why tensorflow backend is 4x slower than theano when i training a word vector level lstm. #2869

Closed

Closed

why tensorflow backend is 4x slower than theano when i training a word vector level lstm.#2869

opened

Same comments say that tensorflow backend is slower than theano when training lstm.I meet the same question. I want to know why this happen? @fchollet

Contributor

Are you using TF 0.8+? The main reason that TF is sometimes slower than theano is that it is newer. It's currently much faster than after release but there's still room for optimization. This takes time, though.

Author

@talpay Yes, the latest version. Interesting, hope the time will be short.

Also experiencing extreme slowness on TF side.

On Thu, Jun 2, 2016 at 5:21 AM, yanyuxiaoxiao notifications@github.com
wrote:

@talpay https://github.com/talpay Yes, the latest version. Interesting,
hope the time will be short.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#2869 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AFdLCGQRRVi2gtR4qj2kd45ygtWcPGijks5qHj4kgaJpZM4IrjXl
.

added

on May 23, 2017

closed this as completed

on Jun 22, 2017

to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Assignees

No one assigned

Labels

No labels

No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Participants