Hugging Face's logo

keras
/
moonshine_tiny_en

KerasHub
Model card Files
xet
Community
Divyasreepat commited on
Commit
a7c2def
·
verified ·
1 Parent(s): 27fc170

Upload folder using huggingface_hub

Browse files
Files changed (10) hide show
  1. .gitattributes +1 -0
  2. README.md +31 -0
  3. assets/tokenizer/vocabulary.spm +3 -0
  4. audio_converter.json +20 -0
  5. config.json +29 -0
  6. metadata.json +9 -0
  7. model.weights.h5 +3 -0
  8. preprocessor.json +60 -0
  9. task.json +97 -0
  10. tokenizer.json +22 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ assets/tokenizer/vocabulary.spm filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: keras-hub
3
+ ---
4
+ This is a [`Moonshine` model](https://keras.io/api/keras_hub/models/moonshine) uploaded using the KerasHub library and can be used with JAX, TensorFlow, and PyTorch backends.
5
+ This model is related to a `AudioToText` task.
6
+
7
+ Model config:
8
+ * **name:** moonshine_backbone_1
9
+ * **trainable:** True
10
+ * **vocabulary_size:** 32768
11
+ * **filter_dim:** 288
12
+ * **encoder_num_layers:** 6
13
+ * **decoder_num_layers:** 6
14
+ * **hidden_dim:** 288
15
+ * **intermediate_dim:** 1152
16
+ * **encoder_num_heads:** 8
17
+ * **decoder_num_heads:** 8
18
+ * **feedforward_expansion_factor:** 4
19
+ * **encoder_use_swiglu_activation:** False
20
+ * **decoder_use_swiglu_activation:** True
21
+ * **max_position_embeddings:** 194
22
+ * **pad_head_dim_to_multiple_of:** None
23
+ * **partial_rotary_factor:** 0.9
24
+ * **dropout:** 0.0
25
+ * **initializer_range:** 0.02
26
+ * **rope_theta:** 10000.0
27
+ * **attention_bias:** False
28
+ * **attention_dropout:** 0.0
29
+ * **dtype:** float32
30
+
31
+ This model card has been generated automatically and should be completed by the model author. See [Model Cards documentation](https://huggingface.co/docs/hub/model-cards) for more information.
assets/tokenizer/vocabulary.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
audio_converter.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.moonshine.moonshine_audio_converter",
3
+ "class_name": "MoonshineAudioConverter",
4
+ "config": {
5
+ "name": "moonshine_audio_converter_1",
6
+ "trainable": true,
7
+ "dtype": {
8
+ "module": "keras",
9
+ "class_name": "DTypePolicy",
10
+ "config": {
11
+ "name": "float32"
12
+ },
13
+ "registered_name": null
14
+ },
15
+ "sampling_rate": 16000,
16
+ "padding_value": 0.0,
17
+ "do_normalize": false
18
+ },
19
+ "registered_name": "keras_hub>MoonshineAudioConverter"
20
+ }
config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.moonshine.moonshine_backbone",
3
+ "class_name": "MoonshineBackbone",
4
+ "config": {
5
+ "name": "moonshine_backbone_1",
6
+ "trainable": true,
7
+ "vocabulary_size": 32768,
8
+ "filter_dim": 288,
9
+ "encoder_num_layers": 6,
10
+ "decoder_num_layers": 6,
11
+ "hidden_dim": 288,
12
+ "intermediate_dim": 1152,
13
+ "encoder_num_heads": 8,
14
+ "decoder_num_heads": 8,
15
+ "feedforward_expansion_factor": 4,
16
+ "encoder_use_swiglu_activation": false,
17
+ "decoder_use_swiglu_activation": true,
18
+ "max_position_embeddings": 194,
19
+ "pad_head_dim_to_multiple_of": null,
20
+ "partial_rotary_factor": 0.9,
21
+ "dropout": 0.0,
22
+ "initializer_range": 0.02,
23
+ "rope_theta": 10000.0,
24
+ "attention_bias": false,
25
+ "attention_dropout": 0.0,
26
+ "dtype": "float32"
27
+ },
28
+ "registered_name": "keras_hub>MoonshineBackbone"
29
+ }
metadata.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "keras_version": "3.8.0",
3
+ "keras_hub_version": "0.21.0.dev0",
4
+ "parameter_count": 27092768,
5
+ "date_saved": "2025-05-14@09:36:09",
6
+ "tasks": [
7
+ "Seq2SeqLM"
8
+ ]
9
+ }
model.weights.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c9d6d5446139e30b4ee3308934a91fe807f2307de9fb99fcf342767c80167f0
3
+ size 108953032
preprocessor.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.moonshine.moonshine_audio_to_text_preprocessor",
3
+ "class_name": "MoonshineAudioToTextPreprocessor",
4
+ "config": {
5
+ "name": "moonshine_audio_to_text_preprocessor_1",
6
+ "trainable": true,
7
+ "dtype": {
8
+ "module": "keras",
9
+ "class_name": "DTypePolicy",
10
+ "config": {
11
+ "name": "float32"
12
+ },
13
+ "registered_name": null
14
+ },
15
+ "tokenizer": {
16
+ "module": "keras_hub.src.models.moonshine.moonshine_tokenizer",
17
+ "class_name": "MoonshineTokenizer",
18
+ "config": {
19
+ "name": "moonshine_tokenizer_1",
20
+ "trainable": true,
