JBZhang2342 commited on
Commit
a92cc05
1 Parent(s): ca7e587

Model save

Browse files
README.md CHANGED
@@ -1,26 +1,21 @@
1
  ---
2
- language:
3
- - en
4
  license: mit
5
  base_model: microsoft/speecht5_tts
6
  tags:
7
- - en_accent,mozilla,t5,common_voice_1_0
8
  - generated_from_trainer
9
- datasets:
10
- - mozilla-foundation/common_voice_1_0
11
  model-index:
12
- - name: SpeechT5 TTS English Accented
13
  results: []
14
  ---
15
 
16
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
  should probably proofread and complete it, then remove this comment. -->
18
 
19
- # SpeechT5 TTS English Accented
20
 
21
- This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the Common Voice dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.6650
24
 
25
  ## Model description
26
 
@@ -46,73 +41,133 @@ The following hyperparameters were used during training:
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: linear
48
  - lr_scheduler_warmup_steps: 500
49
- - training_steps: 15000
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
  | Training Loss | Epoch | Step | Validation Loss |
55
  |:-------------:|:-----:|:-----:|:---------------:|
56
- | No log | 0.53 | 250 | 1.1630 |
57
- | 1.3203 | 1.06 | 500 | 0.8518 |
58
- | 1.3203 | 1.6 | 750 | 0.7785 |
59
- | 0.8972 | 2.13 | 1000 | 0.7502 |
60
- | 0.8972 | 2.66 | 1250 | 0.7369 |
61
- | 0.8135 | 3.19 | 1500 | 0.7193 |
62
- | 0.8135 | 3.72 | 1750 | 0.7152 |
63
- | 0.777 | 4.26 | 2000 | 0.7082 |
64
- | 0.777 | 4.79 | 2250 | 0.7076 |
65
- | 0.7586 | 5.32 | 2500 | 0.6965 |
66
- | 0.7586 | 5.85 | 2750 | 0.6894 |
67
- | 0.747 | 6.38 | 3000 | 0.6790 |
68
- | 0.747 | 6.91 | 3250 | 0.6858 |
69
- | 0.7315 | 7.45 | 3500 | 0.6906 |
70
- | 0.7315 | 7.98 | 3750 | 0.6687 |
71
- | 0.7153 | 8.51 | 4000 | 0.6731 |
72
- | 0.7153 | 9.04 | 4250 | 0.6732 |
73
- | 0.7119 | 9.57 | 4500 | 0.6706 |
74
- | 0.7119 | 10.11 | 4750 | 0.6648 |
75
- | 0.6952 | 10.64 | 5000 | 0.6638 |
76
- | 0.6952 | 11.17 | 5250 | 0.6652 |
77
- | 0.6904 | 11.7 | 5500 | 0.6667 |
78
- | 0.6904 | 12.23 | 5750 | 0.6629 |
79
- | 0.6774 | 12.77 | 6000 | 0.6614 |
80
- | 0.6774 | 13.3 | 6250 | 0.6644 |
81
- | 0.6812 | 13.83 | 6500 | 0.6638 |
82
- | 0.6812 | 14.36 | 6750 | 0.6621 |
83
- | 0.6644 | 14.89 | 7000 | 0.6621 |
84
- | 0.6644 | 15.43 | 7250 | 0.6604 |
85
- | 0.6615 | 15.96 | 7500 | 0.6690 |
86
- | 0.6615 | 16.49 | 7750 | 0.6540 |
87
- | 0.6636 | 17.02 | 8000 | 0.6613 |
88
- | 0.6636 | 17.55 | 8250 | 0.6637 |
89
- | 0.6523 | 18.09 | 8500 | 0.6687 |
90
- | 0.6523 | 18.62 | 8750 | 0.6582 |
91
- | 0.6462 | 19.15 | 9000 | 0.6597 |
92
- | 0.6462 | 19.68 | 9250 | 0.6586 |
93
- | 0.6437 | 20.21 | 9500 | 0.6614 |
94
- | 0.6437 | 20.74 | 9750 | 0.6627 |
95
- | 0.6418 | 21.28 | 10000 | 0.6641 |
96
- | 0.6418 | 21.81 | 10250 | 0.6633 |
97
- | 0.6416 | 22.34 | 10500 | 0.6636 |
98
- | 0.6416 | 22.87 | 10750 | 0.6623 |
99
- | 0.6341 | 23.4 | 11000 | 0.6609 |
100
- | 0.6341 | 23.94 | 11250 | 0.6615 |
101
- | 0.6328 | 24.47 | 11500 | 0.6656 |
102
- | 0.6328 | 25.0 | 11750 | 0.6609 |
103
- | 0.6277 | 25.53 | 12000 | 0.6672 |
104
- | 0.6277 | 26.06 | 12250 | 0.6636 |
105
- | 0.6216 | 26.6 | 12500 | 0.6603 |
106
- | 0.6216 | 27.13 | 12750 | 0.6673 |
107
- | 0.6311 | 27.66 | 13000 | 0.6700 |
108
- | 0.6311 | 28.19 | 13250 | 0.6616 |
109
- | 0.6211 | 28.72 | 13500 | 0.6638 |
110
- | 0.6211 | 29.26 | 13750 | 0.6610 |
111
- | 0.6192 | 29.79 | 14000 | 0.6670 |
112
- | 0.6192 | 30.32 | 14250 | 0.6679 |
113
- | 0.6205 | 30.85 | 14500 | 0.6703 |
114
- | 0.6205 | 31.38 | 14750 | 0.6636 |
115
- | 0.6161 | 31.91 | 15000 | 0.6650 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
116
 
