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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
|
// RUN: mlir-opt -split-input-file -verify-diagnostics %s | FileCheck %s
// CHECK-LABEL: func @depthwise_conv_1d_nwc_wcm
func.func @depthwise_conv_1d_nwc_wcm(%input: tensor<1x12x8xf32>, %filter: tensor<3x8x8xf32>) -> tensor<1x10x8x8xf32> {
%zero = arith.constant 0.000000e+00 : f32
%init = tensor.empty() : tensor<1x10x8x8xf32>
%fill = linalg.fill ins(%zero : f32) outs(%init : tensor<1x10x8x8xf32>) -> tensor<1x10x8x8xf32>
// CHECK: depthwise_conv_1d_nwc_wcm
%0 = linalg.depthwise_conv_1d_nwc_wcm {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %filter : tensor<1x12x8xf32>, tensor<3x8x8xf32>)
outs(%fill : tensor<1x10x8x8xf32>) -> tensor<1x10x8x8xf32>
return %0 : tensor<1x10x8x8xf32>
}
// -----
// CHECK-LABEL: func @depthwise_conv_1d_nwc_wc
func.func @depthwise_conv_1d_nwc_wc(%input: tensor<1x12x8xf32>, %filter: tensor<3x8xf32>) -> tensor<1x10x8xf32> {
%zero = arith.constant 0.000000e+00 : f32
%init = tensor.empty() : tensor<1x10x8xf32>
%fill = linalg.fill ins(%zero : f32) outs(%init : tensor<1x10x8xf32>) -> tensor<1x10x8xf32>
// CHECK: depthwise_conv_1d_nwc_wc
%0 = linalg.depthwise_conv_1d_nwc_wc {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %filter : tensor<1x12x8xf32>, tensor<3x8xf32>)
outs(%fill : tensor<1x10x8xf32>) -> tensor<1x10x8xf32>
return %0 : tensor<1x10x8xf32>
}
// -----
// CHECK-LABEL: func @depthwise_conv_2d_nhwc_hwcm_tensor
func.func @depthwise_conv_2d_nhwc_hwcm_tensor(%input: tensor<2x4x5x2xf32>, %filter: tensor<2x2x2x3xf32>) -> tensor<2x3x4x2x3xf32> {
%zero = arith.constant 0.000000e+00 : f32
%init = tensor.empty() : tensor<2x3x4x2x3xf32>
%fill = linalg.fill ins(%zero : f32) outs(%init : tensor<2x3x4x2x3xf32>) -> tensor<2x3x4x2x3xf32>
// CHECK: %{{.+}} = linalg.depthwise_conv_2d_nhwc_hwcm
// CHECK-SAME: {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<2x4x5x2xf32>, tensor<2x2x2x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<2x3x4x2x3xf32>)
%0 = linalg.depthwise_conv_2d_nhwc_hwcm
{ dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64> }
ins(%input, %filter : tensor<2x4x5x2xf32>, tensor<2x2x2x3xf32>)
outs(%fill : tensor<2x3x4x2x3xf32>) -> tensor<2x3x4x2x3xf32>
return %0 : tensor<2x3x4x2x3xf32>
}
// CHECK-LABEL: func @depthwise_conv_2d_nhwc_hwcm_memref
func.func @depthwise_conv_2d_nhwc_hwcm_memref(%input: memref<2x4x5x2xf32>, %filter: memref<2x2x2x3xf32>, %output: memref<2x3x4x2x3xf32>) {
// CHECK: linalg.depthwise_conv_2d_nhwc_hwcm
// CHECK-SAME: {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<2x4x5x2xf32>, memref<2x2x2x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<2x3x4x2x3xf32>)
linalg.depthwise_conv_2d_nhwc_hwcm
{ dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64> }
ins(%input, %filter : memref<2x4x5x2xf32>, memref<2x2x2x3xf32>)
outs(%output : memref<2x3x4x2x3xf32>)
return
}
// CHECK-LABEL: func @depthwise_conv_1d_nw_tensor
func.func @depthwise_conv_1d_nw_tensor(%input: tensor<1x113x96xf32>, %filter: tensor<3x96xf32>) -> tensor<1x56x96xf32> {
%init = tensor.empty() : tensor<1x56x96xf32>
// CHECK: %{{.+}} = linalg.depthwise_conv_1d_nw
// CHECK-SAME: {dilations = dense<1> : vector<1xi64>, strides = dense<2> : vector<1xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x113x96xf32>, tensor<3x96xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x56x96xf32>) -> tensor<1x56x96xf32>
%0 = linalg.depthwise_conv_1d_nwc_wc {dilations = dense<1> : vector<1xi64>, strides = dense<2> : vector<1xi64>}
ins(%input, %filter: tensor<1x113x96xf32>, tensor<3x96xf32>)
