New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Let us decide the index of the blank in tf.nn.ctc_greedy_decoder #32903
Comments
Hello! Is anyone working on this? If not could I have a try? |
Yes; please feel free.
…On Fri, Apr 3, 2020 at 1:11 PM Panagiotis Varouktsis < ***@***.***> wrote:
Hello! Is anyone working on this? If not could I have a try?
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
<#32903 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AANWFGZDNTMIT4BS2NLDLE3RKY7HHANCNFSM4I3RR32A>
.
|
Thank you! |
Although I really want to do this, I don’t know how to contribute it. I have gone to view the implementation of this function, but I can't find the gen_ctc_ops module. Maybe it implemented by c/c++ code, I am not clear how the python code calls the c/c++ code. There also have a function should add the parameter: tf.nn.ctc_beam_search_decoder. |
I believe it's in core/ops/ctc_ops.cc.
…On Fri, Apr 3, 2020, 8:41 PM Yiming Huang ***@***.***> wrote:
Although I really want to do this, I don’t know how to contribute it. I
have gone to view the implementation of this function, but I can't find the
gen_ctc_ops module.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
<#32903 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AANWFG6TGNMHOFABFJTCSJLRK2T77ANCNFSM4I3RR32A>
.
|
@FLming , |
This issue has been automatically marked as stale because it has no recent activity. It will be closed if no further activity occurs. Thank you. |
OK, I will check it in this week. |
This issue has been automatically marked as stale because it has no recent activity. It will be closed if no further activity occurs. Thank you. |
LGFM. Thanks. |
System information
Describe the feature and the current behavior/state.
I am using ctc_greedy_decoder function to decode my output from logits. In the ctc_loss function, we have the opportunity to choose blank_index equals what, but in the ctc_greedy_decoder function, it seems equals (num_classes-1) by default. Thus I have to write a map function by myself or change my lookup table
Will this change the current api? How?
Please add a parameter to this function in order to choose which index we use.
The text was updated successfully, but these errors were encountered: