Skip to content

Commit 2557d71

Browse files
mravanellimravanelli
committed
Merge remote-tracking branch 'upstream/develop' into compile_module_keys
2 parents 723e38e + edff43d commit 2557d71

225 files changed

Lines changed: 1899 additions & 1144 deletions

File tree

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

‎recipes/AISHELL-1/ASR/CTC/extra_requirements.txt‎

Lines changed: 0 additions & 2 deletions
This file was deleted.

‎recipes/AISHELL-1/ASR/transformer/extra_requirements.txt‎

Lines changed: 0 additions & 2 deletions
This file was deleted.

‎recipes/AMI/Diarization/README.md‎

Lines changed: 7 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,13 @@
11
# Speaker Diarization on AMI corpus
22
This directory contains the scripts for speaker diarization on the AMI corpus (http://groups.inf.ed.ac.uk/ami/corpus/).
33

4-
## Extra requirements
5-
The code requires scikit-learn as an additional dependency.
6-
To install it, type: `pip install scikit-learn`
4+
## Installing Extra Dependencies
5+
6+
Before proceeding, ensure you have installed the necessary additional dependencies. To do this, simply run the following command in your terminal:
7+
8+
```
9+
pip install -r extra_requirements.txt
10+
```
711

812
## How to run
913
Use the following command to run diarization on AMI corpus.

‎recipes/AMI/Diarization/experiment.py‎

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -191,8 +191,7 @@ def diarize_dataset(full_meta, split_type, n_lambdas, pval, n_neighbors=10):
191191

192192
if len(all_rec_ids) <= 0:
193193
msg = "No recording IDs found! Please check if meta_data json file is properly generated."
194-
logger.error(msg)
195-
sys.exit()
194+
raise ValueError(msg)
196195

197196
# Diarizing different recordings in a dataset.
198197
for rec_id in tqdm(all_rec_ids):

‎recipes/Aishell1Mix/extra-dependencies.txt‎

Lines changed: 0 additions & 10 deletions
This file was deleted.

‎recipes/Aishell1Mix/extra_requirements.txt‎

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,6 @@
1+
matplotlib>=3.1.3
2+
mir-eval==0.6
3+
pyloudnorm
4+
pyloudnorm>=0.1.0
5+
pysndfx>=0.3.6
6+

‎recipes/Aishell1Mix/separation/README.md‎

Lines changed: 13 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -3,6 +3,11 @@ This folder contains some popular recipes for the Aishell1Mix dataset similar to
33

44
* This recipe supports train with several source separation models on Aishell1Mix, including [Sepformer](https://arxiv.org/abs/2010.13154), [DPRNN](https://arxiv.org/abs/1910.06379), [ConvTasnet](https://arxiv.org/abs/1809.07454), [DPTNet](https://arxiv.org/abs/2007.13975).
55

6+
## Installing Extra Dependencies
7+
8+
Before proceeding, ensure you have installed the necessary additional dependencies. To do this, simply run the following command in your terminal:
9+
10+
611
Make sure that SoX is installed on your machine.
712

813
* For windows :
@@ -15,17 +20,23 @@ conda install -c conda-forge sox
1520
```
1621
Additional dependencies:
1722
```
18-
pip install -r ../extra-dependencies.txt
23+
pip install -r ../extra-requirements.txt
1924
```
2025

2126
To run it:
2227

23-
```
28+
```shell
2429
python train.py hparams/sepformer-aishell1mix2.yaml --data_folder /yourdatapath
2530
python train.py hparams/sepformer-aishell1mix3.yaml --data_folder /yourdatapath
2631
```
2732
Note that during training we print the negative SI-SNR (as we treat this value as the loss).
2833

34+
If you want to run it on the test sets only, you can add the flag `--test_only` to the following command:
35+
36+
```shell
37+
python train.py hparams/sepformer-aishell1mix2.yaml --data_folder /yourdatapath --test_only
38+
python train.py hparams/sepformer-aishell1mix3.yaml --data_folder /yourdatapath --test_only
39+
```
2940

3041
# Aishell1Mix2/3
3142
* Your data folder should contain data_aishell (aishell1), resource_aishell (aishell1), wham_noise and aishell1mix, which can be created using the scripts at `https://github.com/huangzj421/Aishell1Mix`. Otherwise train.py will download and prepare data into your data path automatically.

‎recipes/Aishell1Mix/separation/hparams/sepformer-aishell1mix2-wham.yaml‎

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -36,7 +36,6 @@ ckpt_interval_minutes: 60
3636

3737
# Experiment params
3838
auto_mix_prec: True # Set it to True for mixed precision
39-
test_only: False
4039
noprogressbar: False
4140
save_audio: False # Save estimated sources on disk
4241
sample_rate: 8000

‎recipes/Aishell1Mix/separation/hparams/sepformer-aishell1mix2.yaml‎

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -36,7 +36,6 @@ ckpt_interval_minutes: 60
3636

3737
# Experiment params
3838
auto_mix_prec: True # Set it to True for mixed precision
39-
test_only: False
4039
noprogressbar: False
4140
save_audio: False # Save estimated sources on disk
4241
sample_rate: 8000

‎recipes/Aishell1Mix/separation/hparams/sepformer-aishell1mix3-wham.yaml‎

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -36,7 +36,6 @@ ckpt_interval_minutes: 60
3636

3737
# Experiment params
3838
auto_mix_prec: True # Set it to True for mixed precision
39-
test_only: False
4039
noprogressbar: False
4140
save_audio: False # Save estimated sources on disk
4241
sample_rate: 8000

0 commit comments

Comments
 (0)