outs(%init: tensor<1x56x96xf32>) -> tensor<1x56x96xf32>
return %0: tensor<1x56x96xf32>
}
// CHECK-LABEL: func @depthwise_conv_2d_nhwc_hwc_tensor
func.func @depthwise_conv_2d_nhwc_hwc_tensor(%input: tensor<1x113x113x96xf32>, %filter: tensor<3x3x96xf32>) -> tensor<1x56x56x96xf32> {
%init = tensor.empty() : tensor<1x56x56x96xf32>
// CHECK: %{{.+}} = linalg.depthwise_conv_2d_nhwc_hwc
// CHECK-SAME: {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<2xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x113x113x96xf32>, tensor<3x3x96xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x56x56x96xf32>) -> tensor<1x56x56x96xf32>
%0 = linalg.depthwise_conv_2d_nhwc_hwc {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<2xi64>}
ins(%input, %filter: tensor<1x113x113x96xf32>, tensor<3x3x96xf32>)
outs(%init: tensor<1x56x56x96xf32>) -> tensor<1x56x56x96xf32>
return %0: tensor<1x56x56x96xf32>
}
// CHECK-LABEL: func @depthwise_conv_2d_nhwc_hwc_memref
func.func @depthwise_conv_2d_nhwc_hwc_memref(%input: memref<1x113x113x96xf32>, %filter: memref<3x3x96xf32>, %output: memref<1x56x56x96xf32>) {
// CHECK: linalg.depthwise_conv_2d_nhwc_hwc
// CHECK-SAME: {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<2xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x113x113x96xf32>, memref<3x3x96xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x56x56x96xf32>)
linalg.depthwise_conv_2d_nhwc_hwc {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<2xi64>}
ins(%input, %filter: memref<1x113x113x96xf32>, memref<3x3x96xf32>)
outs(%output: memref<1x56x56x96xf32>)
return
}
// CHECK-LABEL: func @depthwise_conv_2d_nchw_chw_tensor
func.func @depthwise_conv_2d_nchw_chw_tensor(%input: tensor<1x96x113x113xf32>, %filter: tensor<96x3x3xf32>) -> tensor<1x96x56x56xf32> {
%init = tensor.empty() : tensor<1x96x56x56xf32>
// CHECK: %{{.+}} = linalg.depthwise_conv_2d_nchw_chw
// CHECK-SAME: {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<2xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x96x113x113xf32>, tensor<96x3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x96x56x56xf32>) -> tensor<1x96x56x56xf32>
%0 = linalg.depthwise_conv_2d_nchw_chw {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<2xi64>}
ins(%input, %filter: tensor<1x96x113x113xf32>, tensor<96x3x3xf32>)
outs(%init: tensor<1x96x56x56xf32>) -> tensor<1x96x56x56xf32>
return %0: tensor<1x96x56x56xf32>
}
// CHECK-LABEL: func @depthwise_conv_2d_nchw_chw_memref
func.func @depthwise_conv_2d_nchw_chw_memref(%input: memref<1x96x113x113xf32>, %filter: memref<96x3x3xf32>, %output: memref<1x96x56x56xf32>) {
// CHECK: linalg.depthwise_conv_2d_nchw_chw
// CHECK-SAME: {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<2xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x96x113x113xf32>, memref<96x3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x96x56x56xf32>)
linalg.depthwise_conv_2d_nchw_chw {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<2xi64>}
ins(%input, %filter: memref<1x96x113x113xf32>, memref<96x3x3xf32>)
outs(%output: memref<1x96x56x56xf32>)
return
}
func.func @depthwise_conv_2d_nhwc_hwcm_tensor_dilated(%input: tensor<2x8x9x2xf32>, %filter: tensor<2x2x2x3xf32>) -> tensor<2x6x7x2x3xf32> {
%zero = arith.constant 0.000000e+00 : f32
%init = tensor.empty() : tensor<2x6x7x2x3xf32>
%fill = linalg.fill ins(%zero : f32) outs(%init : tensor<2x6x7x2x3xf32>) -> tensor<2x6x7x2x3xf32>
// CHECK: %{{.+}} = linalg.depthwise_conv_2d_nhwc_hwcm
// CHECK-SAME: {dilations = dense<2> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<2x8x9x2xf32>, tensor<2x2x2x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<2x6x7x2x3xf32>)
%0 = linalg.depthwise_conv_2d_nhwc_hwcm
{ dilations = dense<2> : tensor<2xi64>, strides = dense<1> : tensor<2xi64> }
ins(%input, %filter : tensor<2x8x9x2xf32>, tensor<2x2x2x3xf32>)
outs(%fill : tensor<2x6x7x2x3xf32>) -> tensor<2x6x7x2x3xf32>
return %0 : tensor<2x6x7x2x3xf32>
}
// CHECK-LABEL: func @depthwise_conv_2d_nhwc_hwcm_memref_dilated
func.func @depthwise_conv_2d_nhwc_hwcm_memref_dilated(%input: memref<2x8x9x2xf32>, %filter: memref<2x2x2x3xf32>, %output: memref<2x6x7x2x3xf32>) {
// CHECK: linalg.depthwise_conv_2d_nhwc_hwcm
// CHECK-SAME: {dilations = dense<2> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<2x8x9x2xf32>, memref<2x2x2x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<2x6x7x2x3xf32>)
linalg.depthwise_conv_2d_nhwc_hwcm
{ dilations = dense<2> : tensor<2xi64>, strides = dense<1> : tensor<2xi64> }
ins(%input, %filter : memref<2x8x9x2xf32>, memref<2x2x2x3xf32>)
outs(%output : memref<2x6x7x2x3xf32>)
return
}
// -----
// CHECK-LABEL: func @depthwise_conv_2d_input_nhwc_filter_default_attributes
func.func @depthwise_conv_2d_input_nhwc_filter_default_attributes(%input: memref<1x113x113x96xf32>, %filter: memref<3x3x96xf32>, %output: memref<1x56x56x96xf32>) {
// CHECK: linalg.depthwise_conv_2d_nhwc_hwc
// CHECK-NOT: strides =
// CHECK-NOT: dilations =
linalg.depthwise_conv_2d_nhwc_hwc
ins(%input, %filter: memref<1x113x113x96xf32>, memref<3x3x96xf32>)
outs(%output: memref<1x56x56x96xf32>)
return
}
// -----
func.func @depthwise_conv_2d_input_nhwc_filter_wrong_stride_element_type(%input: memref<1x113x113x96xf32>, %filter: memref<3x3x96xf32>, %output: memref<1x56x56x96xf32>) {
// expected-error @+1 {{op attribute 'strides' failed to satisfy constraint: 64-bit signless int elements attribute of shape [2]}}
linalg.depthwise_conv_2d_nhwc_hwc {dilations = dense<1> : vector<2xi64>, strides = dense<2.0> : vector<2xf32>}
ins(%input, %filter: memref<1x113x113x96xf32>, memref<3x3x96xf32>)
outs(%output: memref<1x56x56x96xf32>)
return
}
// -----
func.func @depthwise_conv_2d_input_nhwc_filter_wrong_stride_size(%input: memref<1x113x113x96xf32>, %filter: memref<3x3x96xf32>, %output: memref<1x56x56x96xf32>) {
// expected-error @+1 {{op attribute 'strides' failed to satisfy constraint: 64-bit signless int elements attribute of shape [2]}}
linalg.depthwise_conv_2d_nhwc_hwc {dilations = dense<1> : vector<2xi64>, strides = dense<2> : vector<3xi64> }
ins(%input, %filter: memref<1x113x113x96xf32>, memref<3x3x96xf32>)
outs(%output: memref<1x56x56x96xf32>)
return
}
// -----
// CHECK-LABEL: func @depthwise_conv_3d_ndhwc_dhwcm
func.func @depthwise_conv_3d_ndhwc_dhwcm(%input: tensor<2x6x13x12x6xf32>, %filter: tensor<2x1x3x6x6xf32>) -> tensor<2x3x13x4x6x6xf32> {
%zero = arith.constant 0.000000e+00 : f32
%init = tensor.empty() : tensor<2x3x13x4x6x6xf32>
%fill = linalg.fill ins(%zero : f32) outs(%init : tensor<2x3x13x4x6x6xf32>) -> tensor<2x3x13x4x6x6xf32>
// CHECK: depthwise_conv_3d_ndhwc_dhwcm
%0 = linalg.depthwise_conv_3d_ndhwc_dhwcm {dilations = dense<1> : tensor<3xi64>, strides = dense<[2, 1, 3]> : tensor<3xi64>}
ins(%input, %filter : tensor<2x6x13x12x6xf32>, tensor<2x1x3x6x6xf32>)
outs(%fill : tensor<2x3x13x4x6x6xf32>) -> tensor<2x3x13x4x6x6xf32>
return %0 : tensor<2x3x13x4x6x6xf32>
}
// -----
// CHECK-LABEL: func @depthwise_conv_3d_ndhwc_dhwc
func.func @depthwise_conv_3d_ndhwc_dhwc(%input: tensor<2x6x13x12x6xf32>, %filter: tensor<2x1x3x6xf32>) -> tensor<2x3x13x4x6xf32> {
%zero = arith.constant 0.000000e+00 : f32
