Hugging Face's logo

Manel
/
X-Codec

Transformers
Safetensors
xcodec
Model card Files
xet
Community
Manel commited on
Commit
e07916c
·
verified ·
1 Parent(s): 6b70842

Upload model

Browse files
Files changed (2) hide show
  1. config.json +136 -0
  2. model.safetensors +3 -0
config.json ADDED
@@ -0,0 +1,136 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "acoustic_model_config": {
3
+ "codebook_dim": 8,
4
+ "codebook_loss_weight": 1.0,
5
+ "codebook_size": 1024,
6
+ "commitment_loss_weight": 0.25,
7
+ "decoder_hidden_size": 1024,
8
+ "downsampling_ratios": [
9
+ 8,
10
+ 5,
11
+ 4,
12
+ 2
13
+ ],
14
+ "encoder_hidden_size": 64,
15
+ "hidden_size": 256,
16
+ "hop_length": 320,
17
+ "model_type": "dac",
18
+ "n_codebooks": 9,
19
+ "quantizer_dropout": 0,
20
+ "sampling_rate": 16000,
21
+ "upsampling_ratios": [
22
+ 8,
23
+ 5,
24
+ 4,
25
+ 2
26
+ ]
27
+ },
28
+ "architectures": [
29
+ "XcodecModel"
30
+ ],
31
+ "audio_channels": 1,
32
+ "bandwidth": 4,
33
+ "block_dilations": [
34
+ 1,
35
+ 1
36
+ ],
37
+ "channel_ratios": [
38
+ 1,
39
+ 1
40
+ ],
41
+ "code_dim": 768,
42
+ "codebook_dim": 1024,
43
+ "codebook_size": 1024,
44
+ "decoder_channels": 768,
45
+ "encoder_channels": 768,
46
+ "fc1_in_dim": 1024,
47
+ "fc1_out_dim": 768,
48
+ "fc2_in_dim": 1024,
49
+ "fc2_out_dim": 256,
50
+ "fc_in_dim": 1024,
51
+ "fc_out_dim": 1024,
52
+ "input_channels": 768,
53
+ "kernel_size": 3,
54
+ "model_type": "xcodec",
55
+ "num_quantizers": 8,
56
+ "output_channels": 768,
57
+ "sample_rate": 16000,
58
+ "semantic_model_config": {
59
+ "activation_dropout": 0.1,
60
+ "apply_spec_augment": true,
61
+ "attention_dropout": 0.1,
62
+ "classifier_proj_size": 256,
63
+ "conv_bias": false,
64
+ "conv_dim": [
65
+ 512,
66
+ 512,
67
+ 512,
68
+ 512,
69
+ 512,
70
+ 512,
71
+ 512
72
+ ],
73
+ "conv_kernel": [
74
+ 10,
75
+ 3,
76
+ 3,
77
+ 3,
78
+ 3,
79
+ 2,
80
+ 2
81
+ ],
82
+ "conv_pos_batch_norm": false,
83
+ "conv_stride": [
84
+ 5,
85
+ 2,
86
+ 2,
87
+ 2,
88
+ 2,
89
+ 2,
90
+ 2
91
+ ],
92
+ "ctc_loss_reduction": "sum",
93
+ "ctc_zero_infinity": false,
94
+ "do_stable_layer_norm": false,
95
+ "feat_extract_activation": "gelu",
96
+ "feat_extract_norm": "group",
97
+ "feat_proj_dropout": 0.0,
98
+ "feat_proj_layer_norm": true,
99
+ "final_dropout": 0.1,
100
+ "hidden_act": "gelu",
101
+ "hidden_dropout": 0.1,
102
+ "hidden_size": 768,
103
+ "initializer_range": 0.02,
104
+ "intermediate_size": 3072,
105
+ "layer_norm_eps": 1e-05,
106
+ "layerdrop": 0.1,
107
+ "mask_feature_length": 10,
108
+ "mask_feature_min_masks": 0,
109
+ "mask_feature_prob": 0.0,
110
+ "mask_time_length": 10,
111
+ "mask_time_min_masks": 2,
112
+ "mask_time_prob": 0.05,
113
+ "model_type": "hubert",
114
+ "num_attention_heads": 12,
115
+ "num_conv_pos_embedding_groups": 16,
116
+ "num_conv_pos_embeddings": 128,
117
+ "num_feat_extract_layers": 7,
118
+ "num_hidden_layers": 12,
119
+ "use_weighted_layer_sum": false,
120
+ "vocab_size": 32
121
+ },
122
+ "strides": [
123
+ 1,
124
+ 1
125
+ ],
126
+ "target_bandwidths": [
127
+ 0.5,
128
+ 1,
129
+ 1.5,
130
+ 2,
131
+ 4
132
+ ],
133
+ "torch_dtype": "float32",
134
+ "transformers_version": "4.52.0.dev0",
135
+ "unit_kernel_size": 3
136
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6191971294c216bf6fb7f38fb39e69267a2e3eea6b5469a0952806923b901ce5
3
+ size 709991220