Sony
/

Mayank-2220 commited on
Commit
553ae4a
1 Parent(s): ae66473

upload_models

Browse files
16_khz/97561_iteration/dec_c.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7581ba4e4282f88d15b972fd8b7cb45e280e7238837a5e4c8b009478a1784cd7
3
+ size 2008794
16_khz/97561_iteration/dec_m_0.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf443c7298f6994284082c0039bf824c70560cd437869bf6f2643b1482fcba71
3
+ size 9543554
16_khz/97561_iteration/enc_c.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6473d71ddf69262a0cbe51aa833d47fad6ba36436e4434392e633069093bb27
3
+ size 170429
16_khz/97561_iteration/hparams.yaml ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ AUDIO_LEN: 12
2
+ A_weighting: false
3
+ HOP_LENGTH: 1024
4
+ NMR: 0.0
5
+ N_FFT: 2048
6
+ SR: 16000
7
+ TEMP_PATH: /group/project25/Mayank/Models/HideAndSpeak/ablate_beat_wavmark_47dB/negative_message_50_percent/temp_files/
8
+ TEST_AUDIO_LEN: 12
9
+ aac_aug: false
10
+ add_carrier_noise: gaussian
11
+ adv: false
12
+ batch_size: 4
13
+ block_type: normal
14
+ carrier_detach: -1
15
+ carrier_noise_norm: 0.01
16
+ cdpam_metric: false
17
+ curriculum: 0
18
+ dataloader_length: 189
19
+ dataset: vctk
20
+ dataset_based_SDR: false
21
+ dec_c_n_layers: 4
22
+ decoder_access_source: false
23
+ enc_n_layers: 3
24
+ encoder_decide_SDR: false
25
+ ensure_constrained_message: false
26
+ ensure_negative_message: true
27
+ frame_level_normalization: false
28
+ lambda_carrier_loss: 20.0
29
+ lambda_msg_loss: 20.0
30
+ load_ckpt: /group/project25/Mayank/Models/HideAndSpeak/ablate_beat_wavmark_47dB/negative_message/ckpt/81560_iteration
31
+ loss_type: mse
32
+ lr: 0.001
33
+ mel_loss: 0
34
+ message_band_size: 512
35
+ message_dim: 4
36
+ message_len: 16
37
+ message_sdr: 47
38
+ microphoneEQ: false
39
+ mix_train: false
40
+ mode: train
41
+ model_type: n_msg
42
+ mp3_aug: false
43
+ msg_loss_weighting_with_energy: false
44
+ n_messages: 1
45
+ no_normalization: false
46
+ num_iters: 5
47
+ num_workers: 8
48
+ ogg_aug: false
49
+ opt: adam
50
+ otat: false
51
+ phase_shift: null
52
+ remove_heuristic_message: false
53
+ rir: /group/project25/Mayank/RIR/rir_small.wav.npz
54
+ rir_train: false
55
+ run_dir: /group/project25/Mayank/Models/HideAndSpeak/ablate_beat_wavmark_47dB/negative_message_50_percent
56
+ sample_every: null
57
+ save_model_every: null
58
+ seed: 0
59
+ stft: new_stft
60
+ test_path: Data/test.txt
61
+ test_results_path: Results/results
62
+ train_path: Data/train.txt
63
+ utterance_level_normalization: true
64
+ val_path: null
16_khz/97561_iteration/opt.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6f7563f8a08e3a4653dc38e8abe8236c4478c51cd51bff5165cfd3c078af9b67
3
+ size 23396974
44_1_khz/73999_iteration/dec_c.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c23b57635172b2fbd3a8531d15ba76b5885a10d0fe902ccb823c521c56041b33
3
+ size 2008794
44_1_khz/73999_iteration/dec_m_0.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:829540c058270d29788f05294894d45bf436e44add0e1099422242ebb94b7088
3
+ size 9554818
44_1_khz/73999_iteration/enc_c.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff64f80d2391fdfc888e4103c499be4a2c958e59626587a1eab6db93204814c7
3
+ size 184765
44_1_khz/73999_iteration/hparams.yaml ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ AUDIO_LEN: 12
2
+ A_weighting: false
3
+ HOP_LENGTH: 2048
4
+ NMR: 0.0
5
+ N_FFT: 4096
6
+ SR: 44100
7
+ TEMP_PATH: /group/project25/Mayank/Models/HideAndSpeak/high_sr_model_40dB/temp_files/
8
+ TEST_AUDIO_LEN: 12
9
+ aac_aug: 64k,128k,256k
10
+ add_carrier_noise: gaussian
11
+ adv: false
12
+ batch_size: 2
13
+ carrier_detach: -1
14
+ carrier_noise_norm: 0.01
15
+ cdpam_metric: false
16
+ curriculum: 0
17
+ dataloader_length: 261
18
+ dataset: vctk
19
+ dataset_based_SDR: false
20
+ dec_c_n_layers: 4
21
+ decoder_access_source: false
22
+ enc_n_layers: 3
23
+ encoder_decide_SDR: false
24
+ ensure_constrained_message: false
25
+ ensure_negative_message: true
26
+ frame_level_normalization: false
27
+ lambda_carrier_loss: 20.0
28
+ lambda_msg_loss: 20.0
29
+ load_ckpt: /group/project25/Mayank/Models/HideAndSpeak/high_sr_model_40dB/ckpt/139058_iteration
30
+ loss_type: mse
31
+ lr: 0.001
32
+ mel_loss: 0
33
+ message_band_size: 1024
34
+ message_dim: 5
35
+ message_len: 21
36
+ message_sdr: 47
37
+ microphoneEQ: false
38
+ mix_train: false
39
+ mode: train
40
+ model_type: n_msg
41
+ mp3_aug: 64k,128k,256k
42
+ msg_loss_weighting_with_energy: false
43
+ n_messages: 1
44
+ no_normalization: false
45
+ num_iters: 5
46
+ num_workers: 8
47
+ ogg_aug: 64k,128k,256k
48
+ opt: adam
49
+ otat: false
50
+ phase_shift: null
51
+ remove_heuristic_message: false
52
+ rir: /group/project25/Mayank/RIR/rir_small.wav.npz
53
+ rir_train: false
54
+ run_dir: /group/project25/Mayank/Models/HideAndSpeak/high_sr_model_40dB
55
+ sample_every: null
56
+ save_model_every: null
57
+ seed: 121
58
+ stft: new_stft
59
+ test_path: Data/test.txt
60
+ test_results_path: Results/results
61
+ train_path: Data/train.txt
62
+ utterance_level_normalization: true
63
+ val_path: null
44_1_khz/73999_iteration/opt.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f129597d7b5459be0455c43d9df12b6a842cbb95344024ebe297fb2516cfd6d1
3
+ size 23448174