21
+ "dtype": {
22
+ "module": "keras",
23
+ "class_name": "DTypePolicy",
24
+ "config": {
25
+ "name": "int32"
26
+ },
27
+ "registered_name": null
28
+ },
29
+ "config_file": "tokenizer.json",
30
+ "proto": null,
31
+ "sequence_length": null,
32
+ "add_bos": false,
33
+ "add_eos": false
34
+ },
35
+ "registered_name": "keras_hub>MoonshineTokenizer"
36
+ },
37
+ "audio_converter": {
38
+ "module": "keras_hub.src.models.moonshine.moonshine_audio_converter",
39
+ "class_name": "MoonshineAudioConverter",
40
+ "config": {
41
+ "name": "moonshine_audio_converter_1",
42
+ "trainable": true,
43
+ "dtype": {
44
+ "module": "keras",
45
+ "class_name": "DTypePolicy",
46
+ "config": {
47
+ "name": "float32"
48
+ },
49
+ "registered_name": null
50
+ },
51
+ "sampling_rate": 16000,
52
+ "padding_value": 0.0,
53
+ "do_normalize": false
54
+ },
55
+ "registered_name": "keras_hub>MoonshineAudioConverter"
56
+ },
57
+ "config_file": "preprocessor.json"
58
+ },
59
+ "registered_name": "keras_hub>MoonshineAudioToTextPreprocessor"
60
+ }
task.json ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.moonshine.moonshine_audio_to_text",
3
+ "class_name": "MoonshineAudioToText",
4
+ "config": {
5
+ "backbone": {
6
+ "module": "keras_hub.src.models.moonshine.moonshine_backbone",
7
+ "class_name": "MoonshineBackbone",
8
+ "config": {
9
+ "name": "moonshine_backbone_1",
10
+ "trainable": true,
11
+ "vocabulary_size": 32768,
12
+ "filter_dim": 288,
13
+ "encoder_num_layers": 6,
14
+ "decoder_num_layers": 6,
15
+ "hidden_dim": 288,
16
+ "intermediate_dim": 1152,
17
+ "encoder_num_heads": 8,
18
+ "decoder_num_heads": 8,
19
+ "feedforward_expansion_factor": 4,
20
+ "encoder_use_swiglu_activation": false,
21
+ "decoder_use_swiglu_activation": true,
22
+ "max_position_embeddings": 194,
23
+ "pad_head_dim_to_multiple_of": null,
24
+ "partial_rotary_factor": 0.9,
25
+ "dropout": 0.0,
26
+ "initializer_range": 0.02,
27
+ "rope_theta": 10000.0,
28
+ "attention_bias": false,
29
+ "attention_dropout": 0.0,
30
+ "dtype": "float32"
31
+ },
32
+ "registered_name": "keras_hub>MoonshineBackbone"
33
+ },
34
+ "preprocessor": {
35
+ "module": "keras_hub.src.models.moonshine.moonshine_audio_to_text_preprocessor",
36
+ "class_name": "MoonshineAudioToTextPreprocessor",
37
+ "config": {
38
+ "name": "moonshine_audio_to_text_preprocessor_1",
39
+ "trainable": true,
40
+ "dtype": {
41
+ "module": "keras",
42
+ "class_name": "DTypePolicy",
43
+ "config": {
44
+ "name": "float32"
45
+ },
46
+ "registered_name": null
47
+ },
48
+ "tokenizer": {
49
+ "module": "keras_hub.src.models.moonshine.moonshine_tokenizer",
50
+ "class_name": "MoonshineTokenizer",
51
+ "config": {
52
+ "name": "moonshine_tokenizer_1",
53
+ "trainable": true,
54
+ "dtype": {
55
+ "module": "keras",
56
+ "class_name": "DTypePolicy",
57
+ "config": {
58
+ "name": "int32"
59
+ },
60
+ "registered_name": null
61
+ },
62
+ "config_file": "tokenizer.json",
63
+ "proto": null,
64
+ "sequence_length": null,
65
+ "add_bos": false,
66
+ "add_eos": false
67
+ },
68
+ "registered_name": "keras_hub>MoonshineTokenizer"
69
+ },
70
+ "audio_converter": {
71
+ "module": "keras_hub.src.models.moonshine.moonshine_audio_converter",
72
+ "class_name": "MoonshineAudioConverter",
73
+ "config": {
74
+ "name": "moonshine_audio_converter_1",
75
+ "trainable": true,
76
+ "dtype": {
77
+ "module": "keras",
78
+ "class_name": "DTypePolicy",
79
+ "config": {
80
+ "name": "float32"
81
+ },
82
+ "registered_name": null
83
+ },
84
+ "sampling_rate": 16000,
85
+ "padding_value": 0.0,
86
+ "do_normalize": false
87
+ },
88
+ "registered_name": "keras_hub>MoonshineAudioConverter"
89
+ },
90
+ "config_file": "preprocessor.json"
91
+ },
92
+ "registered_name": "keras_hub>MoonshineAudioToTextPreprocessor"
93
+ },
94
+ "name": "moonshine_audio_to_text_1"
95
+ },
96
+ "registered_name": "keras_hub>MoonshineAudioToText"
97
+ }
tokenizer.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.moonshine.moonshine_tokenizer",
3
+ "class_name": "MoonshineTokenizer",
4
+ "config": {
5
+ "name": "moonshine_tokenizer_1",
6
+ "trainable": true,
7
+ "dtype": {
8
+ "module": "keras",
9
+ "class_name": "DTypePolicy",
10
+ "config": {
11
+ "name": "int32"
12
+ },
13
+ "registered_name": null
14
+ },
15
+ "config_file": "tokenizer.json",
16
+ "proto": null,
17
+ "sequence_length": null,
18
+ "add_bos": false,
19
+ "add_eos": false
20
+ },
21
+ "registered_name": "keras_hub>MoonshineTokenizer"
22
+ }