Hugging Face's logo

DeepSahelAI
/
DeepSahel-Kikuyu-ASR-Optimized

Automatic Speech Recognition
Safetensors
Kikuyu
wav2vec2
kikuyu
gikuyu
asr
deepsahel
Model card Files
xet
 Community
InterstellarCG commited on
Commit
d3fc119
·
verified ·
1 Parent(s): 9655ffb

Publish DeepSahel MMS CTC Kikuyu Muon Ext3

Browse files
Files changed (7) hide show
  1. README.md +29 -0
  2. added_tokens.json +8 -0
  3. config.json +108 -0
  4. model.safetensors +3 -0
  5. processor_config.json +12 -0
  6. tokenizer_config.json +101 -0
  7. vocab.json +0 -0
README.md ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - kik
4
+ license: apache-2.0
5
+ tags:
6
+ - automatic-speech-recognition
7
+ - kikuyu
8
+ - gikuyu
9
+ - asr
10
+ - deepsahel
11
+ pipeline_tag: automatic-speech-recognition
12
+ ---
13
+
14
+ # DeepSahel MMS CTC Kikuyu Muon Ext3
15
+
16
+ A Kikuyu ASR model for speech recognition experiments and evaluation workflows.
17
+
18
+ This model is part of DeepSahel open African-language AI infrastructure work, supporting reusable model-development, evaluation, and documentation workflows for low-resource African languages.
19
+
20
+ ## Intended Use
21
+
22
+ - African-language speech and text research
23
+ - Kikuyu model evaluation and benchmarking
24
+ - Public-interest language technology experiments
25
+ - Downstream prototyping for education, health communication, agriculture, and civic information workflows
26
+
27
+ ## Limitations
28
+
29
+ Performance can vary across domains, dialects, speaker conditions, recording quality, and code-switched usage. Human review by fluent speakers is recommended before production use.
added_tokens.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "3": 39,
3
+ "5": 40,
4
+ "6": 41,
5
+ "7": 42,
6
+ "8": 43,
7
+ "9": 44
8
+ }
config.json ADDED
@@ -0,0 +1,108 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.05,
3
+ "adapter_attn_dim": 16,
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": 0.05,
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,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
44
+ "ctc_loss_reduction": "mean",
45
+ "ctc_zero_infinity": false,
46
+ "diversity_loss_weight": 0.1,
47
+ "do_stable_layer_norm": true,
48
+ "dtype": "float32",
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.05,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.05,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.05,
58
+ "hidden_size": 1280,
59
+ "initializer_range": 0.02,
60
+ "intermediate_size": 5120,
61
+ "layer_norm_eps": 1e-05,
62
+ "layerdrop": 0.05,
63
+ "mask_feature_length": 10,
64
+ "mask_feature_min_masks": 0,
65
+ "mask_feature_prob": 0.004,
66
+ "mask_time_length": 10,
67
+ "mask_time_min_masks": 2,
68
+ "mask_time_prob": 0.05,
69
+ "model_type": "wav2vec2",
70
+ "num_adapter_layers": 3,
71
+ "num_attention_heads": 16,
72
+ "num_codevector_groups": 2,
73
+ "num_codevectors_per_group": 320,
74
+ "num_conv_pos_embedding_groups": 16,
75
+ "num_conv_pos_embeddings": 128,
76
+ "num_feat_extract_layers": 7,
77
+ "num_hidden_layers": 48,
78
+ "num_negatives": 100,
79
+ "output_hidden_size": 1280,
80
+ "pad_token_id": 0,
81
+ "proj_codevector_dim": 1024,
82
+ "tdnn_dilation": [
83
+ 1,
84
+ 2,
85
+ 3,
86
+ 1,
87
+ 1
88
+ ],
89
+ "tdnn_dim": [
90
+ 512,
91
+ 512,
92
+ 512,
93
+ 512,
94
+ 1500
95
+ ],
96
+ "tdnn_kernel": [
97
+ 5,
98
+ 3,
99
+ 3,
100
+ 1,
101
+ 1
102
+ ],
103
+ "transformers_version": "5.3.0",
104
+ "use_cache": false,
105
+ "use_weighted_layer_sum": false,
106
+ "vocab_size": 45,
107
+ "xvector_output_dim": 512
108
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fdffb73c9b8a05d254a97a4b13b0078d0ed0b88687cd7fb799a2e3798272cd68
3
+ size 3858962604
processor_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "feature_extractor": {
3
+ "do_normalize": true,
4
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
5
+ "feature_size": 1,
6
+ "padding_side": "right",
7
+ "padding_value": 0,
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ },
11
+ "processor_class": "Wav2Vec2Processor"
12
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,101 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<pad>",
5
+ "lstrip": true,
6
+ "normalized": false,
7
+ "rstrip": true,
8
+ "single_word": false,
9
+ "special": false
10
+ },
11
+ "1": {
12
+ "content": "<s>",
13
+ "lstrip": true,
14
+ "normalized": false,
15
+ "rstrip": true,
16
+ "single_word": false,
17
+ "special": false
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": true,
22
+ "normalized": false,
23
+ "rstrip": true,
24
+ "single_word": false,
25
+ "special": false
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": true,
30
+ "normalized": false,
31
+ "rstrip": true,
32
+ "single_word": false,
33
+ "special": false
34
+ },
35
+ "39": {
36
+ "content": "3",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": false
42
+ },
43
+ "40": {
44
+ "content": "5",
45
+ "lstrip": false,
46
+ "normalized": false,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": false
50
+ },
51
+ "41": {
52
+ "content": "6",
53
+ "lstrip": false,
54
+ "normalized": false,
55
+ "rstrip": false,
56
+ "single_word": false,
57
+ "special": false
58
+ },
59
+ "42": {
60
+ "content": "7",
61
+ "lstrip": false,
62
+ "normalized": false,
63
+ "rstrip": false,
64
+ "single_word": false,
65
+ "special": false
66
+ },
67
+ "43": {
68
+ "content": "8",
69
+ "lstrip": false,
70
+ "normalized": false,
71
+ "rstrip": false,
72
+ "single_word": false,
73
+ "special": false
74
+ },
75
+ "44": {
76
+ "content": "9",
77
+ "lstrip": false,
78
+ "normalized": false,
79
+ "rstrip": false,
80
+ "single_word": false,
81
+ "special": false
82
+ }
83
+ },
84
+ "backend": "custom",
85
+ "bos_token": "<s>",
86
+ "clean_up_tokenization_spaces": true,
87
+ "do_lower_case": false,
88
+ "eos_token": "</s>",
89
+ "is_local": false,
90
+ "model_max_length": 1000000000000000019884624838656,
91
+ "model_specific_special_tokens": {
92
+ "word_delimiter_token": "|"
93
+ },
94
+ "pad_token": "<pad>",
95
+ "processor_class": "Wav2Vec2Processor",
96
+ "replace_word_delimiter_char": " ",
97
+ "target_lang": "kik",
98
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
99
+ "unk_token": "<unk>",
100
+ "word_delimiter_token": "|"
101
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff