Jarbas commited on
Commit
5c3717e
·
verified ·
1 Parent(s): 1950533

Add CoreML export: stt-los-conformer-transducer-large-punctuated-coreml-4bit

Browse files
README.md ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ tags:
4
+ - automatic-speech-recognition
5
+ - nemo
6
+ - coreml
7
+ - apple-silicon
8
+ - ovos
9
+ ---
10
+
11
+ # stt-los-conformer-transducer-large-punctuated-coreml
12
+
13
+ CoreML export of [BSC-LT/stt_los_conformer_transducer_large_punctuated](https://huggingface.co/BSC-LT/stt_los_conformer_transducer_large_punctuated) for Apple Silicon.
14
+
15
+ Exported with [ovos-stt-plugin-coreml](https://github.com/TigreGotico/ovos-stt-plugin-coreml).
16
+
17
+ ## Usage
18
+
19
+ ```python
20
+ from ovos_stt_plugin_coreml import CoremlSTT
21
+ stt = CoremlSTT(config={"repo_id": "OpenVoiceOS/stt-los-conformer-transducer-large-punctuated-coreml"})
22
+ ```
metadata.json ADDED
@@ -0,0 +1,113 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_id": "BSC-LT/stt_los_conformer_transducer_large_punctuated",
3
+ "model_type": "parakeet_rnnt",
4
+ "language": "",
5
+ "sample_rate": 16000,
6
+ "max_audio_seconds": 15.0,
7
+ "max_audio_samples": 240000,
8
+ "vocab_size": 300,
9
+ "blank_id": 300,
10
+ "checkpoint": {
11
+ "type": "pretrained",
12
+ "model_id": "BSC-LT/stt_los_conformer_transducer_large_punctuated"
13
+ },
14
+ "coreml": {
15
+ "compute_precision": "FLOAT32",
16
+ "quantization": "4bit_palettize_kmeans"
17
+ },
18
+ "components": {
19
+ "mel_encoder": {
20
+ "path": "parakeet_mel_encoder.mlpackage",
21
+ "inputs": {
22
+ "audio_signal": [
23
+ 1,
24
+ 240000
25
+ ],
26
+ "audio_length": [
27
+ 1
28
+ ]
29
+ },
30
+ "outputs": {
31
+ "encoder": [
32
+ 1,
33
+ 512,
34
+ 376
35
+ ],
36
+ "encoder_length": [
37
+ 1
38
+ ]
39
+ }
40
+ },
41
+ "decoder": {
42
+ "path": "parakeet_decoder.mlpackage",
43
+ "inputs": {
44
+ "targets": [
45
+ 1,
46
+ 1
47
+ ],
48
+ "target_length": [
49
+ 1
50
+ ],
51
+ "h_in": [
52
+ 1,
53
+ 1,
54
+ 640
55
+ ],
56
+ "c_in": [
57
+ 1,
58
+ 1,
59
+ 640
60
+ ]
61
+ },
62
+ "outputs": {
63
+ "decoder": [
64
+ 1,
65
+ 640,
66
+ 1
67
+ ],
68
+ "h_out": [
69
+ 1,
70
+ 1,
71
+ 640
72
+ ],
73
+ "c_out": [
74
+ 1,
75
+ 1,
76
+ 640
77
+ ]
78
+ }
79
+ },
80
+ "joint_decision_single_step": {
81
+ "path": "parakeet_joint_decision_single_step.mlpackage",
82
+ "inputs": {
83
+ "encoder_step": [
84
+ 1,
85
+ 512,
86
+ 1
87
+ ],
88
+ "decoder_step": [
89
+ 1,
90
+ 640,
91
+ 1
92
+ ]
