Hugging Face's logo

espnet
/
fastspeech2_conformer

Text-to-Audio
Transformers
PyTorch
English
fastspeech2_conformer
audio
Model card Files
xet
Community
1
connor-henderson commited on
Commit
ecc8dc1
·
1 Parent(s): ea0b415

Upload FastSpeech2ConformerHifiGan

Browse files
Files changed (2) hide show
  1. config.json +42 -61
  2. pytorch_model.bin +2 -2
config.json CHANGED
@@ -1,67 +1,48 @@
1
  {
2
  "architectures": [
3
- "FastSpeech2ConformerModel"
4
  ],
5
- "bos_token_id": 77,
6
- "decoder_attention_dropout_rate": 0.2,
7
- "decoder_concat_after": false,
8
- "decoder_dropout_rate": 0.2,
9
- "decoder_kernel_size": 31,
10
- "decoder_layers": 4,
11
- "decoder_linear_units": 1536,
12
- "decoder_normalize_before": false,
13
- "decoder_num_attention_heads": 2,
14
- "decoder_positional_dropout_rate": 0.2,
15
- "duration_predictor_channels": 256,
16
- "duration_predictor_dropout_rate": 0.2,
17
- "duration_predictor_kernel_size": 3,
18
- "duration_predictor_layers": 2,
19
- "encoder_attention_dropout_rate": 0.2,
20
- "encoder_concat_after": false,
21
- "encoder_dropout_rate": 0.2,
22
- "encoder_kernel_size": 7,
23
- "encoder_layers": 4,
24
- "encoder_linear_units": 1536,
25
- "encoder_normalize_before": false,
26
- "encoder_num_attention_heads": 2,
27
- "encoder_positional_dropout_rate": 0.2,
28
- "energy_embed_dropout": 0.0,
29
- "energy_embed_kernel_size": 1,
30
- "energy_predictor_channels": 256,
31
- "energy_predictor_dropout": 0.5,
32
- "energy_predictor_kernel_size": 3,
33
- "energy_predictor_layers": 2,
34
- "eos_token_id": 77,
35
- "hidden_size": 384,
36
- "input_dim": 78,
37
- "is_encoder_decoder": true,
38
- "max_source_positions": 5000,
39
- "model_type": "fastspeech2_conformer",
40
- "num_languages": null,
41
- "num_mel_bins": 80,
42
- "num_speakers": null,
43
- "pad_token_id": 0,
44
- "pitch_embed_dropout": 0.0,
45
- "pitch_embed_kernel_size": 1,
46
- "pitch_predictor_channels": 256,
47
- "pitch_predictor_dropout": 0.5,
48
- "pitch_predictor_kernel_size": 5,
49
- "pitch_predictor_layers": 5,
50
- "positionwise_conv_kernel_size": 3,
51
- "reduction_factor": 1,
52
- "speaker_embed_dim": null,
53
- "speaking_speed": 1.0,
54
- "speech_decoder_postnet_dropout": 0.5,
55
- "speech_decoder_postnet_kernel": 5,
56
- "speech_decoder_postnet_layers": 5,
57
- "speech_decoder_postnet_units": 256,
58
- "stop_gradient_from_energy_predictor": false,
59
- "stop_gradient_from_pitch_predictor": true,
60
  "torch_dtype": "float32",
61
  "transformers_version": "4.31.0.dev0",
62
- "use_cnn_in_conformer": true,
63
- "use_macaron_style_in_conformer": true,
64
- "use_masking": true,
65
- "use_weighted_masking": false,
66
- "vocab_size": 78
 
 
 
 
 
 
 
 
67
  }
 
1
  {
2
  "architectures": [
3
+ "FastSpeech2ConformerHifiGan"
4
  ],
5
+ "initializer_range": 0.01,
6
+ "leaky_relu_slope": 0.1,
7
+ "model_in_dim": 80,
8
+ "model_type": "hifigan",
9
+ "normalize_before": false,
10
+ "resblock_dilation_sizes": [
11
+ [
12
+ 1,
13
+ 3,
14
+ 5
15
+ ],
16
+ [
17
+ 1,
18
+ 3,
19
+ 5
20
+ ],
21
+ [
22
+ 1,
23
+ 3,
24
+ 5
25
+ ]
26
+ ],
27
+ "resblock_kernel_sizes": [
28
+ 3,
29
+ 7,
30
+ 11
31
+ ],
32
+ "sampling_rate": 22050,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33
  "torch_dtype": "float32",
34
  "transformers_version": "4.31.0.dev0",
35
+ "upsample_initial_channel": 512,
36
+ "upsample_kernel_sizes": [
37
+ 16,
38
+ 16,
39
+ 4,
40
+ 4
41
+ ],
42
+ "upsample_rates": [
43
+ 8,
44
+ 8,
45
+ 2,
46
+ 2
47
+ ]
48
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ba9a9ca759148eee734db773c89ca8502ff84b2f632a8e867759a7d5c2cf2ceb
3
- size 281214857
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b9bc64b2dcd2a463a2bb6bcf89c9482015fe8e75cbabee59e314ae7cad1cb47
3
+ size 55751869