Skip to content

Performance issues in the definition of programs.dataset,tensorflow_dl_models/official/mnist/dataset.py #191

@DLPerf

Description

@DLPerf

Hello,I found some performance issues.
The first one is in the definition of dataset ,tensorflow_dl_models/official/mnist/dataset.py,
tf.data.FixedLengthRecordDataset(images_file, 28 * 28, header_bytes=16).map was called without num_parallel_calls.
I think it will increase the efficiency of your program if you add this.

The same issues also exist in tf.data.FixedLengthRecordDataset(
labels_file, 1, header_bytes=8).map
,
.map(decode_csv)),
dataset = dataset.map(decode),
dataset = dataset.map(decode)
and dataset = dataset.map(_parse_line)

Here is the documemtation of tensorflow to support this thing.

The socond one is in thedefinition of Eval,tensorflow_dl_models/research/street/python/vgsl_model.py.
sess = tf.Session('') was repeatedly called and was not closed.
I think it will increase the efficiency and avoid out of memory if you close this session after using it.

The same issues also exist in sess = tf.Session()

Here are two files to support this issue,support1 and support2

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

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions