forked from triton-inference-server/server
-
Notifications
You must be signed in to change notification settings - Fork 0
/
model_config.proto
1736 lines (1576 loc) · 58.4 KB
/
model_config.proto
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
// Copyright (c) 2018-2020, NVIDIA CORPORATION. All rights reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions
// are met:
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer in the
// documentation and/or other materials provided with the distribution.
// * Neither the name of NVIDIA CORPORATION nor the names of its
// contributors may be used to endorse or promote products derived
// from this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
// PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
// CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
// PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
// OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//
// Copyright (c) 2018, TensorFlow Authors. All rights reserved.
syntax = "proto3";
package inference;
//@@.. cpp:namespace:: inference
//@@
//@@.. cpp:enum:: DataType
//@@
//@@ Data types supported for input and output tensors.
//@@
enum DataType {
//@@ .. cpp:enumerator:: DataType::INVALID = 0
TYPE_INVALID = 0;
//@@ .. cpp:enumerator:: DataType::BOOL = 1
TYPE_BOOL = 1;
//@@ .. cpp:enumerator:: DataType::UINT8 = 2
TYPE_UINT8 = 2;
//@@ .. cpp:enumerator:: DataType::UINT16 = 3
TYPE_UINT16 = 3;
//@@ .. cpp:enumerator:: DataType::UINT32 = 4
TYPE_UINT32 = 4;
//@@ .. cpp:enumerator:: DataType::UINT64 = 5
TYPE_UINT64 = 5;
//@@ .. cpp:enumerator:: DataType::INT8 = 6
TYPE_INT8 = 6;
//@@ .. cpp:enumerator:: DataType::INT16 = 7
TYPE_INT16 = 7;
//@@ .. cpp:enumerator:: DataType::INT32 = 8
TYPE_INT32 = 8;
//@@ .. cpp:enumerator:: DataType::INT64 = 9
TYPE_INT64 = 9;
//@@ .. cpp:enumerator:: DataType::FP16 = 10
TYPE_FP16 = 10;
//@@ .. cpp:enumerator:: DataType::FP32 = 11
TYPE_FP32 = 11;
//@@ .. cpp:enumerator:: DataType::FP64 = 12
TYPE_FP64 = 12;
//@@ .. cpp:enumerator:: DataType::STRING = 13
TYPE_STRING = 13;
}
//@@
//@@ .. cpp:var:: message ModelRateLimiter
//@@
//@@ The specifications required by the rate limiter to properly
//@@ schedule the inference requests across the different models
//@@ and their instances.
//@@
message ModelRateLimiter
{
//@@ .. cpp:var:: message Resource
//@@
//@@ The resource property.
//@@
message Resource
{
//@@ .. cpp:var:: string name
//@@
//@@ The name associated with the resource.
//@@
string name = 1;
//@@ .. cpp:var:: bool global
//@@
//@@ Whether or not the resource is global. If true then the resource
//@@ is assumed to be shared among the devices otherwise specified
//@@ count of the resource is assumed for each device associated
//@@ with the instance.
//@@
bool global = 2;
//@@ .. cpp:var:: uint32 count
//@@
//@@ The number of resources required for the execution of the model
//@@ instance.
//@@
uint32 count = 3;
}
//@@ .. cpp:var:: Resource resources (repeated)
//@@
//@@ The resources required to execute the request on a model instance.
//@@ Resources are just names with a corresponding count. The execution
//@@ of the instance will be blocked until the specificied resources are
//@@ available. By default an instance uses no rate-limiter resources.
//@@
repeated Resource resources = 1;
//@@ .. cpp:var:: uint32 priority
//@@
//@@ The weighting value to be used for prioritizing across instances.
//@@ An instance with priority 2 will be given 1/2 the number of
//@@ scheduling chances as an instance_group with priority 1. The
//@@ default priority is 1.
//@@
uint32 priority = 2;
}
//@@
//@@.. cpp:var:: message ModelInstanceGroup
//@@
//@@ A group of one or more instances of a model and resources made
//@@ available for those instances.
//@@
message ModelInstanceGroup
{
//@@
//@@ .. cpp:enum:: Kind
//@@
//@@ Kind of this instance group.
//@@
enum Kind {
//@@ .. cpp:enumerator:: Kind::KIND_AUTO = 0
//@@
//@@ This instance group represents instances that can run on either
//@@ CPU or GPU. If all GPUs listed in 'gpus' are available then
//@@ instances will be created on GPU(s), otherwise instances will
//@@ be created on CPU.
//@@
KIND_AUTO = 0;
//@@ .. cpp:enumerator:: Kind::KIND_GPU = 1
//@@
//@@ This instance group represents instances that must run on the
//@@ GPU.
//@@
KIND_GPU = 1;
//@@ .. cpp:enumerator:: Kind::KIND_CPU = 2
//@@
//@@ This instance group represents instances that must run on the
//@@ CPU.
//@@
KIND_CPU = 2;
//@@ .. cpp:enumerator:: Kind::KIND_MODEL = 3
//@@
//@@ This instance group represents instances that should run on the
//@@ CPU and/or GPU(s) as specified by the model or backend itself.
//@@ The inference server will not override the model/backend
//@@ settings.
//@@ Currently, this option is supported only for Tensorflow models.
//@@
KIND_MODEL = 3;
}
//@@ .. cpp:var:: string name
//@@
//@@ Optional name of this group of instances. If not specified the
//@@ name will be formed as <model name>_<group number>. The name of
//@@ individual instances will be further formed by a unique instance
//@@ number and GPU index:
//@@
string name = 1;
//@@ .. cpp:var:: Kind kind
//@@
//@@ The kind of this instance group. Default is KIND_AUTO. If
//@@ KIND_AUTO or KIND_GPU then both 'count' and 'gpu' are valid and
//@@ may be specified. If KIND_CPU or KIND_MODEL only 'count' is valid
//@@ and 'gpu' cannot be specified.
//@@
Kind kind = 4;
//@@ .. cpp:var:: int32 count
//@@
//@@ For a group assigned to GPU, the number of instances created for
//@@ each GPU listed in 'gpus'. For a group assigned to CPU the number
//@@ of instances created. Default is 1.
int32 count = 2;
//@@ .. cpp:var:: ModelRateLimiter rate_limiter
//@@
//@@ The rate limiter specific settings to be associated with this
//@@ instance group. Optional, if not specified no rate limiting
//@@ will be applied to this instance group.
//@@
ModelRateLimiter rate_limiter = 6;
//@@ .. cpp:var:: int32 gpus (repeated)
//@@
//@@ GPU(s) where instances should be available. For each GPU listed,
//@@ 'count' instances of the model will be available. Setting 'gpus'
//@@ to empty (or not specifying at all) is eqivalent to listing all
//@@ available GPUs.
//@@
repeated int32 gpus = 3;
//@@ .. cpp:var:: string profile (repeated)
//@@
//@@ For TensorRT models containing multiple optimization profile, this
//@@ parameter specifies a set of optimization profiles available to this
//@@ instance group. The inference server will choose the optimal profile
//@@ based on the shapes of the input tensors. This field should lie
//@@ between 0 and <TotalNumberOfOptimizationProfilesInPlanModel> - 1
//@@ and be specified only for TensorRT backend, otherwise an error will
//@@ be generated. If not specified, the server will select the first
//@@ optimization profile by default.
//@@
repeated string profile = 5;
}
//@@
//@@.. cpp:var:: message ModelTensorReshape
//@@
//@@ Reshape specification for input and output tensors.
//@@
message ModelTensorReshape
{
//@@ .. cpp:var:: int64 shape (repeated)
//@@
//@@ The shape to use for reshaping.
//@@
repeated int64 shape = 1;
}
//@@
//@@.. cpp:var:: message ModelInput
//@@
//@@ An input required by the model.
//@@
message ModelInput
{
//@@
//@@ .. cpp:enum:: Format
//@@
//@@ The format for the input.
//@@
enum Format {
//@@ .. cpp:enumerator:: Format::FORMAT_NONE = 0
//@@
//@@ The input has no specific format. This is the default.
//@@
FORMAT_NONE = 0;
//@@ .. cpp:enumerator:: Format::FORMAT_NHWC = 1
//@@
//@@ HWC image format. Tensors with this format require 3 dimensions
//@@ if the model does not support batching (max_batch_size = 0) or 4
//@@ dimensions if the model does support batching (max_batch_size
//@@ >= 1). In either case the 'dims' below should only specify the
//@@ 3 non-batch dimensions (i.e. HWC or CHW).
//@@
FORMAT_NHWC = 1;
//@@ .. cpp:enumerator:: Format::FORMAT_NCHW = 2
//@@
//@@ CHW image format. Tensors with this format require 3 dimensions
//@@ if the model does not support batching (max_batch_size = 0) or 4
//@@ dimensions if the model does support batching (max_batch_size
//@@ >= 1). In either case the 'dims' below should only specify the
//@@ 3 non-batch dimensions (i.e. HWC or CHW).
//@@
FORMAT_NCHW = 2;
}
//@@ .. cpp:var:: string name
//@@
//@@ The name of the input.
//@@
string name = 1;
//@@ .. cpp:var:: DataType data_type
//@@
//@@ The data-type of the input.
//@@
DataType data_type = 2;
//@@ .. cpp:var:: Format format
//@@
//@@ The format of the input. Optional.
//@@
Format format = 3;
//@@ .. cpp:var:: int64 dims (repeated)
//@@
//@@ The dimensions/shape of the input tensor that must be provided
//@@ when invoking the inference API for this model.
//@@
repeated int64 dims = 4;
//@@ .. cpp:var:: ModelTensorReshape reshape
//@@
//@@ The shape expected for this input by the backend. The input will
//@@ be reshaped to this before being presented to the backend. The
//@@ reshape must have the same number of elements as the input shape
//@@ specified by 'dims'. Optional.
//@@
ModelTensorReshape reshape = 5;
//@@ .. cpp:var:: bool is_shape_tensor
//@@
//@@ Whether or not the input is a shape tensor to the model. This field
//@@ is currently supported only for the TensorRT model. An error will be
//@@ generated if this specification does not comply with underlying
//@@ model.
//@@
bool is_shape_tensor = 6;
//@@ .. cpp:var:: bool allow_ragged_batch
//@@
//@@ Whether or not the input is allowed to be "ragged" in a dynamically
//@@ created batch. Default is false indicating that two requests will
//@@ only be batched if this tensor has the same shape in both requests.
//@@ True indicates that two requests can be batched even if this tensor
//@@ has a different shape in each request. A true value is currently
//@@ supported only for custom models.
//@@
bool allow_ragged_batch = 7;
}
//@@
//@@.. cpp:var:: message ModelOutput
//@@
//@@ An output produced by the model.
//@@
message ModelOutput
{
//@@ .. cpp:var:: string name
//@@
//@@ The name of the output.
//@@
string name = 1;
//@@ .. cpp:var:: DataType data_type
//@@
//@@ The data-type of the output.
//@@
DataType data_type = 2;
//@@ .. cpp:var:: int64 dims (repeated)
//@@
//@@ The dimensions/shape of the output tensor.
//@@
repeated int64 dims = 3;
//@@ .. cpp:var:: ModelTensorReshape reshape
//@@
//@@ The shape produced for this output by the backend. The output will
//@@ be reshaped from this to the shape specifed in 'dims' before being
//@@ returned in the inference response. The reshape must have the same
//@@ number of elements as the output shape specified by 'dims'. Optional.
//@@
ModelTensorReshape reshape = 5;
//@@ .. cpp:var:: string label_filename
//@@
//@@ The label file associated with this output. Should be specified only
//@@ for outputs that represent classifications. Optional.
//@@
string label_filename = 4;
//@@ .. cpp:var:: bool is_shape_tensor
//@@
//@@ Whether or not the output is a shape tensor to the model. This field
//@@ is currently supported only for the TensorRT model. An error will be
//@@ generated if this specification does not comply with underlying
//@@ model.
//@@
bool is_shape_tensor = 6;
}
//@@ .. cpp:var:: message BatchInput
//@@
//@@ A batch input is an additional input that must be added by
//@@ the backend based on all the requests in a batch.
//@@
message BatchInput
{
//@@
//@@ .. cpp:enum:: Kind
//@@
//@@ The kind of the batch input.
//@@
enum Kind {
//@@ .. cpp:enumerator:: Kind::BATCH_ELEMENT_COUNT = 0
//@@
//@@ The element count of the 'source_input' will be added as
//@@ input with shape [1].
//@@
BATCH_ELEMENT_COUNT = 0;
//@@ .. cpp:enumerator:: Kind::BATCH_ACCUMULATED_ELEMENT_COUNT = 1
//@@
//@@ The accumulated element count of the 'source_input' will be
//@@ added as input with shape [1]. For example, if there is a
//@@ batch of two request, each with 2 elements, an input of value
//@@ 2 will be added to the first request, and an input of value
//@@ 4 will be added to the second request.
//@@
BATCH_ACCUMULATED_ELEMENT_COUNT = 1;
//@@ .. cpp:enumerator::
//@@ Kind::BATCH_ACCUMULATED_ELEMENT_COUNT_WITH_ZERO = 2
//@@
//@@ The accumulated element count of the 'source_input' will be
//@@ added as input with shape [1], except for the first request
//@@ in the batch. For the first request in the batch, the input
//@@ will have shape [2] where the first element is value 0.
//@@
BATCH_ACCUMULATED_ELEMENT_COUNT_WITH_ZERO = 2;
//@@ .. cpp:enumerator:: Kind::BATCH_MAX_ELEMENT_COUNT_AS_SHAPE = 3
//@@
//@@ Among the requests in the batch, the max element count of the
//@@ 'source_input' will be added as input with shape
//@@ [max_element_count] for the first request in the batch.
//@@ For other requests, such input will be with shape [0].
//@@ The data of the tensor will be uninitialized.
//@@
BATCH_MAX_ELEMENT_COUNT_AS_SHAPE = 3;
}
//@@ .. cpp:var:: Kind kind
//@@
//@@ The kind of this batch input.
//@@
Kind kind = 1;
//@@ .. cpp:var:: string target_name (repeated)
//@@
//@@ The name of the model inputs that the backend will create
//@@ for this batch input.
//@@
repeated string target_name = 2;
//@@ .. cpp:var:: DataType data_type
//@@
//@@ The input's datatype. The data type can be TYPE_INT32 or
//@@ TYPE_FP32.
//@@
DataType data_type = 3;
//@@ .. cpp:var:: string source_input (repeated)
//@@
//@@ The backend derives the value for each batch input from one or
//@@ more other inputs. 'source_input' gives the names of those
//@@ inputs.
//@@
repeated string source_input = 4;
}
//@@.. cpp:var:: message BatchOutput
//@@
//@@ A batch output is an output produced by the model that must be handled
//@@ differently by the backend based on all the requests in a batch.
//@@
message BatchOutput
{
//@@
//@@ .. cpp:enum:: Kind
//@@
//@@ The kind of the batch output.
//@@
enum Kind {
//@@ .. cpp:enumerator:: Kind::BATCH_SCATTER_WITH_INPUT_SHAPE = 0
//@@
//@@ The output should be scattered according to the shape of
//@@ 'source_input'. The dynamic dimension of the output will
//@@ be set to the value of the same dimension in the input.
//@@
BATCH_SCATTER_WITH_INPUT_SHAPE = 0;
}
//@@ .. cpp:var:: string target_name (repeated)
//@@
//@@ The name of the outputs to be produced by this batch output
//@@ specification.
//@@
repeated string target_name = 1;
//@@ .. cpp:var:: Kind kind
//@@
//@@ The kind of this batch output.
//@@
Kind kind = 2;
//@@ .. cpp:var:: string source_input (repeated)
//@@
//@@ The backend derives each batch output from one or more inputs.
//@@ 'source_input' gives the names of those inputs.
//@@
repeated string source_input = 3;
}
//@@
//@@.. cpp:var:: message ModelVersionPolicy
//@@
//@@ Policy indicating which versions of a model should be made
//@@ available by the inference server.
//@@
message ModelVersionPolicy
{
//@@ .. cpp:var:: message Latest
//@@
//@@ Serve only the latest version(s) of a model. This is
//@@ the default policy.
//@@
message Latest
{
//@@ .. cpp:var:: uint32 num_versions
//@@
//@@ Serve only the 'num_versions' highest-numbered versions. T
//@@ The default value of 'num_versions' is 1, indicating that by
//@@ default only the single highest-number version of a
//@@ model will be served.
//@@
uint32 num_versions = 1;
}
//@@ .. cpp:var:: message All
//@@
//@@ Serve all versions of the model.
//@@
message All {}
//@@ .. cpp:var:: message Specific
//@@
//@@ Serve only specific versions of the model.
//@@
message Specific
{
//@@ .. cpp:var:: int64 versions (repeated)
//@@
//@@ The specific versions of the model that will be served.
//@@
repeated int64 versions = 1;
}
//@@ .. cpp:var:: oneof policy_choice
//@@
//@@ Each model must implement only a single version policy. The
//@@ default policy is 'Latest'.
//@@
oneof policy_choice
{
//@@ .. cpp:var:: Latest latest
//@@
//@@ Serve only latest version(s) of the model.
//@@
Latest latest = 1;
//@@ .. cpp:var:: All all
//@@
//@@ Serve all versions of the model.
//@@
All all = 2;
//@@ .. cpp:var:: Specific specific
//@@
//@@ Serve only specific version(s) of the model.
//@@
Specific specific = 3;
}
}
//@@
//@@.. cpp:var:: message ModelOptimizationPolicy
//@@
//@@ Optimization settings for a model. These settings control if/how a
//@@ model is optimized and prioritized by the backend framework when
//@@ it is loaded.
//@@
message ModelOptimizationPolicy
{
//@@
//@@ .. cpp:var:: message Graph
//@@
//@@ Enable generic graph optimization of the model. If not specified
//@@ the framework's default level of optimization is used. Supports
//@@ TensorFlow graphdef and savedmodel and Onnx models. For TensorFlow
//@@ causes XLA to be enabled/disabled for the model. For Onnx defaults
//@@ to enabling all optimizations, -1 enables only basic optimizations,
//@@ +1 enables only basic and extended optimizations.
//@@
message Graph
{
//@@ .. cpp:var:: int32 level
//@@
//@@ The optimization level. Defaults to 0 (zero) if not specified.
//@@
//@@ - -1: Disabled
//@@ - 0: Framework default
//@@ - 1+: Enable optimization level (greater values indicate
//@@ higher optimization levels)
//@@
int32 level = 1;
}
//@@
//@@ .. cpp:enum:: ModelPriority
//@@
//@@ Model priorities. A model will be given scheduling and execution
//@@ preference over models at lower priorities. Current model
//@@ priorities only work for TensorRT models.
//@@
enum ModelPriority {
//@@ .. cpp:enumerator:: ModelPriority::PRIORITY_DEFAULT = 0
//@@
//@@ The default model priority.
//@@
PRIORITY_DEFAULT = 0;
//@@ .. cpp:enumerator:: ModelPriority::PRIORITY_MAX = 1
//@@
//@@ The maximum model priority.
//@@
PRIORITY_MAX = 1;
//@@ .. cpp:enumerator:: ModelPriority::PRIORITY_MIN = 2
//@@
//@@ The minimum model priority.
//@@
PRIORITY_MIN = 2;
}
//@@
//@@ .. cpp:var:: message Cuda
//@@
//@@ CUDA-specific optimization settings.
//@@
message Cuda
{
//@@ .. cpp:var:: message GraphSpec
//@@
//@@ Specification of the CUDA graph to be captured.
//@@
message GraphSpec
{
//@@ .. cpp:var:: message Dims
//@@
//@@ Specification of tensor dimension.
//@@
message Shape
{
//@@ .. cpp:var:: int64 dim (repeated)
//@@
//@@ The dimension.
//@@
repeated int64 dim = 1;
}
message LowerBound
{
//@@ .. cpp:var:: int32 batch_size
//@@
//@@ The batch size of the CUDA graph. If 'max_batch_size' is 0,
//@@ 'batch_size' must be set to 0. Otherwise, 'batch_size' must
//@@ be set to value between 1 and 'max_batch_size'.
//@@
int32 batch_size = 1;
//@@ .. cpp:var:: map<string, Shape> input
//@@
//@@ The specification of the inputs. 'Shape' is the shape of
//@@ the input without batching dimension.
//@@
map<string, Shape> input = 2;
}
//@@ .. cpp:var:: int32 batch_size
//@@
//@@ The batch size of the CUDA graph. If 'max_batch_size' is 0,
//@@ 'batch_size' must be set to 0. Otherwise, 'batch_size' must
//@@ be set to value between 1 and 'max_batch_size'.
//@@
int32 batch_size = 1;
//@@ .. cpp:var:: map<string, Shape> input
//@@
//@@ The specification of the inputs. 'Shape' is the shape of the
//@@ input without batching dimension.
//@@
map<string, Shape> input = 2;
//@@ .. cpp:var:: LowerBound graph_lower_bound
//@@
//@@ Specify the lower bound of the CUDA graph. Optional.
//@@ If specified, the graph can be used for input shapes and
//@@ batch sizes that are in closed interval between the lower
//@@ bound specification and graph specification. For dynamic
//@@ shape model, this allows CUDA graphs to be launched
//@@ frequently without capturing all possible shape combinations.
//@@ However, using graph for shape combinations different from
//@@ the one used for capturing introduces uninitialized data for
//@@ execution and it may distort the inference result if
//@@ the model is sensitive to uninitialized data.
//@@
LowerBound graph_lower_bound = 3;
}
//@@ .. cpp:var:: bool graphs
//@@
//@@ Use CUDA graphs API to capture model operations and execute
//@@ them more efficiently. Default value is false.
//@@ Currently only recognized by TensorRT backend.
//@@
bool graphs = 1;
//@@ .. cpp:var:: bool busy_wait_events
//@@
//@@ Use busy-waiting to synchronize CUDA events to achieve minimum
//@@ latency from event complete to host thread to be notified, with
//@@ the cost of high CPU load. Default value is false.
//@@ Currently only recognized by TensorRT backend.
//@@
bool busy_wait_events = 2;
//@@ .. cpp:var:: GraphSpec graph_spec (repeated)
//@@
//@@ Specification of the CUDA graph to be captured. If not specified
//@@ and 'graphs' is true, the default CUDA graphs will be captured
//@@ based on model settings.
//@@ Currently only recognized by TensorRT backend.
//@@
repeated GraphSpec graph_spec = 3;
//@@ .. cpp:var:: bool output_copy_stream
//@@
//@@ Uses a CUDA stream separate from the inference stream to copy the
//@@ output to host. Default value is false.
//@@ Currently only recognized by TensorRT backend.
//@@
bool output_copy_stream = 4;
}
//@@
//@@ .. cpp:var:: message ExecutionAccelerators
//@@
//@@ Specify the preferred execution accelerators to be used to execute
//@@ the model. Currently only recognized by ONNX Runtime backend and
//@@ TensorFlow backend.
//@@
//@@ For ONNX Runtime backend, it will deploy the model with the execution
//@@ accelerators by priority, the priority is determined based on the
//@@ order that they are set, i.e. the provider at the front has highest
//@@ priority. Overall, the priority will be in the following order:
//@@ <gpu_execution_accelerator> (if instance is on GPU)
//@@ CUDA Execution Provider (if instance is on GPU)
//@@ <cpu_execution_accelerator>
//@@ Default CPU Execution Provider
//@@
message ExecutionAccelerators
{
//@@
//@@ .. cpp:var:: message Accelerator
//@@
//@@ Specify the accelerator to be used to execute the model.
//@@ Accelerator with the same name may accept different parameters
//@@ depending on the backends.
//@@
message Accelerator
{
//@@ .. cpp:var:: string name
//@@
//@@ The name of the execution accelerator.
//@@
string name = 1;
//@@ .. cpp:var:: map<string, string> parameters
//@@
//@@ Additional paremeters used to configure the accelerator.
//@@
map<string, string> parameters = 2;
}
//@@ .. cpp:var:: Accelerator gpu_execution_accelerator (repeated)
//@@
//@@ The preferred execution provider to be used if the model instance
//@@ is deployed on GPU.
//@@
//@@ For ONNX Runtime backend, possible value is "tensorrt" as name,
//@@ and no parameters are required.
//@@
//@@ For TensorFlow backend, possible values are "tensorrt",
//@@ "auto_mixed_precision", "gpu_io".
//@@
//@@ For "tensorrt", the following parameters can be specified:
//@@ "precision_mode": The precision used for optimization.
//@@ Allowed values are "FP32" and "FP16". Default value is "FP32".
//@@
//@@ "max_cached_engines": The maximum number of cached TensorRT
//@@ engines in dynamic TensorRT ops. Default value is 100.
//@@
//@@ "minimum_segment_size": The smallest model subgraph that will
//@@ be considered for optimization by TensorRT. Default value is 3.
//@@
//@@ "max_workspace_size_bytes": The maximum GPU memory the model
//@@ can use temporarily during execution. Default value is 1GB.
//@@
//@@ For "auto_mixed_precision", no parameters are required. If set,
//@@ the model will try to use FP16 for better performance.
//@@ This optimization can not be set with "tensorrt".
//@@
//@@ For "gpu_io", no parameters are required. If set, the model will
//@@ be executed using TensorFlow Callable API to set input and output
//@@ tensors in GPU memory if possible, which can reduce data transfer
//@@ overhead if the model is used in ensemble. However, the Callable
//@@ object will be created on model creation and it will request all
//@@ outputs for every model execution, which may impact the
//@@ performance if a request does not require all outputs. This
//@@ optimization will only take affect if the model instance is
//@@ created with KIND_GPU.
//@@
repeated Accelerator gpu_execution_accelerator = 1;
//@@ .. cpp:var:: Accelerator cpu_execution_accelerator (repeated)
//@@
//@@ The preferred execution provider to be used if the model instance
//@@ is deployed on CPU.
//@@
//@@ For ONNX Runtime backend, possible value is "openvino" as name,
//@@ and no parameters are required.
//@@
repeated Accelerator cpu_execution_accelerator = 2;
}
//@@
//@@ .. cpp:var:: message PinnedMemoryBuffer
//@@
//@@ Specify whether to use a pinned memory buffer when transferring data
//@@ between non-pinned system memory and GPU memory. Using a pinned
//@@ memory buffer for system from/to GPU transfers will typically provide
//@@ increased performance. For example, in the common use case where the
//@@ request provides inputs and delivers outputs via non-pinned system
//@@ memory, if the model instance accepts GPU IOs, the inputs will be
//@@ processed by two copies: from non-pinned system memory to pinned
//@@ memory, and from pinned memory to GPU memory. Similarly, pinned
//@@ memory will be used for delivering the outputs.
//@@
message PinnedMemoryBuffer
{
//@@ .. cpp:var:: bool enable
//@@
//@@ Use pinned memory buffer. Default is true.
//@@
bool enable = 1;
}
//@@ .. cpp:var:: Graph graph
//@@
//@@ The graph optimization setting for the model. Optional.
//@@
Graph graph = 1;
//@@ .. cpp:var:: ModelPriority priority
//@@
//@@ The priority setting for the model. Optional.
//@@
ModelPriority priority = 2;
//@@ .. cpp:var:: Cuda cuda
//@@
//@@ CUDA-specific optimization settings. Optional.
//@@
Cuda cuda = 3;
//@@ .. cpp:var:: ExecutionAccelerators execution_accelerators
//@@
//@@ The accelerators used for the model. Optional.
//@@
ExecutionAccelerators execution_accelerators = 4;
//@@ .. cpp:var:: PinnedMemoryBuffer input_pinned_memory
//@@
//@@ Use pinned memory buffer when the data transfer for inputs
//@@ is between GPU memory and non-pinned system memory.
//@@ Default is true.
//@@
PinnedMemoryBuffer input_pinned_memory = 5;
//@@ .. cpp:var:: PinnedMemoryBuffer output_pinned_memory
//@@
//@@ Use pinned memory buffer when the data transfer for outputs
//@@ is between GPU memory and non-pinned system memory.
//@@ Default is true.
//@@
PinnedMemoryBuffer output_pinned_memory = 6;
//@@ .. cpp:var:: uint32 gather_kernel_buffer_threshold
//@@
//@@ The backend may use a gather kernel to gather input data if the
//@@ device has direct access to the source buffer and the destination
//@@ buffer. In such case, the gather kernel will be used only if the
//@@ number of buffers to be gathered is greater or equal to
//@@ the specifed value. If 0, the gather kernel will be disabled.
//@@ Default value is 0.
//@@ Currently only recognized by TensorRT backend.
//@@
uint32 gather_kernel_buffer_threshold = 7;
//@@ .. cpp:var:: bool eager_batching
//@@
//@@ Start preparing the next batch before the model instance is ready
//@@ for the next inference. This option can be used to overlap the
//@@ batch preparation with model execution, with the trade-off that
//@@ the next batch might be smaller than what it could have been.
//@@ Default value is false.
//@@ Currently only recognized by TensorRT backend.
//@@
bool eager_batching = 8;
}
//@@
//@@.. cpp:var:: message ModelQueuePolicy
//@@
//@@ Queue policy for inference requests.
//@@
message ModelQueuePolicy
{
//@@
//@@ .. cpp:enum:: TimeoutAction
//@@
//@@ The action applied to timed-out requests.
//@@
enum TimeoutAction {
//@@ .. cpp:enumerator:: Action::REJECT = 0
//@@
//@@ Reject the request and return error message accordingly.
//@@
REJECT = 0;
//@@ .. cpp:enumerator:: Action::DELAY = 1
//@@
//@@ Delay the request until all other requests at the same
//@@ (or higher) priority levels that have not reached their timeouts
//@@ are processed. A delayed request will eventually be processed,
//@@ but may be delayed indefinitely due to newly arriving requests.
//@@
DELAY = 1;
}
//@@
//@@ .. cpp:var:: TimeoutAction timeout_action
//@@
//@@ The action applied to timed-out request.
//@@ The default action is REJECT.
//@@
TimeoutAction timeout_action = 1;
//@@
//@@ .. cpp:var:: uint64 default_timeout_microseconds
//@@
//@@ The default timeout for every request, in microseconds.
//@@ The default value is 0 which indicates that no timeout is set.
//@@
uint64 default_timeout_microseconds = 2;
//@@
//@@ .. cpp:var:: bool allow_timeout_override
//@@
//@@ Whether individual request can override the default timeout value.
//@@ When true, individual requests can set a timeout that is less than
//@@ the default timeout value but may not increase the timeout.
//@@ The default value is false.
//@@
bool allow_timeout_override = 3;
//@@
//@@ .. cpp:var:: uint32 max_queue_size
//@@
//@@ The maximum queue size for holding requests. A request will be
//@@ rejected immediately if it can't be enqueued because the queue is
//@@ full. The default value is 0 which indicates that no maximum
//@@ queue size is enforced.