Hugging Face's logo

UsefulSensors
/
moonshine-streaming-small

Automatic Speech Recognition
Transformers
Safetensors
English
moonshine_streaming
Eval Results
Model card Files
xet
 Community
2
keveman commited on
Commit
c17c188
·
verified ·
1 Parent(s): 86927a0

Upload folder using huggingface_hub

Browse files
Files changed (6) hide show
  1. config.json +88 -0
  2. generation_config.json +8 -0
  3. model.safetensors +3 -0
  4. processor_config.json +12 -0
  5. tokenizer.json +0 -0
  6. tokenizer_config.json +7 -0
config.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "adapter_block_size": 4,
3
+ "adapter_dropout": 0.1,
4
+ "adapter_max_positions": 4096,
5
+ "architectures": [
6
+ "MoonshineStreamingForConditionalGeneration"
7
+ ],
8
+ "attention_bias": false,
9
+ "attn_backend": "auto",
10
+ "attn_dropout": 0.0,
11
+ "bos_token_id": 1,
12
+ "decoder_dim": 512,
13
+ "decoder_num_attention_heads": 8,
14
+ "decoder_num_hidden_layers": 10,
15
+ "decoder_rotary_dim": 32,
16
+ "decoder_start_token_id": 1,
17
+ "dtype": "float32",
18
+ "encoder_dim": 620,
19
+ "encoder_num_attention_heads": 8,
20
+ "encoder_num_hidden_layers": 10,
21
+ "encoder_rotary_dim": 0,
22
+ "encoder_window": [
23
+ [
24
+ 16,
25
+ 4
26
+ ],
27
+ [
28
+ 16,
29
+ 4
30
+ ],
31
+ [
32
+ 16,
33
+ 0
34
+ ],
35
+ [
36
+ 16,
37
+ 0
38
+ ],
39
+ [
40
+ 16,
41
+ 0
42
+ ],
43
+ [
44
+ 16,
45
+ 0
46
+ ],
47
+ [
48
+ 16,
49
+ 0
50
+ ],
51
+ [
52
+ 16,
53
+ 0
54
+ ],
55
+ [
56
+ 16,
57
+ 4
58
+ ],
59
+ [
60
+ 16,
61
+ 4
62
+ ]
63
+ ],
64
+ "eos_token_id": 2,
65
+ "ff_dropout": 0.1,
66
+ "ffn_mult": 4,
67
+ "frame_ms": 5.0,
68
+ "head_dim": 64,
69
+ "is_encoder_decoder": true,
70
+ "model_type": "moonshine_streaming",
71
+ "num_tokens_per_sec": 6.5,
72
+ "pad_token_id": 0,
73
+ "preprocessor_asinh_k_init": 0.75,
74
+ "preprocessor_c1": 1240,
75
+ "preprocessor_c2": 620,
76
+ "preprocessor_input_dropout_p": 0.1,
77
+ "preprocessor_input_dropout_sigma": 0.3,
78
+ "preprocessor_k1": 5,
79
+ "preprocessor_k2": 5,
80
+ "rotary_base": 10000.0,
81
+ "rotary_interpolation_factor": 1.0,
82
+ "sample_rate": 16000,
83
+ "transformers_version": "5.0.0.dev0",
84
+ "use_cache": true,
85
+ "use_swiglu_decoder": true,
86
+ "use_swiglu_encoder": false,
87
+ "vocab_size": 32768
88
+ }
generation_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "decoder_start_token_id": 1,
5
+ "eos_token_id": 2,
6
+ "pad_token_id": 0,
7
+ "transformers_version": "5.0.0.dev0"
8
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46e7a700cac5bde66d08741dc4f881ae61574dcba00bd187cf729bb31be5e480
3
+ size 493471740
processor_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "feature_extractor": {
3
+ "do_normalize": false,
4
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
5
+ "feature_size": 1,
6
+ "padding_side": "right",
7
+ "padding_value": 0.0,
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ },
11
+ "processor_class": "Wav2Vec2Processor"
12
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "is_local": false,
4
+ "model_max_length": 1000000000000000019884624838656,
5
+ "processor_class": "Wav2Vec2Processor",
6
+ "tokenizer_class": "TokenizersBackend"
7
+ }