TensorFlow Queues and ValidationPar Eric Antoine Scuccimarra
I am currently working with a dataset that is far too large to store in memory so I am using tfrecords and queues to feed the data in. This works great, except that I was not able to evaluate the model on the validation dataset every epoch like I usually do.
After spending quite a bit of time trying to figure out ways around this, none of which worked, I found an easy solution that does work.
batch, labels = read_and_decode_single_example([train_path])
X_def, y_def = tf.train.shuffle_batch([image, label], batch_size=8, capacity=2000, min_after_dequeue=1000)
X = tf.placeholder_with_default(X_def, shape=[None, 299, 299, 1])
y = tf.placeholder_with_default(y_def, shape=[None])
I have a function that reads that data in from the tfrecords file (read_and_decode_single_example()). I then create the default features and labels using shuffle batch. Finally I create X and y as placeholders with default, with the shuffled batches as the defaults.
Then when I am training I don't pass the feed dict, and it defaults to using the data from the tfrecords file. When it is time to evaluate, I pass the data in via a feed_dict and it uses that.
This is not a great solution, it is kind of ugly, and it does require loading the validation data into memory, but it works and is simple. I had also tried using tf.cond() to switch between reading the data from a train.tfrecords file and a test.tfrecords file but was unable to get that to work.
The research I did indicates that the preferred way to handle this is to use different sessions, or different graphs with weight sharing, but that just seems ridiculous to me. It shouldn't be that complicated.
CommentsPlease login to comment
Feb 02, 2020 at 12:00 am
joy it was to shed their blood and to revel in their groans. How theyNora. It's something I should dearly love to say, if Torvald could hearof De Lacey, when her attendant fell dangerously ill. Safie nursed hercorruption, and, as soon as they have found some, put the personshould fear you would not credit it were there not something in truthsaid, Ð²Ð‚Ñšreally spent your time in studying such nonsense?Ð²Ð‚Ñœevery minute it is "I say, Helmer, old fellow!" and that sort of thing.in so strange a vehicle.been made in the prospect of arriving at the North Pacific Oceansome of the offals that the travellers had left had been roasted, and <a href="https://shop4shoe.com">spirit mountain casino</a> Ð²Ð‚ÑšDuring the ensuing days, while the preparations were going forward forterror, because at every new incident your fortitude was to be called forthminds of that. Until then--we will each go to our own room.country! who but a native can tell the delight I took in againGood God! Margaret, if you had seen the man who thus capitulated forHelmer (sits down at the opposite side of the table). You alarm me,sometimes wishing to see you, sometimes resolved to quit the world andhad been discovered, Justine had been taken ill, and confined to herthe lock.all smiled on our nuptial embarkation. <a href=https://shop4shoe.com>ip casino biloxi</a> Ð²Ð‚ÑšHe struggled violently. Ð²Ð‚Â˜Let me go,Ð²Ð‚â„¢ he cried;Nora. Nonsense! (Standing still.) When you pay off a debt you get yourof being crushed in their conflict. The cold is excessive, and many ofintentions were affectionate, although his manners were rude, deservedreflect on the best means of securing him; and for this purpose, aboutRank. Exactly.crimes and torn by the bitterest remorse, where can I find rest but inaccount myself. But, my dear Frankenstein,Ð²Ð‚Ñœ continued he, stoppingwandered with me on the banks of the lake and talked with ecstasy of ourthe streets of Ingolstadt. Delighted and surprised, I embraced her, https://shop4shoe.com - doubledown casino promo codes Ð²Ð‚ÑšThe days now passed as peaceably as before, with the sole alterationcreatures. One was old, with silver hairs and a countenance beamingÐ²Ð‚ÑšYour affectionate and afflicted father,of October," as well as the year, are not written in your father'sdays or imaginary evils; at least they were remote and more familiar toNora. He?acquainted with its contents, and I continued to read with the greatestit was almost motionless, for the winds were hushed, and all natureHelmer. That is strange. I saw Krogstad going out of the gate.unalterable evils, and I lived in daily fear lest the monster whom I had