Hugging Face's logo

bartelds
/
gos-gpu6-cp1_adp0_144m-silver_48-orig_5e-4_cp-11000

Automatic Speech Recognition
Transformers
PyTorch
Gronings
wav2vec2
Inference Endpoints
Model card Files Community
bartelds commited on
Commit
166417c
•
1 Parent(s): b8ebd15

Upload Wav2Vec2ForCTC

Browse files
Files changed (2) hide show
  1. config.json +120 -0
  2. pytorch_model.bin +3 -0
config.json ADDED
@@ -0,0 +1,120 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/nvme/martijn/fine-tuning/gos/cp1_adp0_144m-silver_48-orig/wav2vec2-xls-r-300m-5e-4/checkpoint-11000",
3
+ "activation_dropout": 0.0,
4
+ "adapter_kernel_size": 3,
5
+ "adapter_stride": 2,
6
+ "add_adapter": false,
7
+ "apply_spec_augment": true,
8
+ "architectures": [
9
+ "Wav2Vec2ForCTC"
10
+ ],
11
+ "attention_dropout": 0.1,
12
+ "bos_token_id": 1,
13
+ "bottleneck_adapter_act": "gelu",
14
+ "bottleneck_adapter_dim": 256,
15
+ "classifier_proj_size": 256,
16
+ "cnn_adapter_do_norm": true,
17
+ "cnn_adapter_in_conv_dim": 512,
18
+ "cnn_adapter_kernel": 1,
19
+ "cnn_adapter_out_conv_dim": 512,
20
+ "cnn_adapter_stride": 1,
21
+ "codevector_dim": 768,
22
+ "contrastive_logits_temperature": 0.1,
23
+ "conv_bias": true,
24
+ "conv_dim": [
25
+ 512,
26
+ 512,
27
+ 512,
28
+ 512,
29
+ 512,
30
+ 512,
31
+ 512
32
+ ],
33
+ "conv_kernel": [
34
+ 10,
35
+ 3,
36
+ 3,
37
+ 3,
38
+ 3,
39
+ 2,
40
+ 2
41
+ ],
42
+ "conv_stride": [
43
+ 5,
44
+ 2,
45
+ 2,
46
+ 2,
47
+ 2,
48
+ 2,
49
+ 2
50
+ ],
51
+ "ctc_loss_reduction": "mean",
52
+ "ctc_zero_infinity": true,
53
+ "diversity_loss_weight": 0.1,
54
+ "do_stable_layer_norm": true,
55
+ "eos_token_id": 2,
56
+ "feat_extract_activation": "gelu",
57
+ "feat_extract_dropout": 0.0,
58
+ "feat_extract_norm": "layer",
59
+ "feat_proj_dropout": 0.1,
60
+ "feat_quantizer_dropout": 0.0,
61
+ "final_dropout": 0.0,
62
+ "gradient_checkpointing": false,
63
+ "hidden_act": "gelu",
64
+ "hidden_dropout": 0.1,
65
+ "hidden_size": 1024,
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 4096,
68
+ "layer_norm_eps": 1e-05,
69
+ "layerdrop": 0.1,
70
+ "mask_feature_length": 10,
71
+ "mask_feature_min_masks": 0,
72
+ "mask_feature_prob": 0.004,
73
+ "mask_time_length": 10,
74
+ "mask_time_min_masks": 2,
75
+ "mask_time_prob": 0.075,
76
+ "model_type": "wav2vec2",
77
+ "num_adapter_layers": 3,
78
+ "num_attention_heads": 16,
79
+ "num_codevector_groups": 2,
80
+ "num_codevectors_per_group": 320,
81
+ "num_conv_pos_embedding_groups": 16,
82
+ "num_conv_pos_embeddings": 128,
83
+ "num_feat_extract_layers": 7,
84
+ "num_hidden_layers": 24,
85
+ "num_negatives": 100,
86
+ "output_hidden_size": 1024,
87
+ "pad_token_id": 0,
88
+ "pretrained_model_name_or_path": "/nvme/martijn/pre-training/wav2vec2-xls-r-300m-gos-full-cpt-no-test-1e-5/checkpoint-100000",
89
+ "proj_codevector_dim": 768,
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": "4.24.0",
113
+ "unfreeze_encoder": false,
114
+ "unfreeze_layernorm": false,
115
+ "use_bottleneck_adapter": false,
116
+ "use_cnn_adapter": false,
117
+ "use_weighted_layer_sum": false,
118
+ "vocab_size": 33,
119
+ "xvector_output_dim": 512
120
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca70f9b52a514ee366d041700b2f2a4455d53005b322ac93eaf1b4b1aa2a5555
3
+ size 1262031277