how to calculate a net's FLOPs in CNN [closed]

Answer 1

For future visitors, if you use Keras and TensorFlow as Backend then you can try the following example. It calculates the FLOPs for the MobileNet.

import tensorflow as tf
import keras.backend as K
from keras.applications.mobilenet import MobileNet

run_meta = tf.RunMetadata()
with tf.Session(graph=tf.Graph()) as sess:
    K.set_session(sess)
    net = MobileNet(alpha=.75, input_tensor=tf.placeholder('float32', shape=(1,32,32,3)))

    opts = tf.profiler.ProfileOptionBuilder.float_operation()    
    flops = tf.profiler.profile(sess.graph, run_meta=run_meta, cmd='op', options=opts)

    opts = tf.profiler.ProfileOptionBuilder.trainable_variables_parameter()    
    params = tf.profiler.profile(sess.graph, run_meta=run_meta, cmd='op', options=opts)

    print("{:,} --- {:,}".format(flops.total_float_ops, params.total_parameters))

Answer 2

Tobias Scheck's answer works if you are using TensorFlow v1.x, but if you are using TensorFlow v2.x you can use the following code:

import tensorflow as tf

def get_flops(model_h5_path):
    session = tf.compat.v1.Session()
    graph = tf.compat.v1.get_default_graph()
        

    with graph.as_default():
        with session.as_default():
            model = tf.keras.models.load_model(model_h5_path)

            run_meta = tf.compat.v1.RunMetadata()
            opts = tf.compat.v1.profiler.ProfileOptionBuilder.float_operation()
        
            # We use the Keras session graph in the call to the profiler.
            flops = tf.compat.v1.profiler.profile(graph=graph,
                                                  run_meta=run_meta, cmd='op', options=opts)
        
            return flops.total_float_ops

The above function takes the path of a saved model in h5 format. You can save your model and use the function this way:

model.save('path_to_my_model.h5')
tf.compat.v1.reset_default_graph()
print(get_flops('path_to_my_model.h5'))

Note that we use tf.compat.v1.reset_default_graph() for not to accumulate FLOPS each time we call the fuction.

Answer 3

For online tool see http://dgschwend.github.io/netscope/#/editor . For alexnet see http://dgschwend.github.io/netscope/#/preset/alexnet . This supports most wide known layers. For custom layers you will have to calculate yourself.

Answer 4

If you're using Keras, you could just use the patch in this pull request: https://github.com/fchollet/keras/pull/6203

Then just call print_summary() and you'll see both the flops per layer and the total.

Even if not using Keras, it may be worth it to recreate your nets in Keras just so you can get the flops counts.