Skip to content
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

[doc&tools] librispeech recipe #30

Merged
merged 1 commit into from
Dec 28, 2020
Merged

[doc&tools] librispeech recipe #30

merged 1 commit into from
Dec 28, 2020

Conversation

whiteshirt0429
Copy link
Collaborator

No description provided.

test clean
| decoding mode/chunk size | full |
|--------------------------|------|
| attention rescoring | 2,85 |
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

2.85


# decoder related
decoder: transformer
decoder_conf:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

whether the output size should be modified to 512 for a large conformer?


collate_conf:
spec_aug: true
norm_mean: true # if true, must specify cmvn file in cmd line
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

what's the default configure in espnet for cmvn, spec augmentation?

@@ -0,0 +1,61 @@
# network architecture
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

remove this if you don't have any experiment result on this recipe

@@ -0,0 +1,56 @@
# network architecture
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

remove this

@@ -0,0 +1,63 @@
# network architecture
Copy link
Collaborator

@robin1001 robin1001 Dec 24, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

remove this, add it later

datadir=/nfsa/diwu/open-dir
nj=16
dict=data/dict/lang_char.txt
# Optional train_config
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

refine this comment or remove this

# use your own data path
datadir=/nfsa/diwu/open-dir
nj=16
dict=data/dict/lang_char.txt
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

conflicts with the dict declared below

test_dir=$dir/${test}_${mode}
mkdir -p $test_dir
python wenet/bin/recognize.py --gpu 0 \
--mode $mode \
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

indent here

@@ -2,3 +2,5 @@ flake8
pyyaml
tensorboardX
typeguard
Pillow
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

where Pillow is used?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

spec aug time warp, will be added

@@ -0,0 +1,49 @@
#!/usr/bin/env python
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

what's the benefits of wrapping spm_encode/spm_decode/spm_train?
If just for easier input output interface, the raw spm_* are recommended.

@whiteshirt0429 whiteshirt0429 force-pushed the diwu-librispeech branch 2 times, most recently from 3e40d37 to b946596 Compare December 25, 2020 15:07
[tools] specaugmentation config
@robin1001 robin1001 merged commit b4abffb into main Dec 28, 2020
@robin1001 robin1001 deleted the diwu-librispeech branch December 28, 2020 02:03
robin1001 pushed a commit that referenced this pull request Jan 13, 2021
[tools] specaugmentation config
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants