flozi00 commited on
Commit
056e7a6
1 Parent(s): 02f411b
Files changed (2) hide show
  1. config.json +23 -13
  2. pytorch_model.bin +2 -2
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "wav2vec2-xls-r-1b-5g-german",
3
  "activation_dropout": 0.0,
4
  "adapter_kernel_size": 3,
5
  "adapter_stride": 2,
@@ -11,9 +11,9 @@
11
  "attention_dropout": 0.0,
12
  "bos_token_id": 1,
13
  "classifier_proj_size": 256,
14
- "codevector_dim": 1024,
15
  "contrastive_logits_temperature": 0.1,
16
- "conv_bias": true,
17
  "conv_dim": [
18
  512,
19
  512,
@@ -44,40 +44,50 @@
44
  "ctc_loss_reduction": "mean",
45
  "ctc_zero_infinity": false,
46
  "diversity_loss_weight": 0.1,
47
- "do_stable_layer_norm": true,
48
  "eos_token_id": 2,
49
  "feat_extract_activation": "gelu",
50
- "feat_extract_dropout": 0.0,
51
- "feat_extract_norm": "layer",
52
  "feat_proj_dropout": 0.0,
53
  "feat_quantizer_dropout": 0.0,
54
  "final_dropout": 0.0,
 
55
  "hidden_act": "gelu",
56
  "hidden_dropout": 0.0,
57
- "hidden_size": 1280,
58
  "initializer_range": 0.02,
59
- "intermediate_size": 5120,
60
  "layer_norm_eps": 1e-05,
61
  "layerdrop": 0.0,
 
 
 
 
 
62
  "mask_feature_length": 10,
63
  "mask_feature_min_masks": 0,
64
  "mask_feature_prob": 0.0,
65
  "mask_time_length": 10,
66
  "mask_time_min_masks": 2,
 
 
67
  "mask_time_prob": 0.1,
 
68
  "model_type": "wav2vec2",
 
 
69
  "num_adapter_layers": 3,
70
- "num_attention_heads": 16,
71
  "num_codevector_groups": 2,
72
  "num_codevectors_per_group": 320,
73
  "num_conv_pos_embedding_groups": 16,
74
  "num_conv_pos_embeddings": 128,
75
  "num_feat_extract_layers": 7,
76
- "num_hidden_layers": 48,
77
  "num_negatives": 100,
78
- "output_hidden_size": 1280,
79
  "pad_token_id": 64,
80
- "proj_codevector_dim": 1024,
81
  "tdnn_dilation": [
82
  1,
83
  2,
@@ -100,7 +110,7 @@
100
  1
101
  ],
102
  "torch_dtype": "float32",
103
- "transformers_version": "4.17.0",
104
  "use_weighted_layer_sum": false,
105
  "vocab_size": 67,
106
  "xvector_output_dim": 512
 
1
  {
2
+ "_name_or_path": "wav2vec2-base-german",
3
  "activation_dropout": 0.0,
4
  "adapter_kernel_size": 3,
5
  "adapter_stride": 2,
 
11
  "attention_dropout": 0.0,
12
  "bos_token_id": 1,
13
  "classifier_proj_size": 256,
14
+ "codevector_dim": 256,
15
  "contrastive_logits_temperature": 0.1,
16
+ "conv_bias": false,
17
  "conv_dim": [
18
  512,
19
  512,
 
44
  "ctc_loss_reduction": "mean",
45
  "ctc_zero_infinity": false,
46
  "diversity_loss_weight": 0.1,
47
+ "do_stable_layer_norm": false,
48
  "eos_token_id": 2,
49
  "feat_extract_activation": "gelu",
50
+ "feat_extract_norm": "group",
 
51
  "feat_proj_dropout": 0.0,
52
  "feat_quantizer_dropout": 0.0,
53
  "final_dropout": 0.0,
54
+ "freeze_feat_extract_train": true,
55
  "hidden_act": "gelu",
56
  "hidden_dropout": 0.0,
57
+ "hidden_size": 768,
58
  "initializer_range": 0.02,
59
+ "intermediate_size": 3072,
60
  "layer_norm_eps": 1e-05,
61
  "layerdrop": 0.0,
62
+ "mask_channel_length": 10,
63
+ "mask_channel_min_space": 1,
64
+ "mask_channel_other": 0.0,
65
+ "mask_channel_prob": 0.0,
66
+ "mask_channel_selection": "static",
67
  "mask_feature_length": 10,
68
  "mask_feature_min_masks": 0,
69
  "mask_feature_prob": 0.0,
70
  "mask_time_length": 10,
71
  "mask_time_min_masks": 2,
72
+ "mask_time_min_space": 1,
73
+ "mask_time_other": 0.0,
74
  "mask_time_prob": 0.1,
75
+ "mask_time_selection": "static",
76
  "model_type": "wav2vec2",
77
+ "no_mask_channel_overlap": false,
78
+ "no_mask_time_overlap": false,
79
  "num_adapter_layers": 3,
80
+ "num_attention_heads": 12,
81
  "num_codevector_groups": 2,
82
  "num_codevectors_per_group": 320,
83
  "num_conv_pos_embedding_groups": 16,
84
  "num_conv_pos_embeddings": 128,
85
  "num_feat_extract_layers": 7,
86
+ "num_hidden_layers": 12,
87
  "num_negatives": 100,
88
+ "output_hidden_size": 768,
89
  "pad_token_id": 64,
90
+ "proj_codevector_dim": 256,
91
  "tdnn_dilation": [
92
  1,
93
  2,
 
110
  1
111
  ],
112
  "torch_dtype": "float32",
113
+ "transformers_version": "4.19.0.dev0",
114
  "use_weighted_layer_sum": false,
115
  "vocab_size": 67,
116
  "xvector_output_dim": 512
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bc8a3137238c32f4bb3beec765f8b36e16db7ba75d2f909a888a52d860b573d1
3
- size 3850649329
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf8b03e848773744ee0f27223a98818b03c2e23e3049e9982c7dac3e46437d13
3
+ size 377761559