SKNahin commited on
Commit
3392874
1 Parent(s): e59c1e9

Upload Wav2Vec2ConformerForCTC (#2)

Browse files

- Upload Wav2Vec2ConformerForCTC (6e992e1ca8143cb60b330ce5a8fb12d2a66facc5)

Files changed (2) hide show
  1. config.json +115 -0
  2. pytorch_model.bin +3 -0
config.json ADDED
@@ -0,0 +1,115 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Bengali_Wav2vec2_Conformer/checkpoint-250000",
3
+ "activation_dropout": 0.1,
4
+ "adapter_kernel_size": 3,
5
+ "adapter_stride": 2,
6
+ "add_adapter": false,
7
+ "apply_spec_augment": true,
8
+ "architectures": [
9
+ "Wav2Vec2ConformerForCTC"
10
+ ],
11
+ "attention_dropout": 0.1,
12
+ "bos_token_id": 1,
13
+ "classifier_proj_size": 256,
14
+ "codevector_dim": 768,
15
+ "conformer_conv_dropout": 0.1,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
18
+ "conv_depthwise_kernel_size": 31,
19
+ "conv_dim": [
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512,
26
+ 512
27
+ ],
28
+ "conv_kernel": [
29
+ 10,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 3,
34
+ 2,
35
+ 2
36
+ ],
37
+ "conv_stride": [
38
+ 5,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2,
44
+ 2
45
+ ],
46
+ "ctc_loss_reduction": "mean",
47
+ "ctc_zero_infinity": true,
48
+ "diversity_loss_weight": 0.1,
49
+ "do_stable_layer_norm": true,
50
+ "eos_token_id": 2,
51
+ "feat_extract_activation": "gelu",
52
+ "feat_extract_dropout": 0.0,
53
+ "feat_extract_norm": "layer",
54
+ "feat_proj_dropout": 0.1,
55
+ "feat_quantizer_dropout": 0.0,
56
+ "final_dropout": 0.1,
57
+ "gradient_checkpointing": false,
58
+ "hidden_act": "swish",
59
+ "hidden_dropout": 0.1,
60
+ "hidden_dropout_prob": 0.1,
61
+ "hidden_size": 1024,
62
+ "initializer_range": 0.02,
63
+ "intermediate_size": 4096,
64
+ "layer_norm_eps": 1e-05,
65
+ "layerdrop": 0.1,
66
+ "mask_feature_length": 10,
67
+ "mask_feature_min_masks": 0,
68
+ "mask_feature_prob": 0.0,
69
+ "mask_time_length": 10,
70
+ "mask_time_min_masks": 2,
71
+ "mask_time_prob": 0.05,
72
+ "max_source_positions": 5000,
73
+ "model_type": "wav2vec2-conformer",
74
+ "num_adapter_layers": 3,
75
+ "num_attention_heads": 16,
76
+ "num_codevector_groups": 2,
77
+ "num_codevectors_per_group": 320,
78
+ "num_conv_pos_embedding_groups": 16,
79
+ "num_conv_pos_embeddings": 128,
80
+ "num_feat_extract_layers": 7,
81
+ "num_hidden_layers": 24,
82
+ "num_negatives": 100,
83
+ "output_hidden_size": 1024,
84
+ "pad_token_id": 109,
85
+ "position_embeddings_type": "relative",
86
+ "proj_codevector_dim": 768,
87
+ "rotary_embedding_base": 10000,
88
+ "tdnn_dilation": [
89
+ 1,
90
+ 2,
91
+ 3,
92
+ 1,
93
+ 1
94
+ ],
95
+ "tdnn_dim": [
96
+ 512,
97
+ 512,
98
+ 512,
99
+ 512,
100
+ 1500
101
+ ],
102
+ "tdnn_kernel": [
103
+ 5,
104
+ 3,
105
+ 3,
106
+ 1,
107
+ 1
108
+ ],
109
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
110
+ "torch_dtype": "float32",
111
+ "transformers_version": "4.34.0.dev0",
112
+ "use_weighted_layer_sum": false,
113
+ "vocab_size": 112,
114
+ "xvector_output_dim": 512
115
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98271af63b0c119bb60db83610a363061c5d5e0e35a354251789de21892216a3
3
+ size 2475225229