%init = tensor.empty() : tensor<2x3x13x4x6xf32>
%fill = linalg.fill ins(%zero : f32) outs(%init : tensor<2x3x13x4x6xf32>) -> tensor<2x3x13x4x6xf32>
// CHECK: depthwise_conv_3d_ndhwc_dhwc
%0 = linalg.depthwise_conv_3d_ndhwc_dhwc {dilations = dense<1> : tensor<3xi64>, strides = dense<[2, 1, 3]> : tensor<3xi64>}
ins(%input, %filter : tensor<2x6x13x12x6xf32>, tensor<2x1x3x6xf32>)
outs(%fill : tensor<2x3x13x4x6xf32>) -> tensor<2x3x13x4x6xf32>
return %0 : tensor<2x3x13x4x6xf32>
}
// -----
// CHECK-LABEL: func @conv_1d_nwc_wcf
func.func @conv_1d_nwc_wcf(%input: tensor<?x?x?xf32>, %filter: tensor<?x?x?xf32>, %init: tensor<?x?x?xf32>) -> tensor<?x?x?xf32> {
// CHECK: %{{.+}} = linalg.conv_1d_nwc_wcf
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<?x?x?xf32>, tensor<?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
%0 = linalg.conv_1d_nwc_wcf {dilations = dense<1> : tensor<1xi64>,
strides = dense<1> : tensor<1xi64>}
ins (%input, %filter: tensor<?x?x?xf32>, tensor<?x?x?xf32>)
outs (%init: tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
return %0 : tensor<?x?x?xf32>
}
// -----
// CHECK-LABEL: func @conv_1d_nwc_wcf
func.func @conv_1d_nwc_wcf(%input: memref<?x?x?xf32>, %filter: memref<?x?x?xf32>, %output: memref<?x?x?xf32>) {
// CHECK: linalg.conv_1d_nwc_wcf
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<?x?x?xf32>, memref<?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<?x?x?xf32>)
linalg.conv_1d_nwc_wcf {dilations = dense<1> : tensor<1xi64>,
strides = dense<1> : tensor<1xi64>}
ins (%input, %filter: memref<?x?x?xf32>, memref<?x?x?xf32>)
outs (%output: memref<?x?x?xf32>)
return
}
// -----
// CHECK-LABEL: func @conv_1d_ncw_fcw
func.func @conv_1d_ncw_fcw(%input: tensor<?x?x?xf32>, %filter: tensor<?x?x?xf32>, %init: tensor<?x?x?xf32>) -> tensor<?x?x?xf32> {
// CHECK: %{{.+}} = linalg.conv_1d_ncw_fcw
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<?x?x?xf32>, tensor<?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
%0 = linalg.conv_1d_ncw_fcw {dilations = dense<1> : tensor<1xi64>,
strides = dense<1> : tensor<1xi64>}
ins (%input, %filter: tensor<?x?x?xf32>, tensor<?x?x?xf32>)
outs (%init: tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
return %0 : tensor<?x?x?xf32>
}
// -----
// CHECK-LABEL: func @conv_1d_ncw_fcw
func.func @conv_1d_ncw_fcw(%input: memref<?x?x?xf32>, %filter: memref<?x?x?xf32>, %output: memref<?x?x?xf32>) {
// CHECK: linalg.conv_1d_ncw_fcw
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<?x?x?xf32>, memref<?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<?x?x?xf32>)
linalg.conv_1d_ncw_fcw {dilations = dense<1> : tensor<1xi64>,
strides = dense<1> : tensor<1xi64>}
ins (%input, %filter: memref<?x?x?xf32>, memref<?x?x?xf32>)
outs (%output: memref<?x?x?xf32>)
return
}
// -----
// CHECK-LABEL: func @conv_2d_nhwc_hwcf
func.func @conv_2d_nhwc_hwcf(%input: tensor<?x?x?x?xf32>, %filter: tensor<?x?x?x?xf32>, %init: tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32> {
// CHECK: %{{.+}} = linalg.conv_2d_nhwc_hwcf
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<?x?x?x?xf32>, tensor<?x?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
%0 = linalg.conv_2d_nhwc_hwcf {dilations = dense<1> : tensor<2xi64>,
strides = dense<1> : tensor<2xi64>}
ins (%input, %filter: tensor<?x?x?x?xf32>, tensor<?x?x?x?xf32>)
outs (%init: tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
return %0 : tensor<?x?x?x?xf32>
}
// -----
// CHECK-LABEL: func @conv_2d_ngchw_fgchw
func.func @conv_2d_ngchw_fgchw(%input: tensor<?x?x?x?x?xf32>, %filter: tensor<?x?x?x?x?xf32>, %init: tensor<?x?x?x?x?xf32>) -> tensor<?x?x?x?x?xf32> {