117
 
118
  ### Framework versions
 
1
  ---
 
 
2
  license: mit
3
  base_model: microsoft/speecht5_tts
4
  tags:
 
5
  - generated_from_trainer
 
 
6
  model-index:
7
+ - name: speecht5_tts
8
  results: []
9
  ---
10
 
11
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
12
  should probably proofread and complete it, then remove this comment. -->
13
 
14
+ # speecht5_tts
15
 
16
+ This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the None dataset.
17
  It achieves the following results on the evaluation set:
18
+ - Loss: 0.6815
19
 
20
  ## Model description
21
 
 
41
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
42
  - lr_scheduler_type: linear
43
  - lr_scheduler_warmup_steps: 500
44
+ - training_steps: 30000
45
  - mixed_precision_training: Native AMP
46
 
47
  ### Training results
48
 
49
  | Training Loss | Epoch | Step | Validation Loss |
50
  |:-------------:|:-----:|:-----:|:---------------:|
51
+ | No log | 0.53 | 250 | 1.1437 |
52
+ | 1.3289 | 1.06 | 500 | 0.8521 |
53
+ | 1.3289 | 1.6 | 750 | 0.7901 |
54
+ | 0.8977 | 2.13 | 1000 | 0.7478 |
55
+ | 0.8977 | 2.66 | 1250 | 0.7437 |
56
+ | 0.8131 | 3.19 | 1500 | 0.7243 |
57
+ | 0.8131 | 3.72 | 1750 | 0.7106 |
58
+ | 0.771 | 4.26 | 2000 | 0.7072 |
59
+ | 0.771 | 4.79 | 2250 | 0.7008 |
60
+ | 0.7562 | 5.32 | 2500 | 0.6916 |
61
+ | 0.7562 | 5.85 | 2750 | 0.6850 |
62
+ | 0.7472 | 6.38 | 3000 | 0.6876 |
63
+ | 0.7472 | 6.91 | 3250 | 0.6807 |
64
+ | 0.7266 | 7.45 | 3500 | 0.6804 |
65
+ | 0.7266 | 7.98 | 3750 | 0.6763 |
66
+ | 0.715 | 8.51 | 4000 | 0.6769 |
67
+ | 0.715 | 9.04 | 4250 | 0.6698 |
68
+ | 0.7005 | 9.57 | 4500 | 0.6690 |
69
+ | 0.7005 | 10.11 | 4750 | 0.6653 |
70
+ | 0.6932 | 10.64 | 5000 | 0.6656 |
71
+ | 0.6932 | 11.17 | 5250 | 0.6684 |
72
+ | 0.6854 | 11.7 | 5500 | 0.6645 |
73
+ | 0.6854 | 12.23 | 5750 | 0.6634 |
74
+ | 0.6739 | 12.77 | 6000 | 0.6674 |
75
+ | 0.6739 | 13.3 | 6250 | 0.6606 |
76
+ | 0.6754 | 13.83 | 6500 | 0.6663 |
77
+ | 0.6754 | 14.36 | 6750 | 0.6681 |
78
+ | 0.6592 | 14.89 | 7000 | 0.6589 |
79
+ | 0.6592 | 15.43 | 7250 | 0.6601 |
80
