Current Streaming Accuracy: %.4f', sess.run(accuracy))

            evaluating
            if step % 10 == 0:
                eval_step(sess, metrics_op=metrics_op, global_step=sv.global_step)
                summaries = sess.run(my_summary_op)
                sv.summary_computed(sess, summaries)


            else:
                eval_step(sess, metrics_op=metrics_op, global_step=sv.global_step)


        logging.info('Final eval Accuracy: %.4f', sess.run(accuracy))

我希望看到并保存输出，类似于keras如此轻松地完成输出：

https://blog.keras.io/how-convolutional-neural-networks-see-the-world.html