Hugging Face's logo

Oysiyl
/
speecht5_tts_common_voice_uk

Text-to-Speech
Transformers
Safetensors
Ukrainian
speecht5
text-to-audio
Generated from Trainer
Inference Endpoints
Model card Files Community
Oysiyl commited on
Commit
8e5ebdb
·
verified ·
1 Parent(s): 85f4e5f

Training in progress, step 146

Browse files
Files changed (4) hide show
  1. README.md +14 -9
  2. config.json +1 -1
  3. model.safetensors +1 -1
  4. training_args.bin +1 -1
README.md CHANGED
@@ -17,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the common_voice_16_1 dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 0.5430
21
 
22
  ## Model description
23
 
@@ -37,24 +37,29 @@ More information needed
37
 
38
  The following hyperparameters were used during training:
39
  - learning_rate: 1e-05
40
- - train_batch_size: 64
41
  - eval_batch_size: 8
42
  - seed: 42
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
- - lr_scheduler_warmup_steps: 500
46
- - num_epochs: 5
47
  - mixed_precision_training: Native AMP
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss |
52
  |:-------------:|:-----:|:----:|:---------------:|
53
- | 0.8074 | 1.0 | 422 | 0.5961 |
54
- | 0.571 | 2.0 | 844 | 0.5699 |
55
- | 0.5418 | 2.99 | 1266 | 0.5517 |
56
- | 0.5241 | 3.99 | 1688 | 0.5487 |
57
- | 0.5105 | 4.99 | 2110 | 0.5430 |
 
 
 
 
 
58
 
59
 
60
  ### Framework versions
 
17
 
18
  This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the common_voice_16_1 dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 0.4211
21
 
22
  ## Model description
23
 
 
37
 
38
  The following hyperparameters were used during training:
39
  - learning_rate: 1e-05
40
+ - train_batch_size: 32
41
  - eval_batch_size: 8
42
  - seed: 42
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
+ - lr_scheduler_warmup_steps: 100
46
+ - num_epochs: 10
47
  - mixed_precision_training: Native AMP
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss |
52
  |:-------------:|:-----:|:----:|:---------------:|
53
+ | 0.8081 | 0.99 | 130 | 0.5394 |
54
+ | 0.5563 | 1.98 | 260 | 0.4533 |
55
+ | 0.5041 | 2.98 | 390 | 0.4468 |
56
+ | 0.4962 | 3.97 | 520 | 0.4628 |
57
+ | 0.495 | 4.96 | 650 | 0.4335 |
58
+ | 0.4777 | 5.95 | 780 | 0.4259 |
59
+ | 0.4752 | 6.95 | 910 | 0.4234 |
60
+ | 0.4673 | 7.94 | 1040 | 0.4229 |
61
+ | 0.4741 | 8.93 | 1170 | 0.4232 |
62
+ | 0.4666 | 9.92 | 1300 | 0.4211 |
63
 
64
 
65
  ### Framework versions
config.json CHANGED
@@ -86,7 +86,7 @@
86
  "speech_decoder_prenet_units": 256,
87
  "torch_dtype": "float32",
88
  "transformers_version": "4.37.2",
89
- "use_cache": false,
90
  "use_guided_attention_loss": true,
91
  "vocab_size": 81
92
  }
 
86
  "speech_decoder_prenet_units": 256,
87
  "torch_dtype": "float32",
88
  "transformers_version": "4.37.2",
89
+ "use_cache": true,
90
  "use_guided_attention_loss": true,
91
  "vocab_size": 81
92
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b0b4edbd24676ad7243a68bb623141090e3a7c7dd229f565844ecfaeb96577b5
3
  size 577789320
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f15792b548370ef392939922942049805149b36822d05642a58a68de3af6f9a8
3
  size 577789320
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1ae5808f1c706a390a93d2838f84685363d3126a61cb8af9f463e11e5bfaa775
3
  size 4399
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a86284ad5f8409a30c10d95d03c6cc21ad42c811bb88f14b26776c223e98656d
3
  size 4399