Hugging Face's logo

bezzam
/
xcodec2-hf

Feature Extraction
Transformers
Safetensors
xcodec2
audio
codec
Model card Files
xet
 Community
bezzam HF Staff commited on
Commit
30f1787
·
verified ·
1 Parent(s): 888e478

Upload model

Browse files
Files changed (2) hide show
  1. config.json +128 -0
  2. model.safetensors +3 -0
config.json ADDED
@@ -0,0 +1,128 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.1,
3
+ "architectures": [
4
+ "Xcodec2Model"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "downsampling_ratios": [
9
+ 2,
10
+ 2,
11
+ 4,
12
+ 4,
13
+ 5
14
+ ],
15
+ "dtype": "float32",
16
+ "encoder_hidden_size": 48,
17
+ "head_dim": 64,
18
+ "hidden_act": "silu",
19
+ "hidden_size": 1024,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 4096,
22
+ "max_position_embeddings": 4096,
23
+ "model_type": "xcodec2",
24
+ "num_attention_heads": 16,
25
+ "num_hidden_layers": 12,
26
+ "num_key_value_heads": 16,
27
+ "pad_token_id": null,
28
+ "quantization_dim": 2048,
29
+ "quantization_levels": [
30
+ 4,
31
+ 4,
32
+ 4,
33
+ 4,
34
+ 4,
35
+ 4,
36
+ 4,
37
+ 4
38
+ ],
39
+ "rms_norm_eps": 1e-06,
40
+ "rope_parameters": {
41
+ "rope_theta": 10000.0,
42
+ "rope_type": "default"
43
+ },
44
+ "sampling_rate": 16000,
45
+ "semantic_model_config": {
46
+ "_name_or_path": "facebook/w2v-bert-2.0",
47
+ "activation_dropout": 0.0,
48
+ "adapter_act": "relu",
49
+ "adapter_kernel_size": 3,
50
+ "adapter_stride": 2,
51
+ "add_adapter": false,
52
+ "apply_spec_augment": false,
53
+ "architectures": [
54
+ "Wav2Vec2BertModel"
55
+ ],
56
+ "attention_dropout": 0.0,
57
+ "bos_token_id": 1,
58
+ "classifier_proj_size": 768,
59
+ "codevector_dim": 768,
60
+ "conformer_conv_dropout": 0.1,
61
+ "contrastive_logits_temperature": 0.1,
62
+ "conv_depthwise_kernel_size": 31,
63
+ "ctc_loss_reduction": "sum",
64
+ "ctc_zero_infinity": false,
65
+ "diversity_loss_weight": 0.1,
66
+ "dtype": "float32",
67
+ "eos_token_id": 2,
68
+ "feat_proj_dropout": 0.0,
69
+ "feat_quantizer_dropout": 0.0,
70
+ "feature_projection_input_dim": 160,
71
+ "final_dropout": 0.1,
72
+ "hidden_act": "swish",
73
+ "hidden_dropout": 0.0,
74
+ "hidden_size": 1024,
75
+ "initializer_range": 0.02,
76
+ "intermediate_size": 4096,
77
+ "layer_norm_eps": 1e-05,
78
+ "layerdrop": 0.1,
79
+ "left_max_position_embeddings": 64,
80
+ "mask_feature_length": 10,
81
+ "mask_feature_min_masks": 0,
82
+ "mask_feature_prob": 0.0,
83
+ "mask_time_length": 10,
84
+ "mask_time_min_masks": 2,
85
+ "mask_time_prob": 0.05,
86
+ "max_source_positions": 5000,
87
+ "model_type": "wav2vec2-bert",
88
+ "num_adapter_layers": 1,
89
+ "num_attention_heads": 16,
90
+ "num_codevector_groups": 2,
91
+ "num_codevectors_per_group": 320,
92
+ "num_hidden_layers": 16,
93
+ "num_negatives": 100,
94
+ "output_hidden_size": 1024,
95
+ "pad_token_id": 0,
96
+ "position_embeddings_type": "relative_key",
97
+ "proj_codevector_dim": 768,
98
+ "right_max_position_embeddings": 8,
99
+ "rotary_embedding_base": 10000,
100
+ "tdnn_dilation": [
101
+ 1,
102
+ 2,
103
+ 3,
104
+ 1,
105
+ 1
106
+ ],
107
+ "tdnn_dim": [
108
+ 512,
109
+ 512,
110
+ 512,
111
+ 512,
112
+ 1500
113
+ ],
114
+ "tdnn_kernel": [
115
+ 5,
116
+ 3,
117
+ 3,
118
+ 1,
119
+ 1
120
+ ],
121
+ "use_intermediate_ffn_before_adapter": false,
122
+ "use_weighted_layer_sum": false,
123
+ "vocab_size": null,
124
+ "xvector_output_dim": 512
125
+ },
126
+ "tie_word_embeddings": false,
127
+ "transformers_version": "5.13.0.dev0"
128
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:611a63e4dff70c19bd4718d701bb7bc522acf6293a109ab62f5db2f7ff395114
3
+ size 2517231448