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
ASR2 recipe on Tedlium3 dataset #5331
Conversation
Codecov Report
@@ Coverage Diff @@
## master #5331 +/- ##
==========================================
+ Coverage 72.72% 75.36% +2.64%
==========================================
Files 679 709 +30
Lines 61692 65290 +3598
==========================================
+ Hits 44865 49206 +4341
+ Misses 16827 16084 -743
Flags with carried forward coverage won't be shown. Click here to find out more. see 120 files with indirect coverage changes 📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
@@ -0,0 +1,93 @@ | |||
# Trained with A100 (40 GB) x 2 GPUs. It takes 21 minutes per epoch. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Please update the time information later.
@@ -0,0 +1,93 @@ | |||
# Trained with A100 (40 GB) x 1 GPUs for Kmeans1K+nbpe5K. It takes 32 minutes per epoch. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is the time information correct?
@@ -0,0 +1,93 @@ | |||
# Trained with A100 (40 GB) x 1 GPUs. It takes 24 minutes per epoch. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ditto.
What is the status of this PR? We can make it from a draft to regular PR if it is ready. |
Once @kohei0209 gets the asr2 results with new config, we can bring this PR to regular and proceed to merge. |
Hi @kohei0209 , can you continue this PR and upload you checkpoints? |
I am sorry for the late reply. I'll upload the checkpoints and update this PR |
Hi @simpleoier, is it okay to include the data filtering process for removing empty text at stage 6 in this PR? Since Ted3 has some empty texts, data filtering is necessary in # remove empty text
cat "${data_feats}/org/${dset}/text.ts.en" | awk ' { if( NF != 1 ) print $0; } ' > "${data_feats}/${dset}/text.ts.en"
# align keys
# maybe fix_data_dir.sh should be used, it's dirty
utils/filter_scp.pl "${data_feats}/${dset}/text.ts.en" "${data_feats}/org/${dset}/utt2spk" > "${data_feats}/${dset}/utt2spk"
utils/filter_scp.pl "${data_feats}/${dset}/text.ts.en" "${data_feats}/org/${dset}/text.rm.${kmeans_feature_type}_${layer}_km${nclusters}" > "${data_feats}/${dset}/text.rm.${kmeans_feature_type}_${layer}_km${nclusters}"
utils/utt2spk_to_spk2utt.pl "${data_feats}/${dset}/utt2spk" > "${data_feats}/${dset}/spk2utt" |
Did you observe some improvements with it? Yes, stage 6 is the correct place. |
Thank you for your answer. I'll try |
I see. You do not have to do it, but one approach would be to add a special silence token for such utterances. |
I am very sorry for the late reply.
I tried to use
|
I see. |
Thank you for your advice. I've added the code to show how many samples are removed:
The log is as follows:
|
@simpleoier BTW, do you plan to switch the input orders of |
@kohei0209 Thanks for the reminder. You can adjust the order in this PR. |
I reflected your comments. I've also uploaded the model parameters on huggingface. |
Thanks, @kohei0209! |
tedlium3/asr2 recipe
Implementation of tedlium3/asr2 recipe