// CHECK: %{{.+}} = linalg.conv_2d_ngchw_fgchw
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<?x?x?x?x?xf32>, tensor<?x?x?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<?x?x?x?x?xf32>) -> tensor<?x?x?x?x?xf32>
%0 = linalg.conv_2d_ngchw_fgchw {dilations = dense<1> : tensor<2xi64>,
strides = dense<1> : tensor<2xi64>}
ins (%input, %filter: tensor<?x?x?x?x?xf32>, tensor<?x?x?x?x?xf32>)
outs (%init: tensor<?x?x?x?x?xf32>) -> tensor<?x?x?x?x?xf32>
return %0 : tensor<?x?x?x?x?xf32>
}
// -----
// CHECK-LABEL: func @conv_2d_nhwc_fhwc
func.func @conv_2d_nhwc_fhwc(%input: tensor<?x?x?x?xf32>, %filter: tensor<?x?x?x?xf32>, %init: tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32> {
// CHECK: %{{.+}} = linalg.conv_2d_nhwc_fhwc
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<?x?x?x?xf32>, tensor<?x?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
%0 = linalg.conv_2d_nhwc_fhwc {dilations = dense<1> : tensor<2xi64>,
strides = dense<1> : tensor<2xi64>}
ins (%input, %filter: tensor<?x?x?x?xf32>, tensor<?x?x?x?xf32>)
outs (%init: tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
return %0 : tensor<?x?x?x?xf32>
}
// -----
// CHECK-LABEL: func @conv_2d_nhwc_fhwc_static
func.func @conv_2d_nhwc_fhwc_static(%input: tensor<?x128x128x32xf32>, %filter: tensor<64x3x3x32xf32>, %init: tensor<?x126x126x64xf32>) -> tensor<?x126x126x64xf32> {
// CHECK: %{{.+}} = linalg.conv_2d_nhwc_fhwc
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<?x128x128x32xf32>, tensor<64x3x3x32xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<?x126x126x64xf32>) -> tensor<?x126x126x64xf32>
%0 = linalg.conv_2d_nhwc_fhwc {dilations = dense<1> : tensor<2xi64>,
strides = dense<1> : tensor<2xi64>}
ins (%input, %filter: tensor<?x128x128x32xf32>, tensor<64x3x3x32xf32>)
outs (%init: tensor<?x126x126x64xf32>) -> tensor<?x126x126x64xf32>
return %0 : tensor<?x126x126x64xf32>
}
// -----
// CHECK-LABEL: func @conv_2d_nhwc_hwcf
func.func @conv_2d_nhwc_hwcf(%input: memref<?x?x?x?xf32>, %filter: memref<?x?x?x?xf32>, %output: memref<?x?x?x?xf32>) {
// CHECK: linalg.conv_2d_nhwc_hwcf
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<?x?x?x?xf32>, memref<?x?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<?x?x?x?xf32>)
linalg.conv_2d_nhwc_hwcf {dilations = dense<1> : tensor<2xi64>,
strides = dense<1> : tensor<2xi64>}
ins (%input, %filter: memref<?x?x?x?xf32>, memref<?x?x?x?xf32>)
outs (%output: memref<?x?x?x?xf32>)
return
}
// -----
// CHECK-LABEL: func @conv_2d_ngchw_fgchw
func.func @conv_2d_ngchw_fgchw(%input: memref<?x?x?x?x?xf32>, %filter: memref<?x?x?x?x?xf32>, %output: memref<?x?x?x?x?xf32>) {
// CHECK: linalg.conv_2d_ngchw_fgchw
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<?x?x?x?x?xf32>, memref<?x?x?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<?x?x?x?x?xf32>)
linalg.conv_2d_ngchw_fgchw {dilations = dense<1> : tensor<2xi64>,
strides = dense<1> : tensor<2xi64>}
ins (%input, %filter: memref<?x?x?x?x?xf32>, memref<?x?x?x?x?xf32>)
outs (%output: memref<?x?x?x?x?xf32>)
return
}
// -----
// CHECK-LABEL: func @conv_3d_ndhwc_dhwcf
func.func @conv_3d_ndhwc_dhwcf(%input: tensor<?x?x?x?x?xf32>, %filter: tensor<?x?x?x?x?xf32>, %init: tensor<?x?x?x?x?xf32>) -> tensor<?x?x?x?x?xf32> {
// CHECK: %{{.+}} = linalg.conv_3d_ndhwc_dhwcf
// CHECK-SAME: dilations = dense<1> : tensor<3xi64>
// CHECK-SAME: strides = dense<1> : tensor<3xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<?x?x?x?x?xf32>, tensor<?x?x?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<?x?x?x?x?xf32>) -> tensor<?x?x?x?x?xf32>
%0 = linalg.conv_3d_ndhwc_dhwcf {dilations = dense<1> : tensor<3xi64>,
strides = dense<1> : tensor<3xi64>}
ins (%input, %filter: tensor<?x?x?x?x?xf32>, tensor<?x?x?x?x?xf32>)
outs (%init: tensor<?x?x?x?x?xf32>) -> tensor<?x?x?x?x?xf32>
return %0 : tensor<?x?x?x?x?xf32>
}
// -----
// CHECK-LABEL: func @conv_3d_ndhwc_dhwcf
func.func @conv_3d_ndhwc_dhwcf(%input: memref<?x?x?x?x?xf32>, %filter: memref<?x?x?x?x?xf32>, %output: memref<?x?x?x?x?xf32>) {
// CHECK: linalg.conv_3d_ndhwc_dhwcf
// CHECK-SAME: dilations = dense<1> : tensor<3xi64>
// CHECK-SAME: strides = dense<1> : tensor<3xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<?x?x?x?x?xf32>, memref<?x?x?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<?x?x?x?x?xf32>)
linalg.conv_3d_ndhwc_dhwcf {dilations = dense<1> : tensor<3xi64>,
strides = dense<1> : tensor<3xi64>}
ins (%input, %filter: memref<?x?x?x?x?xf32>, memref<?x?x?x?x?xf32>)
outs (%output: memref<?x?x?x?x?xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nhwc_sum_tensor
// CHECK: %{{.+}} = linalg.pooling_nhwc_sum
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x1xf32>, tensor<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
func.func @pooling_nhwc_sum_tensor(%input: tensor<1x4x4x1xf32>) -> tensor<1x2x2x1xf32> {
%fake = tensor.empty() : tensor<3x3xf32>
%init = tensor.empty() : tensor<1x2x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
%res = linalg.pooling_nhwc_sum {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: tensor<1x4x4x1xf32>, tensor<3x3xf32>)
outs(%fill: tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
return %res : tensor<1x2x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_nwc_sum_tensor
// CHECK: %{{.+}} = linalg.pooling_nwc_sum
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x1xf32>, tensor<3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
func.func @pooling_nwc_sum_tensor(%input: tensor<1x4x1xf32>) -> tensor<1x2x1xf32> {
%fake = tensor.empty() : tensor<3xf32>
%init = tensor.empty() : tensor<1x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
%res = linalg.pooling_nwc_sum {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: tensor<1x4x1xf32>, tensor<3xf32>)
outs(%fill: tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
return %res : tensor<1x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_nhwc_sum
// CHECK: linalg.pooling_nhwc_sum
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x1xf32>, memref<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x1xf32>)
func.func @pooling_nhwc_sum(%input: memref<1x4x4x1xf32>, %fake: memref<3x3xf32>, %output: memref<1x2x2x1xf32>) {
linalg.pooling_nhwc_sum {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: memref<1x4x4x1xf32>, memref<3x3xf32>)
outs(%output: memref<1x2x2x1xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nwc_sum
// CHECK: linalg.pooling_nwc_sum
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x1xf32>, memref<3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x1xf32>)
func.func @pooling_nwc_sum(%input: memref<1x4x1xf32>, %fake: memref<3xf32>, %output: memref<1x2x1xf32>) {
linalg.pooling_nwc_sum {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: memref<1x4x1xf32>, memref<3xf32>)
outs(%output: memref<1x2x1xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nchw_sum_tensor
// CHECK: %{{.+}} = linalg.pooling_nchw_sum
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x1x4x4xf32>, tensor<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x1x2x2xf32>) -> tensor<1x1x2x2xf32>
func.func @pooling_nchw_sum_tensor(%input: tensor<1x1x4x4xf32>) -> tensor<1x1x2x2xf32> {
%fake = tensor.empty() : tensor<3x3xf32>
%init = tensor.empty() : tensor<1x1x2x2xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x1x2x2xf32>) -> tensor<1x1x2x2xf32>
%res = linalg.pooling_nchw_sum {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: tensor<1x1x4x4xf32>, tensor<3x3xf32>)
outs(%fill: tensor<1x1x2x2xf32>) -> tensor<1x1x2x2xf32>
return %res : tensor<1x1x2x2xf32>
}
// -----
// CHECK-LABEL: func @pooling_ncw_sum_tensor
// CHECK: %{{.+}} = linalg.pooling_ncw_sum
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x1x4xf32>, tensor<3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x1x2xf32>) -> tensor<1x1x2xf32>
func.func @pooling_ncw_sum_tensor(%input: tensor<1x1x4xf32>) -> tensor<1x1x2xf32> {
%fake = tensor.empty() : tensor<3xf32>
%init = tensor.empty() : tensor<1x1x2xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x1x2xf32>) -> tensor<1x1x2xf32>
%res = linalg.pooling_ncw_sum {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: tensor<1x1x4xf32>, tensor<3xf32>)
outs(%fill: tensor<1x1x2xf32>) -> tensor<1x1x2xf32>
return %res : tensor<1x1x2xf32>
}
// -----
// CHECK-LABEL: func @pooling_nchw_sum
// CHECK: linalg.pooling_nchw_sum
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x1x4x4xf32>, memref<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x1x2x2xf32>)
func.func @pooling_nchw_sum(%input: memref<1x1x4x4xf32>, %fake: memref<3x3xf32>, %output: memref<1x1x2x2xf32>) {
linalg.pooling_nchw_sum {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: memref<1x1x4x4xf32>, memref<3x3xf32>)
outs(%output: memref<1x1x2x2xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_ncw_sum
// CHECK: linalg.pooling_ncw_sum
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x1x4xf32>, memref<3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x1x2xf32>)
func.func @pooling_ncw_sum(%input: memref<1x1x4xf32>, %fake: memref<3xf32>, %output: memref<1x1x2xf32>) {
linalg.pooling_ncw_sum {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: memref<1x1x4xf32>, memref<3xf32>)
outs(%output: memref<1x1x2xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nhwc_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x1xf32>, tensor<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
func.func @pooling_nhwc_max_tensor(%input: tensor<1x4x4x1xf32>) -> tensor<1x2x2x1xf32> {
%fake = tensor.empty() : tensor<3x3xf32>
%init = tensor.empty() : tensor<1x2x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
%res = linalg.pooling_nhwc_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: tensor<1x4x4x1xf32>, tensor<3x3xf32>)
outs(%fill: tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
return %res : tensor<1x2x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_nwc_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nwc_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x1xf32>, tensor<3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
func.func @pooling_nwc_max_tensor(%input: tensor<1x4x1xf32>) -> tensor<1x2x1xf32> {
%fake = tensor.empty() : tensor<3xf32>
%init = tensor.empty() : tensor<1x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
%res = linalg.pooling_nwc_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: tensor<1x4x1xf32>, tensor<3xf32>)
outs(%fill: tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
return %res : tensor<1x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_nchw_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nchw_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x1x4x4xf32>, tensor<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x1x2x2xf32>) -> tensor<1x1x2x2xf32>
