Performance issues in your project (by P3)

[DLPerf]

Hello! I've found two types of performance issues in your program:

• batch() should be called before map().
• tf.Session being defined repeatedly leads to incremental overhead.

You can make your program more efficient by fixing the above two problems. Here are the tensorflow document and the Stack Overflow post to support this.

Below are detailed issues about batch() should be called before map():

• train_nets.py: dataset = dataset.batch(args.batch_size)(here) should be called before dataset = dataset.map(parse_function)(here).
• train_nets_mgpu_new.py: dataset = dataset.batch(args.batch_size)(here) should be called before dataset = dataset.map(parse_function)(here).
• train_nets_mgpu.py: dataset = dataset.batch(args.batch_size)(here) should be called beforedataset = dataset.map(parse_function)(here).
• data/mx2tfrecords.py: dataset = dataset.batch(32)(here) should be called before dataset = dataset.map(parse_function)(here).

Besides, you need to check the function called in map()(e.g., parse_function called in dataset = dataset.map(parse_function)) whether to be affected or not to make the changed code work properly. For example, ifparse_function needs data with shape (x, y, z) as its input before fix, it would require data with shape (batch_size, x, y, z).

Below are detailed issues about tf.Session being defined repeatedly:

• data/eval_data_reader.py: sess = tf.Session()(here) is repeatedly called in a loop for i in range(len(issame_list)*2):(here).

If you define tf.Session out of the loop and pass tf.Session as a parameter to the loop, your program would be much more efficient.

Looking forward to your reply. Btw, I am very glad to create a PR to fix it if you are too busy.

[DLPerf]

Hello, I'm looking forward to your reply~