+ | 0.6528 | 15.96 | 7500 | 0.6739 |
81
+ | 0.6528 | 16.49 | 7750 | 0.6643 |
82
+ | 0.6539 | 17.02 | 8000 | 0.6605 |
83
+ | 0.6539 | 17.55 | 8250 | 0.6614 |
84
+ | 0.6437 | 18.09 | 8500 | 0.6551 |
85
+ | 0.6437 | 18.62 | 8750 | 0.6604 |
86
+ | 0.6341 | 19.15 | 9000 | 0.6606 |
87
+ | 0.6341 | 19.68 | 9250 | 0.6582 |
88
+ | 0.6305 | 20.21 | 9500 | 0.6714 |
89
+ | 0.6305 | 20.74 | 9750 | 0.6618 |
90
+ | 0.627 | 21.28 | 10000 | 0.6600 |
91
+ | 0.627 | 21.81 | 10250 | 0.6636 |
92
+ | 0.6244 | 22.34 | 10500 | 0.6692 |
93
+ | 0.6244 | 22.87 | 10750 | 0.6645 |
94
+ | 0.6178 | 23.4 | 11000 | 0.6670 |
95
+ | 0.6178 | 23.94 | 11250 | 0.6611 |
96
+ | 0.6157 | 24.47 | 11500 | 0.6697 |
97
+ | 0.6157 | 25.0 | 11750 | 0.6651 |
98
+ | 0.6108 | 25.53 | 12000 | 0.6642 |
99
+ | 0.6108 | 26.06 | 12250 | 0.6646 |
100
+ | 0.6008 | 26.6 | 12500 | 0.6672 |
101
+ | 0.6008 | 27.13 | 12750 | 0.6601 |
102
+ | 0.6067 | 27.66 | 13000 | 0.6760 |
103
+ | 0.6067 | 28.19 | 13250 | 0.6639 |
104
+ | 0.5985 | 28.72 | 13500 | 0.6662 |
105
+ | 0.5985 | 29.26 | 13750 | 0.6720 |
106
+ | 0.5957 | 29.79 | 14000 | 0.6710 |
107
+ | 0.5957 | 30.32 | 14250 | 0.6688 |
108
+ | 0.5944 | 30.85 | 14500 | 0.6714 |
109
+ | 0.5944 | 31.38 | 14750 | 0.6760 |
110
+ | 0.5886 | 31.91 | 15000 | 0.6639 |
111
+ | 0.5886 | 32.45 | 15250 | 0.6714 |
112
+ | 0.5868 | 32.98 | 15500 | 0.6722 |
113
+ | 0.5868 | 33.51 | 15750 | 0.6790 |
114
+ | 0.5851 | 34.04 | 16000 | 0.6728 |
115
+ | 0.5851 | 34.57 | 16250 | 0.6812 |
116
+ | 0.5819 | 35.11 | 16500 | 0.6756 |
117
+ | 0.5819 | 35.64 | 16750 | 0.6679 |
118
+ | 0.5811 | 36.17 | 17000 | 0.6719 |
119
+ | 0.5811 | 36.7 | 17250 | 0.6684 |
120
+ | 0.5759 | 37.23 | 17500 | 0.6776 |
121
+ | 0.5759 | 37.77 | 17750 | 0.6743 |
122
+ | 0.5743 | 38.3 | 18000 | 0.6725 |
123
+ | 0.5743 | 38.83 | 18250 | 0.6730 |
124
+ | 0.5761 | 39.36 | 18500 | 0.6712 |
125
+ | 0.5761 | 39.89 | 18750 | 0.6765 |
126
+ | 0.576 | 40.43 | 19000 | 0.6779 |
127
+ | 0.576 | 40.96 | 19250 | 0.6801 |
128
+ | 0.5734 | 41.49 | 19500 | 0.6756 |
129
+ | 0.5734 | 42.02 | 19750 | 0.6761 |
130
