alexwengg commited on
Commit
8f32251
·
verified ·
1 Parent(s): 08bf7d0

Upload 8 files

Browse files
q8/cohere_decoder_cache_external_v2.mlmodelc/analytics/coremldata.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c2f8083545d32e711223ccd852f9bf9a8b5714b7f46513c828ad60c1926d6e90
3
+ size 243
q8/cohere_decoder_cache_external_v2.mlmodelc/coremldata.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dbd9290c1188f7e529f3c16a41428b7d0bc978e662a269385125d996e5516e8f
3
+ size 1562
q8/cohere_decoder_cache_external_v2.mlmodelc/metadata.json ADDED
@@ -0,0 +1,438 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "metadataOutputVersion" : "3.0",
4
+ "shortDescription" : "Cohere Transcribe decoder (cache-external, static shapes, ANE-friendly)",
5
+ "outputSchema" : [
6
+ {
7
+ "hasShapeFlexibility" : "0",
8
+ "isOptional" : "0",
9
+ "dataType" : "Float32",
10
+ "formattedType" : "MultiArray (Float32 1 × 16384)",
11
+ "shortDescription" : "",
12
+ "shape" : "[1, 16384]",
13
+ "name" : "logits",
14
+ "type" : "MultiArray"
15
+ },
16
+ {
17
+ "hasShapeFlexibility" : "0",
18
+ "isOptional" : "0",
19
+ "dataType" : "Float32",
20
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
21
+ "shortDescription" : "",
22
+ "shape" : "[1, 8, 108, 128]",
23
+ "name" : "k_cache_0_out",
24
+ "type" : "MultiArray"
25
+ },
26
+ {
27
+ "hasShapeFlexibility" : "0",
28
+ "isOptional" : "0",
29
+ "dataType" : "Float32",
30
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
31
+ "shortDescription" : "",
32
+ "shape" : "[1, 8, 108, 128]",
33
+ "name" : "v_cache_0_out",
34
+ "type" : "MultiArray"
35
+ },
36
+ {
37
+ "hasShapeFlexibility" : "0",
38
+ "isOptional" : "0",
39
+ "dataType" : "Float32",
40
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
41
+ "shortDescription" : "",
42
+ "shape" : "[1, 8, 108, 128]",
43
+ "name" : "k_cache_1_out",
44
+ "type" : "MultiArray"
45
+ },
46
+ {
47
+ "hasShapeFlexibility" : "0",
48
+ "isOptional" : "0",
49
+ "dataType" : "Float32",
50
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
51
+ "shortDescription" : "",
52
+ "shape" : "[1, 8, 108, 128]",
53
+ "name" : "v_cache_1_out",
54
+ "type" : "MultiArray"
55
+ },
56
+ {
57
+ "hasShapeFlexibility" : "0",
58
+ "isOptional" : "0",
59
+ "dataType" : "Float32",
60
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
61
+ "shortDescription" : "",
62
+ "shape" : "[1, 8, 108, 128]",
63
+ "name" : "k_cache_2_out",
64
+ "type" : "MultiArray"
65
+ },
66
+ {
67
+ "hasShapeFlexibility" : "0",
68
+ "isOptional" : "0",
69
+ "dataType" : "Float32",
70
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
71
+ "shortDescription" : "",
72
+ "shape" : "[1, 8, 108, 128]",
73
+ "name" : "v_cache_2_out",
74
+ "type" : "MultiArray"
75
+ },
76
+ {
77
+ "hasShapeFlexibility" : "0",
78
+ "isOptional" : "0",
79
+ "dataType" : "Float32",
80
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
81
+ "shortDescription" : "",
82
+ "shape" : "[1, 8, 108, 128]",
83
+ "name" : "k_cache_3_out",
84
+ "type" : "MultiArray"
85
+ },
86
+ {
87
+ "hasShapeFlexibility" : "0",
88
+ "isOptional" : "0",
89
+ "dataType" : "Float32",
90
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
91
+ "shortDescription" : "",
92
+ "shape" : "[1, 8, 108, 128]",
93
+ "name" : "v_cache_3_out",
94
+ "type" : "MultiArray"
95
+ },
96
+ {
97
+ "hasShapeFlexibility" : "0",
98
+ "isOptional" : "0",
99
+ "dataType" : "Float32",
100
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
101
+ "shortDescription" : "",
102
+ "shape" : "[1, 8, 108, 128]",
103
+ "name" : "k_cache_4_out",
104
+ "type" : "MultiArray"
105
+ },
106
+ {
107
+ "hasShapeFlexibility" : "0",
108
+ "isOptional" : "0",
109
+ "dataType" : "Float32",
110
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
111
+ "shortDescription" : "",
112
+ "shape" : "[1, 8, 108, 128]",
113
+ "name" : "v_cache_4_out",
114
+ "type" : "MultiArray"
115
+ },
116
+ {
117
+ "hasShapeFlexibility" : "0",
118
+ "isOptional" : "0",
119
+ "dataType" : "Float32",
120
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
121
+ "shortDescription" : "",
122
+ "shape" : "[1, 8, 108, 128]",
123
+ "name" : "k_cache_5_out",
124
+ "type" : "MultiArray"
125
+ },
126
+ {
127
+ "hasShapeFlexibility" : "0",
128
+ "isOptional" : "0",
129
+ "dataType" : "Float32",
130
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
131
+ "shortDescription" : "",
132
+ "shape" : "[1, 8, 108, 128]",
133
+ "name" : "v_cache_5_out",
134
+ "type" : "MultiArray"
135
+ },
136
+ {
137
+ "hasShapeFlexibility" : "0",
138
+ "isOptional" : "0",
139
+ "dataType" : "Float32",
140
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
141
+ "shortDescription" : "",
142
+ "shape" : "[1, 8, 108, 128]",
143
+ "name" : "k_cache_6_out",
144
+ "type" : "MultiArray"
145
+ },
146
+ {
147
+ "hasShapeFlexibility" : "0",
148
+ "isOptional" : "0",
149
+ "dataType" : "Float32",
150
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
151
+ "shortDescription" : "",
152
+ "shape" : "[1, 8, 108, 128]",
153
+ "name" : "v_cache_6_out",
154
+ "type" : "MultiArray"
155
+ },
156
+ {
157
+ "hasShapeFlexibility" : "0",
158
+ "isOptional" : "0",
159
+ "dataType" : "Float32",
160
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
161
+ "shortDescription" : "",
162
+ "shape" : "[1, 8, 108, 128]",
163
+ "name" : "k_cache_7_out",
164
+ "type" : "MultiArray"
165
+ },
166
+ {
167
+ "hasShapeFlexibility" : "0",
168
+ "isOptional" : "0",
169
+ "dataType" : "Float32",
170
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
171
+ "shortDescription" : "",
172
+ "shape" : "[1, 8, 108, 128]",
173
+ "name" : "v_cache_7_out",
174
+ "type" : "MultiArray"
175
+ }
176
+ ],
177
+ "storagePrecision" : "Float16",
178
+ "modelParameters" : [
179
+
180
+ ],
181
+ "author" : "FluidInference",
182
+ "specificationVersion" : 8,
183
+ "mlProgramOperationTypeHistogram" : {
184
+ "Ios17.mul" : 16,
185
+ "Ios17.linear" : 81,
186
+ "Select" : 1,
187
+ "Ios17.transpose" : 64,
188
+ "Ios17.matmul" : 32,
189
+ "Ios17.gather" : 2,
190
+ "Ios17.add" : 42,
191
+ "Ios16.relu" : 8,
192
+ "Tile" : 1,
193
+ "Ios17.greaterEqual" : 1,
194
+ "Ios16.softmax" : 16,
195
+ "Ios17.scatterAlongAxis" : 16,
196
+ "Ios17.squeeze" : 1,
197
+ "Ios17.cast" : 40,
198
+ "Ios17.layerNorm" : 26,
199
+ "Ios17.reshape" : 67
200
+ },
201
+ "computePrecision" : "Mixed (Float16, Float32, Int16, Int32, UInt16)",
202
+ "isUpdatable" : "0",
203
+ "stateSchema" : [
204
+
205
+ ],
206
+ "availability" : {
207
+ "macOS" : "14.0",
208
+ "tvOS" : "17.0",
209
+ "visionOS" : "1.0",
210
+ "watchOS" : "10.0",
211
+ "iOS" : "17.0",
212
+ "macCatalyst" : "17.0"
213
+ },
214
+ "modelType" : {
215
+ "name" : "MLModelType_mlProgram"
216
+ },
217
+ "inputSchema" : [
218
+ {
219
+ "hasShapeFlexibility" : "0",
220
+ "isOptional" : "0",
221
+ "dataType" : "Int32",
222
+ "formattedType" : "MultiArray (Int32 1 × 1)",
223
+ "shortDescription" : "",
224
+ "shape" : "[1, 1]",
225
+ "name" : "input_id",
226
+ "type" : "MultiArray"
227
+ },
228
+ {
229
+ "hasShapeFlexibility" : "0",
230
+ "isOptional" : "0",
231
+ "dataType" : "Int32",
232
+ "formattedType" : "MultiArray (Int32 1 × 1)",
233
+ "shortDescription" : "",
234
+ "shape" : "[1, 1]",
235
+ "name" : "position_id",
236
+ "type" : "MultiArray"
237
+ },
238
+ {
239
+ "hasShapeFlexibility" : "0",
240
+ "isOptional" : "0",
241
+ "dataType" : "Float32",
242
+ "formattedType" : "MultiArray (Float32 1 × 438 × 1024)",
243
+ "shortDescription" : "",
244
+ "shape" : "[1, 438, 1024]",
245
+ "name" : "encoder_hidden_states",
246
+ "type" : "MultiArray"
247
+ },
248
+ {
249
+ "hasShapeFlexibility" : "0",
250
+ "isOptional" : "0",
251
+ "dataType" : "Float32",
252
+ "formattedType" : "MultiArray (Float32 1 × 1 × 1 × 438)",
253
+ "shortDescription" : "",
254
+ "shape" : "[1, 1, 1, 438]",
255
+ "name" : "cross_attention_mask",
256
+ "type" : "MultiArray"
257
+ },
258
+ {
259
+ "hasShapeFlexibility" : "0",
260
+ "isOptional" : "0",
261
+ "dataType" : "Float32",
262
+ "formattedType" : "MultiArray (Float32 1 × 1 × 1 × 108)",
263
+ "shortDescription" : "",
264
+ "shape" : "[1, 1, 1, 108]",
265
+ "name" : "attention_mask",
266
+ "type" : "MultiArray"
267
+ },
268
+ {
269
+ "hasShapeFlexibility" : "0",
270
+ "isOptional" : "0",
271
+ "dataType" : "Float32",
272
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
273
+ "shortDescription" : "",
274
+ "shape" : "[1, 8, 108, 128]",
275
+ "name" : "k_cache_0",
276
+ "type" : "MultiArray"
277
+ },
278
+ {
279
+ "hasShapeFlexibility" : "0",
280
+ "isOptional" : "0",
281
+ "dataType" : "Float32",
282
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
283
+ "shortDescription" : "",
284
+ "shape" : "[1, 8, 108, 128]",
285
+ "name" : "v_cache_0",
286
+ "type" : "MultiArray"
287
+ },
288
+ {
289
+ "hasShapeFlexibility" : "0",
290
+ "isOptional" : "0",
291
+ "dataType" : "Float32",
292
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
293
+ "shortDescription" : "",
294
+ "shape" : "[1, 8, 108, 128]",
295
+ "name" : "k_cache_1",
296
+ "type" : "MultiArray"
297
+ },
298
+ {
299
+ "hasShapeFlexibility" : "0",
300
+ "isOptional" : "0",
301
+ "dataType" : "Float32",
302
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
303
+ "shortDescription" : "",
304
+ "shape" : "[1, 8, 108, 128]",
305
+ "name" : "v_cache_1",
306
+ "type" : "MultiArray"
307
+ },
308
+ {
309
+ "hasShapeFlexibility" : "0",
310
+ "isOptional" : "0",
311
+ "dataType" : "Float32",
312
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
313
+ "shortDescription" : "",
314
+ "shape" : "[1, 8, 108, 128]",
315
+ "name" : "k_cache_2",
316
+ "type" : "MultiArray"
317
+ },
318
+ {
319
+ "hasShapeFlexibility" : "0",
320
+ "isOptional" : "0",
321
+ "dataType" : "Float32",
322
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
323
+ "shortDescription" : "",
324
+ "shape" : "[1, 8, 108, 128]",
325
+ "name" : "v_cache_2",
326
+ "type" : "MultiArray"
327
+ },
328
+ {
329
+ "hasShapeFlexibility" : "0",
330
+ "isOptional" : "0",
331
+ "dataType" : "Float32",
332
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
333
+ "shortDescription" : "",
334
+ "shape" : "[1, 8, 108, 128]",
335
+ "name" : "k_cache_3",
336
+ "type" : "MultiArray"
337
+ },
338
+ {
339
+ "hasShapeFlexibility" : "0",
340
+ "isOptional" : "0",
341
+ "dataType" : "Float32",
342
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
343
+ "shortDescription" : "",
344
+ "shape" : "[1, 8, 108, 128]",
345
+ "name" : "v_cache_3",
346
+ "type" : "MultiArray"
347
+ },
348
+ {
349
+ "hasShapeFlexibility" : "0",
350
+ "isOptional" : "0",
351
+ "dataType" : "Float32",
352
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
353
+ "shortDescription" : "",
354
+ "shape" : "[1, 8, 108, 128]",
355
+ "name" : "k_cache_4",
356
+ "type" : "MultiArray"
357
+ },
358
+ {
359
+ "hasShapeFlexibility" : "0",
360
+ "isOptional" : "0",
361
+ "dataType" : "Float32",
362
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
363
+ "shortDescription" : "",
364
+ "shape" : "[1, 8, 108, 128]",
365
+ "name" : "v_cache_4",
366
+ "type" : "MultiArray"
367
+ },
368
+ {
369
+ "hasShapeFlexibility" : "0",
370
+ "isOptional" : "0",
371
+ "dataType" : "Float32",
372
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
373
+ "shortDescription" : "",
374
+ "shape" : "[1, 8, 108, 128]",
375
+ "name" : "k_cache_5",
376
+ "type" : "MultiArray"
377
+ },
378
+ {
379
+ "hasShapeFlexibility" : "0",
380
+ "isOptional" : "0",
381
+ "dataType" : "Float32",
382
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
383
+ "shortDescription" : "",
384
+ "shape" : "[1, 8, 108, 128]",
385
+ "name" : "v_cache_5",
386
+ "type" : "MultiArray"
387
+ },
388
+ {
389
+ "hasShapeFlexibility" : "0",
390
+ "isOptional" : "0",
391
+ "dataType" : "Float32",
392
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
393
+ "shortDescription" : "",
394
+ "shape" : "[1, 8, 108, 128]",
395
+ "name" : "k_cache_6",
396
+ "type" : "MultiArray"
397
+ },
398
+ {
399
+ "hasShapeFlexibility" : "0",
400
+ "isOptional" : "0",
401
+ "dataType" : "Float32",
402
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
403
+ "shortDescription" : "",
404
+ "shape" : "[1, 8, 108, 128]",
405
+ "name" : "v_cache_6",
406
+ "type" : "MultiArray"
407
+ },
408
+ {
409
+ "hasShapeFlexibility" : "0",
410
+ "isOptional" : "0",
411
+ "dataType" : "Float32",
412
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
413
+ "shortDescription" : "",
414
+ "shape" : "[1, 8, 108, 128]",
415
+ "name" : "k_cache_7",
416
+ "type" : "MultiArray"
417
+ },
418
+ {
419
+ "hasShapeFlexibility" : "0",
420
+ "isOptional" : "0",
421
+ "dataType" : "Float32",
422
+ "formattedType" : "MultiArray (Float32 1 × 8 × 108 × 128)",
423
+ "shortDescription" : "",
424
+ "shape" : "[1, 8, 108, 128]",
425
+ "name" : "v_cache_7",
426
+ "type" : "MultiArray"
427
+ }
428
+ ],
429
+ "userDefinedMetadata" : {
430
+ "com.github.apple.coremltools.conversion_date" : "2026-04-23",
431
+ "com.github.apple.coremltools.source" : "torch==2.7.0",
432
+ "com.github.apple.coremltools.version" : "9.0b1",
433
+ "com.github.apple.coremltools.source_dialect" : "TorchScript"
434
+ },
435
+ "generatedClassName" : "cohere_decoder_cache_external_static",
436
+ "method" : "predict"
437
+ }
438
+ ]
q8/cohere_decoder_cache_external_v2.mlmodelc/model.mil ADDED
The diff for this file is too large to render. See raw diff
 
q8/cohere_decoder_cache_external_v2.mlmodelc/weights/weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86eac4bce6ca5d4ca49f5bf1078a1b42ab75f59d7c809678592a493742c29568
3
+ size 304453120
q8/cohere_decoder_cache_external_v2.mlpackage/Data/com.apple.CoreML/model.mlmodel ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fa37e633da0bea23a8b85c6717d00c2f6fc0d603de157613bf3c3a7bee456e5
3
+ size 176203
q8/cohere_decoder_cache_external_v2.mlpackage/Data/com.apple.CoreML/weights/weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86eac4bce6ca5d4ca49f5bf1078a1b42ab75f59d7c809678592a493742c29568
3
+ size 304453120
q8/cohere_decoder_cache_external_v2.mlpackage/Manifest.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "fileFormatVersion": "1.0.0",
3
+ "itemInfoEntries": {
4
+ "0F21AD6A-C801-4F8A-84F8-24BDD771E1E2": {
5
+ "author": "com.apple.CoreML",
6
+ "description": "CoreML Model Weights",
7
+ "name": "weights",
8
+ "path": "com.apple.CoreML/weights"
9
+ },
10
+ "E775E950-2F3A-448E-BCA7-63B623976AF6": {
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": "E775E950-2F3A-448E-BCA7-63B623976AF6"
18
+ }