JorisCos commited on
Commit
cadd413
1 Parent(s): cd551f0

Initial commit

Browse files
Files changed (3) hide show
  1. .gitattributes +1 -0
  2. README.md +77 -0
  3. pytorch_model.bin +3 -0
.gitattributes CHANGED
@@ -6,3 +6,4 @@
6
  *.tar.gz filter=lfs diff=lfs merge=lfs -text
7
  *.ot filter=lfs diff=lfs merge=lfs -text
8
  *.onnx filter=lfs diff=lfs merge=lfs -text
 
6
  *.tar.gz filter=lfs diff=lfs merge=lfs -text
7
  *.ot filter=lfs diff=lfs merge=lfs -text
8
  *.onnx filter=lfs diff=lfs merge=lfs -text
9
+ pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - asteroid
4
+ - audio
5
+ - ConvTasNet
6
+ datasets:
7
+ - Libri2Mix
8
+ - sep_noisy
9
+ license: CC BY-NC 4.0
10
+ inference: false
11
+ ---
12
+
13
+ ## Asteroid model `JorisCos/ConvTasNet_Libri2Mix_sepnoisy_8k`
14
+ Imported from [Zenodo](https://zenodo.org/record/3874420#.X9I6NcLjJH4)
15
+
16
+ Description:
17
+
18
+ This model was trained by Joris Cosentino using the librimix recipe in [Asteroid](https://github.com/asteroid-team/asteroid).
19
+ It was trained on the `sep_noisy` task of the Libri2Mix dataset.
20
+
21
+ Training config:
22
+
23
+ ```yml
24
+ data:
25
+ n_src: 2
26
+ sample_rate: 8000
27
+ segment: 3
28
+ task: sep_noisy
29
+ train_dir: data/wav8k/min/train-360
30
+ valid_dir: data/wav8k/min/dev
31
+ filterbank:
32
+ kernel_size: 16
33
+ n_filters: 512
34
+ stride: 8
35
+ masknet:
36
+ bn_chan: 128
37
+ hid_chan: 512
38
+ mask_act: relu
39
+ n_blocks: 8
40
+ n_repeats: 3
41
+ skip_chan: 128
42
+ optim:
43
+ lr: 0.001
44
+ optimizer: adam
45
+ weight_decay: 0.0
46
+ training:
47
+ batch_size: 24
48
+ early_stop: True
49
+ epochs: 200
50
+ half_lr: True
51
+ num_workers: 4
52
+ ```
53
+
54
+
55
+ Results:
56
+
57
+ On Libri2Mix min test set :
58
+ ```yml
59
+ si_sdr: 9.944424856077259
60
+ si_sdr_imp: 11.939395359731192
61
+ sdr: 10.701526190782072
62
+ sdr_imp: 12.481757547845662
63
+ sir: 22.633644975545575
64
+ sir_imp: 22.45666740833025
65
+ sar: 11.131644100944868
66
+ sar_imp: 4.248489589311784
67
+ stoi: 0.852048619949357
68
+ stoi_imp: 0.2071994899565506
69
+ ```
70
+
71
+
72
+ License notice:
73
+
74
+ This work "ConvTasNet_Libri2Mix_sepnoisy_8k" is a derivative of LibriSpeech ASR corpus by Vassil Panayotov,
75
+ used under CC BY 4.0; of The WSJ0 Hipster Ambient Mixtures
76
+ dataset by Whisper.ai, used under CC BY-NC 4.0 (Research only).
77
+ "ConvTasNet_Libri2Mix_sepnoisy_8k" is licensed under Attribution-ShareAlike 3.0 Unported by Joris Cosentino
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:582baf5d5fdf1735ddfccbfa5fdd2e52bb23dfd814497bc1a185ec2c90441bb4
3
+ size 20331984