Hugging Face's logo

wadhwani-ai
/
orf_gu_wav2vec2_v3

Automatic Speech Recognition
Transformers
PyTorch
wav2vec2
Inference Endpoints
Model card Files Community
Jatin-WIAI commited on
Commit
75ab23f
•
1 Parent(s): 6873c84

Gujarati Latest model /scratchi/orf/models/asr/gu/pseudo_v9_ft_overlapping_lora/checkpoint-6000, trained on 27/03/24

Browse files
Files changed (2) hide show
  1. config.json +116 -0
  2. pytorch_model.bin +3 -0
config.json ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/scratchi/orf/models/asr/gu/pseudo_v9_ft_overlapping_lora/checkpoint-6000",
3
+ "activation_dropout": 0.1,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_name": "bottleneck",
7
+ "adapter_stride": 2,
8
+ "add_adapter": false,
9
+ "apply_spec_augment": true,
10
+ "architectures": [
11
+ "Wav2Vec2ForCTC"
12
+ ],
13
+ "attention_dropout": 0.1,
14
+ "bos_token_id": 1,
15
+ "classifier_proj_size": 256,
16
+ "codevector_dim": 256,
17
+ "contrastive_logits_temperature": 0.1,
18
+ "conv_bias": true,
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": false,
48
+ "diversity_loss_weight": 0.1,
49
+ "do_stable_layer_norm": true,
50
+ "eos_token_id": 2,
51
+ "feat_enc_adapter": false,
52
+ "feat_extract_activation": "gelu",
53
+ "feat_extract_dropout": 0.0,
54
+ "feat_extract_norm": "layer",
55
+ "feat_proj_dropout": 0.0,
56
+ "feat_quantizer_dropout": 0.0,
57
+ "final_dropout": 0.1,
58
+ "hidden_act": "gelu",
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
+ "lora_adapter": true,
67
+ "lora_r": 8,
68
+ "mask_feature_length": 10,
69
+ "mask_feature_min_masks": 0,
70
+ "mask_feature_prob": 0.0,
71
+ "mask_time_length": 10,
72
+ "mask_time_min_masks": 2,
73
+ "mask_time_prob": 0.05,
74
+ "mh_adapter": false,
75
+ "model_type": "wav2vec2",
76
+ "num_adapter_layers": 3,
77
+ "num_attention_heads": 16,
78
+ "num_codevector_groups": 2,
79
+ "num_codevectors_per_group": 320,
80
+ "num_conv_pos_embedding_groups": 16,
81
+ "num_conv_pos_embeddings": 128,
82
+ "num_feat_extract_layers": 7,
83
+ "num_hidden_layers": 24,
84
+ "num_negatives": 100,
85
+ "output_adapter": false,
86
+ "output_hidden_size": 1024,
87
+ "pad_token_id": 0,
88
+ "prefix_tuning": false,
89
+ "proj_codevector_dim": 256,
90
+ "tdnn_dilation": [
91
+ 1,
92
+ 2,
93
+ 3,
94
+ 1,
95
+ 1
96
+ ],
97
+ "tdnn_dim": [
98
+ 512,
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 1500
103
+ ],
104
+ "tdnn_kernel": [
105
+ 5,
106
+ 3,
107
+ 3,
108
+ 1,
109
+ 1
110
+ ],
111
+ "torch_dtype": "float32",
112
+ "transformers_version": "4.30.0",
113
+ "use_weighted_layer_sum": false,
114
+ "vocab_size": 70,
115
+ "xvector_output_dim": 512
116
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e6658bf272f3731ad535fab4a830178fce88ceba0583ca5917d0886d82d1a1c
3
+ size 1276436205