93
+ },
94
+ "outputs": {
95
+ "token_id": [
96
+ 1,
97
+ 1,
98
+ 1
99
+ ],
100
+ "token_prob": [
101
+ 1,
102
+ 1,
103
+ 1
104
+ ],
105
+ "duration": [
106
+ 1,
107
+ 1,
108
+ 1
109
+ ]
110
+ }
111
+ }
112
+ }
113
+ }
parakeet_decoder.mlpackage/Data/com.apple.CoreML/model.mlmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb3c1cd93bb2f5782d7d627c0edc305ce9015dcd837180326d81562c2146808f
3
+ size 6225
parakeet_decoder.mlpackage/Data/com.apple.CoreML/weights/weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52e1f8b6f65e104595a2953e6beac017c2acc240e12b180fc4a1a63e20c0b27e
3
+ size 1745664
parakeet_decoder.mlpackage/Manifest.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "fileFormatVersion": "1.0.0",
3
+ "itemInfoEntries": {
4
+ "1F0FA901-F3F1-4695-B0AA-CAA66CE5BCE2": {
5
+ "author": "com.apple.CoreML",
6
+ "description": "CoreML Model Weights",
7
+ "name": "weights",
8
+ "path": "com.apple.CoreML/weights"
9
+ },
10
+ "29B3349E-DD01-4E9A-B4F1-0E4DF9BB0FB0": {
11
+ "author": "com.apple.CoreML",
12
+ "description": "CoreML Model Specification",
13
+ "name": "model.mlmodel",
14
+ "path": "com.apple.CoreML/model.mlmodel"
15
+ }
16
+ },
17
+ "rootModelIdentifier": "29B3349E-DD01-4E9A-B4F1-0E4DF9BB0FB0"
18
+ }
parakeet_joint_decision_single_step.mlpackage/Data/com.apple.CoreML/model.mlmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd1a63333a88f2ca2063f2e564c337a8bdf62f3c56763cfdb0881c48aa2d27a5
3
+ size 5575
parakeet_joint_decision_single_step.mlpackage/Data/com.apple.CoreML/weights/weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f1d2143f8f7ee436492eda7e1a2cbfffe2eda9a20e748662b758fd8f8a3ecb9
3
+ size 472128
parakeet_joint_decision_single_step.mlpackage/Manifest.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "fileFormatVersion": "1.0.0",
3
+ "itemInfoEntries": {
4
+ "198B8ADC-92D8-432D-BB72-EF57708F3CC1": {
5
+ "author": "com.apple.CoreML",
6
+ "description": "CoreML Model Weights",
7
+ "name": "weights",
8
+ "path": "com.apple.CoreML/weights"
9
+ },
10
+ "3D4CBDFF-BFBD-4EFD-B499-76008DF8A957": {
11
+ "author": "com.apple.CoreML",
12
+ "description": "CoreML Model Specification",
13
+ "name": "model.mlmodel",
14
+ "path": "com.apple.CoreML/model.mlmodel"
15
+ }
16
+ },
17
+ "rootModelIdentifier": "3D4CBDFF-BFBD-4EFD-B499-76008DF8A957"
18
+ }
parakeet_mel_encoder.mlpackage/Data/com.apple.CoreML/model.mlmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fd3403eee62b52de6f085d385de47f301021a029fb7c2708f426ed689ad146b
