Tensorflow model (.pb) has device information? -
i'm running tf application inference given models. however, it's not running on gpu, on cpu although tensorflow library built cuda enabled. have insight in tf models, tensorflow model (.pb) has device information tf.device(/cpu:0) or tf.device(/gpu:0) ???
from docs (emphasis mine):
sometimes exported meta graph training environment importer doesn't have. example, model might have been trained on gpus, or in distributed environment replicas. when importing such models, it's useful able clear device settings in graph can run on locally available devices. this can achieved calling
import_meta_graph
clear_devices option
settrue
.with tf.session() sess: new_saver = tf.train.import_meta_graph('my-save-dir/my-model-10000.meta', clear_devices=true) new_saver.restore(sess, 'my-save-dir/my-model-10000')
Comments
Post a Comment