Hugging Face's logo

hf-tiny-v2
/
tiny-random-HiggsAudioV2TokenizerModel

Safetensors
higgs_audio_v2_tokenizer
Model card Files
xet
 Community
hf-transformers-bot commited on
Commit
7d2133e
·
verified ·
1 Parent(s): 8286d20

Update tiny models for HiggsAudioV2TokenizerModel

Browse files
Files changed (3) hide show
  1. config.json +125 -0
  2. model.safetensors +3 -0
  3. preprocessor_config.json +9 -0
config.json ADDED
@@ -0,0 +1,125 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "acoustic_model_config": {
3
+ "codebook_dim": 8,
4
+ "codebook_loss_weight": 1.0,
5
+ "codebook_size": 16,
6
+ "commitment_loss_weight": 0.25,
7
+ "decoder_hidden_size": 8,
8
+ "downsampling_ratios": [
9
+ 16,
10
+ 16
11
+ ],
12
+ "encoder_hidden_size": 8,
13
+ "hidden_size": 32,
14
+ "hop_length": 256,
15
+ "model_type": "dac",
16
+ "n_codebooks": 9,
17
+ "quantizer_dropout": 0.0,
18
+ "sampling_rate": 16000,
19
+ "upsampling_ratios": [
20
+ 16,
21
+ 16
22
+ ]
23
+ },
24
+ "architectures": [
25
+ "HiggsAudioV2TokenizerModel"
26
+ ],
27
+ "audio_channels": 1,
28
+ "block_dilations": [
29
+ 1,
30
+ 1
31
+ ],
32
+ "channel_ratios": [
33
+ 1,
34
+ 1
35
+ ],
36
+ "codebook_dim": 64,
37
+ "codebook_size": 1024,
38
+ "downsample_factor": 320,
39
+ "dtype": "float32",
40
+ "initializer_range": 0.02,
41
+ "kernel_size": 3,
42
+ "model_type": "higgs_audio_v2_tokenizer",
43
+ "sample_rate": 16000,
44
+ "semantic_model_config": {
45
+ "activation_dropout": 0.1,
46
+ "apply_spec_augment": true,
47
+ "attention_dropout": 0.1,
48
+ "bos_token_id": 1,
49
+ "classifier_proj_size": 256,
50
+ "conv_bias": false,
51
+ "conv_dim": [
52
+ 4,
53
+ 4,
54
+ 4,
55
+ 4,
56
+ 4,
57
+ 4,
58
+ 4
59
+ ],
60
+ "conv_kernel": [
61
+ 10,
62
+ 3,
63
+ 3,
64
+ 3,
65
+ 3,
66
+ 2,
67
+ 2
68
+ ],
69
+ "conv_pos_batch_norm": false,
70
+ "conv_stride": [
71
+ 5,
72
+ 2,
73
+ 2,
74
+ 2,
75
+ 2,
76
+ 2,
77
+ 2
78
+ ],
79
+ "ctc_loss_reduction": "sum",
80
+ "ctc_zero_infinity": false,
81
+ "do_stable_layer_norm": false,
82
+ "eos_token_id": 2,
83
+ "feat_extract_activation": "gelu",
84
+ "feat_extract_norm": "group",
85
+ "feat_proj_dropout": 0.0,
86
+ "feat_proj_layer_norm": true,
87
+ "final_dropout": 0.1,
88
+ "hidden_act": "gelu",
89
+ "hidden_dropout": 0.1,
90
+ "hidden_size": 32,
91
+ "initializer_range": 0.02,
92
+ "intermediate_size": 12,
93
+ "layer_norm_eps": 1e-05,
94
+ "layerdrop": 0.1,
95
+ "mask_feature_length": 10,
96
+ "mask_feature_min_masks": 0,
97
+ "mask_feature_prob": 0.0,
98
+ "mask_time_length": 10,
99
+ "mask_time_min_masks": 2,
100
+ "mask_time_prob": 0.05,
101
+ "model_type": "hubert",
102
+ "num_attention_heads": 2,
103
+ "num_conv_pos_embedding_groups": 16,
104
+ "num_conv_pos_embeddings": 128,
105
+ "num_feat_extract_layers": 7,
106
+ "num_hidden_layers": 2,
107
+ "pad_token_id": 0,
108
+ "use_weighted_layer_sum": false,
109
+ "vocab_size": 32
110
+ },
111
+ "semantic_sample_rate": 16000,
112
+ "strides": [
113
+ 1,
114
+ 1
115
+ ],
116
+ "target_bandwidths": [
117
+ 0.5,
118
+ 1,
119
+ 1.5,
120
+ 2,
121
+ 4
122
+ ],
123
+ "transformers_version": "5.16.0.dev0",
124
+ "unit_kernel_size": 3
125
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2bfdfbdd2db924f692d01f9dcfc8953f32a04ca2b7b1a5b64c7512e33acfcc1b
3
+ size 3864308
preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "feature_extractor_type": "DacFeatureExtractor",
3
+ "feature_size": 1,
4
+ "hop_length": 960,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": true,
8
+ "sampling_rate": 24000
9
+ }