3
+ size 1321300
parakeet_mel_encoder.mlpackage/Data/com.apple.CoreML/weights/weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a94d8c4a56a620478bc8f07bc0eef3d63e6ea2401bd19d71a6488a4267d85253
3
+ size 59774208
parakeet_mel_encoder.mlpackage/Manifest.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "fileFormatVersion": "1.0.0",
3
+ "itemInfoEntries": {
4
+ "082CD348-11A2-4F31-843D-8F4128F0A327": {
5
+ "author": "com.apple.CoreML",
6
+ "description": "CoreML Model Weights",
7
+ "name": "weights",
8
+ "path": "com.apple.CoreML/weights"
9
+ },
10
+ "19AD95AB-B29D-4E1C-94BC-6BC020F38E35": {
11
+ "author": "com.apple.CoreML",
12
+ "description": "CoreML Model Specification",
13
+ "name": "model.mlmodel",
14
+ "path": "com.apple.CoreML/model.mlmodel"
15
+ }
16
+ },
17
+ "rootModelIdentifier": "19AD95AB-B29D-4E1C-94BC-6BC020F38E35"
18
+ }
quantization_summary.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "parakeet_decoder": {
3
+ "original_mb": 13.3,
4
+ "4bit_mb": 1.7,
5
+ "compression": 7.93
6
+ },
7
+ "parakeet_joint_decision_single_step": {
8
+ "original_mb": 3.6,
9
+ "4bit_mb": 0.5,
10
+ "compression": 7.8
11
+ },
12
+ "parakeet_mel_encoder": {
13
+ "original_mb": 449.8,
14
+ "4bit_mb": 58.3,
15
+ "compression": 7.72
16
+ }
17
+ }
vocab.json ADDED
@@ -0,0 +1,302 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ "<unk>",
3
+ "▁d",
4
+ "en",
5
+ "es",
6
+ "er",
7
+ "▁a",
8
+ "ta",
9
+ "▁p",
10
+ "ar",
11
+ "▁e",
12
+ "▁de",
13
+ "▁c",
14
+ "os",
15
+ "an",
16
+ "▁l",
17
+ "qu",
18
+ "▁s",
19
+ "on",
20
+ "or",
21
+ "in",
22
+ "▁m",
23
+ "▁qu",
24
+ "te",
25
+ "un",
26
+ "re",
27
+ "▁que",
28
+ "al",
29
+ "ci",
30
+ "as",
31
+ "▁n",
32
+ "▁h",
33
+ "▁b",
34
+ "ra",
35
+ "ti",
36
+ "▁es",
37
+ "▁en",
38
+ "tu",
39
+ "ri",
40
+ "▁v",
41
+ "▁i",
42
+ "▁la",
43
+ "▁t",
44
+ "▁f",
45
+ "da",
46
+ "ko",
47
+ "▁g",
48
+ "tz",
49
+ "▁un",
50
+ "is",
51
+ "om",
52
+ "▁el",
53
+ "▁E",
54
+ "ol",
55
+ "de",
56
+ "do",
57
+ "ro",
58
+ "▁se",
59
+ "▁o",
60
+ "la",
61
+ "ent",
62
+ "▁di",
63
+ "ia",
64
+ "ón",
65
+ "▁con",
66
+ "era",
67
+ "us",
68
+ "▁per",
69
+ "▁da",
70
+ "res",
71
+ "il",
72
+ "ur",
73
+ "di",
74
+ "ca",
75
+ "▁P",
76
+ "▁al",
77
+ "to",
78
+ "▁A",
79
+ "▁no",
80
+ "le",
81
+ "ak",
82
+ "▁y",
83
+ "at",
84
+ "gu",
85
+ "▁com",
86
+ "▁in",
87
+ "▁z",
88
+ "▁re",
89
+ "go",
90
+ "▁par",
91
+ "tan",
92
+ "▁ha",
93
+ "▁C",
94
+ "ma",
95
+ "em",
96
+ "el",
97