+ | 0.5743 | 42.55 | 20000 | 0.6857 |
131
+ | 0.5743 | 43.09 | 20250 | 0.6734 |
132
+ | 0.5732 | 43.62 | 20500 | 0.6753 |
133
+ | 0.5732 | 44.15 | 20750 | 0.6803 |
134
+ | 0.5657 | 44.68 | 21000 | 0.6743 |
135
+ | 0.5657 | 45.21 | 21250 | 0.6831 |
136
+ | 0.565 | 45.74 | 21500 | 0.6799 |
137
+ | 0.565 | 46.28 | 21750 | 0.6769 |
138
+ | 0.565 | 46.81 | 22000 | 0.6786 |
139
+ | 0.565 | 47.34 | 22250 | 0.6788 |
140
+ | 0.5583 | 47.87 | 22500 | 0.6830 |
141
+ | 0.5583 | 48.4 | 22750 | 0.6884 |
142
+ | 0.5652 | 48.94 | 23000 | 0.6827 |
143
+ | 0.5652 | 49.47 | 23250 | 0.6795 |
144
+ | 0.5625 | 50.0 | 23500 | 0.6807 |
145
+ | 0.5625 | 50.53 | 23750 | 0.6788 |
146
+ | 0.5605 | 51.06 | 24000 | 0.6862 |
147
+ | 0.5605 | 51.6 | 24250 | 0.6822 |
148
+ | 0.5571 | 52.13 | 24500 | 0.6819 |
149
+ | 0.5571 | 52.66 | 24750 | 0.6797 |
150
+ | 0.5633 | 53.19 | 25000 | 0.6835 |
151
+ | 0.5633 | 53.72 | 25250 | 0.6835 |
152
+ | 0.5572 | 54.26 | 25500 | 0.6881 |
153
+ | 0.5572 | 54.79 | 25750 | 0.6791 |
154
+ | 0.5571 | 55.32 | 26000 | 0.6815 |
155
+ | 0.5571 | 55.85 | 26250 | 0.6868 |
156
+ | 0.5534 | 56.38 | 26500 | 0.6876 |
157
+ | 0.5534 | 56.91 | 26750 | 0.6871 |
158
+ | 0.5525 | 57.45 | 27000 | 0.6836 |
159
+ | 0.5525 | 57.98 | 27250 | 0.6841 |
160
+ | 0.5542 | 58.51 | 27500 | 0.6911 |
161
+ | 0.5542 | 59.04 | 27750 | 0.6835 |
162
+ | 0.5512 | 59.57 | 28000 | 0.6806 |
163
+ | 0.5512 | 60.11 | 28250 | 0.6805 |
164
+ | 0.5474 | 60.64 | 28500 | 0.6858 |
165
+ | 0.5474 | 61.17 | 28750 | 0.6874 |
166
+ | 0.5548 | 61.7 | 29000 | 0.6811 |
167
+ | 0.5548 | 62.23 | 29250 | 0.6808 |
168
+ | 0.5545 | 62.77 | 29500 | 0.6868 |
169
+ | 0.5545 | 63.3 | 29750 | 0.6894 |
170
+ | 0.5522 | 63.83 | 30000 | 0.6815 |
171
 
172
 
173
  ### Framework versions
runs/Dec10_14-09-53_Threadripper/events.out.tfevents.1702235393.Threadripper CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0a108bee4eaa013a170ce5eba817d3020cc84b7935463cdf14b6ac535f5f3625
3
- size 33611
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8f57f5a6f7038b5599f5b8304e86fe4acdf994607bbbca3175d9fb2faf65604
3
+ size 48487