+ "▁é",
98
+ "am",
99
+ "ita",
100
+ "▁eta",
101
+ "▁S",
102
+ "ec",
103
+ "▁du",
104
+ "ción",
105
+ "▁del",
106
+ "▁L",
107
+ "ai",
108
+ "▁pro",
109
+ "▁por",
110
+ "ab",
111
+ "▁j",
112
+ "ten",
113
+ "ts",
114
+ "▁po",
115
+ "li",
116
+ "mp",
117
+ "▁B",
118
+ "▁D",
119
+ "ente",
120
+ "▁M",
121
+ "mb",
122
+ "▁des",
123
+ "ul",
124
+ "▁ta",
125
+ "io",
126
+ "▁G",
127
+ "▁an",
128
+ "ha",
129
+ "ir",
130
+ "▁una",
131
+ "du",
132
+ "▁as",
133
+ "▁to",
134
+ "co",
135
+ "ez",
136
+ "cia",
137
+ "ix",
138
+ "ter",
139
+ "ue",
140
+ "▁do",
141
+ "▁te",
142
+ "iz",
143
+ "ía",
144
+ "▁er",
145
+ "▁para",
146
+ "ga",
147
+ "▁ba",
148
+ "▁eg",
149
+ "▁si",
150
+ "me",
151
+ "▁na",
152
+ "eh",
153
+ "tes",
154
+ "dos",
155
+ "▁H",
156
+ "▁Es",
157
+ "▁ez",
158
+ "▁esta",
159
+ "ka",
160
+ "▁les",
161
+ "ida",
162
+ "▁su",
163
+ "▁me",
164
+ "▁los",
165
+ "ora",
166
+ "tra",
167
+ "▁T",
168
+ "és",
169
+ "ció",
170
+ "va",
171
+ "tr",
172
+ "lo",
173
+ "ari",
174
+ "enta",
175
+ "pa",
176
+ "si",
177
+ "▁N",
178
+ "▁",
179
+ "a",
180
+ "e",
181
+ "o",
182
+ "n",
183
+ "r",
184
+ "i",
185
+ "s",
186
+ "t",
187
+ "u",
188
+ "l",
189
+ "d",
190
+ "c",
191
+ "m",
192
+ "p",
193
+ "b",
194
+ "g",
195
+ ",",
196
+ "k",
197
+ "z",
198
+ ".",
199
+ "h",
200
+ "q",
201
+ "v",
202
+ "f",
203
+ "ó",
204
+ "x",
205
+ "é",
206
+ "í",
207
+ "E",
208
+ "y",
209
+ "á",
210
+ "j",
211
+ "?",
212
+ "P",
213
+ "A",
214
+ "!",
215
+ "C",
216
+ "ñ",
217
+ "S",
218
+ "L",
219
+ "B",
220
+ "ú",
221
+ "D",
222
+ "M",
223
+ "G",
224
+ "à",
225
+ "'",
226
+ ":",
227
+ "è",
228
+ "H",
229
+ "T",
230
+ "ò",
231
+ "N",
232
+ "I",
233
+ "O",
234
+ "F",
235
+ "U",
236
+ "Q",
237
+ "-",
238
+ "R",
239
+ "V",
240
+ "J",
241
+ "ç",
242
+ "Y",
243
+ "K",
244
+ "Z",
245
+ "0",
246
+ "X",
247
+ "É",
248
+ "¿",
249
+ "2",
250
+ "1",
251
+ "w",
252
+ "ü",
253
+ "ï",
254
+ "W",
255
+ "5",
256
+ ";",
257
+ ")",
258
+ "3",
259
+ "4",
260
+ "(",
261
+ "·",
262
+ "7",
263
+ "Á",
264
+ "6",
265
+ "9",
266
+ "%",
267
+ "¡",
268
+ "8",
269
+ "\"",
270
+ "“",
271
+ "”",
272
+ "À",
273
+ "/",
274
+ "Ó",
275
+ "Í",
276
+ "Ú",
277
+ "–",
278
+ "ë",
279
+ "Ò",
280
+ "ã",
281
+ "−",
282
+ "È",
283
+ "Ñ",
284
+ "ö",
285
+ "ê",
286
+ "́",
287
+ "’",
288
+ "Ç",
289
+ "â",
290
+ "Ü",
291
+ "ô",
292
+ "õ",
293
+ "«",
294
+ "°",
295
+ "»",
296
+ "ä",
297
+ "_",
298
+ "ù",
299
+ "©",
300
+ "Â",
301
+ "Ö"
302
+ ]
Free AI Image Generator No sign-up. Instant results. Open Now