func.func @pooling_nchw_max_tensor(%input: tensor<1x1x4x4xf32>) -> tensor<1x1x2x2xf32> {
%fake = tensor.empty() : tensor<3x3xf32>
%init = tensor.empty() : tensor<1x1x2x2xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x1x2x2xf32>) -> tensor<1x1x2x2xf32>
%res = linalg.pooling_nchw_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: tensor<1x1x4x4xf32>, tensor<3x3xf32>)
outs(%fill: tensor<1x1x2x2xf32>) -> tensor<1x1x2x2xf32>
return %res : tensor<1x1x2x2xf32>
}
// -----
// CHECK-LABEL: func @pooling_ncw_max_tensor
// CHECK: %{{.+}} = linalg.pooling_ncw_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x1x4xf32>, tensor<3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x1x2xf32>) -> tensor<1x1x2xf32>
func.func @pooling_ncw_max_tensor(%input: tensor<1x1x4xf32>) -> tensor<1x1x2xf32> {
%fake = tensor.empty() : tensor<3xf32>
%init = tensor.empty() : tensor<1x1x2xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x1x2xf32>) -> tensor<1x1x2xf32>
%res = linalg.pooling_ncw_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: tensor<1x1x4xf32>, tensor<3xf32>)
outs(%fill: tensor<1x1x2xf32>) -> tensor<1x1x2xf32>
return %res : tensor<1x1x2xf32>
}
// -----
// CHECK-LABEL: func @pooling_nhwc_max
// CHECK: linalg.pooling_nhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x1xf32>, memref<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x1xf32>)
func.func @pooling_nhwc_max(%input: memref<1x4x4x1xf32>, %fake: memref<3x3xf32>, %output: memref<1x2x2x1xf32>) {
linalg.pooling_nhwc_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: memref<1x4x4x1xf32>, memref<3x3xf32>)
outs(%output: memref<1x2x2x1xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nwc_max
// CHECK: linalg.pooling_nwc_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x1xf32>, memref<3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x1xf32>)
func.func @pooling_nwc_max(%input: memref<1x4x1xf32>, %fake: memref<3xf32>, %output: memref<1x2x1xf32>) {
linalg.pooling_nwc_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: memref<1x4x1xf32>, memref<3xf32>)
outs(%output: memref<1x2x1xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nhwc_i8_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x1xi8>, tensor<3x3xi8>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x1xi8>) -> tensor<1x2x2x1xi8>
func.func @pooling_nhwc_i8_max_tensor(%input: tensor<1x4x4x1xi8>) -> tensor<1x2x2x1xi8> {
%fake = tensor.empty() : tensor<3x3xi8>
%init = tensor.empty() : tensor<1x2x2x1xi8>
%cst = arith.constant 0 : i8
%fill = linalg.fill ins(%cst : i8) outs(%init : tensor<1x2x2x1xi8>) -> tensor<1x2x2x1xi8>
%res = linalg.pooling_nhwc_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: tensor<1x4x4x1xi8>, tensor<3x3xi8>)
outs(%fill: tensor<1x2x2x1xi8>) -> tensor<1x2x2x1xi8>
return %res : tensor<1x2x2x1xi8>
}
// -----
// CHECK-LABEL: func @pooling_nwc_i8_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nwc_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x1xi8>, tensor<3xi8>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x1xi8>) -> tensor<1x2x1xi8>
func.func @pooling_nwc_i8_max_tensor(%input: tensor<1x4x1xi8>) -> tensor<1x2x1xi8> {
%fake = tensor.empty() : tensor<3xi8>
%init = tensor.empty() : tensor<1x2x1xi8>
%cst = arith.constant 0 : i8
%fill = linalg.fill ins(%cst : i8) outs(%init : tensor<1x2x1xi8>) -> tensor<1x2x1xi8>
%res = linalg.pooling_nwc_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: tensor<1x4x1xi8>, tensor<3xi8>)
outs(%fill: tensor<1x2x1xi8>) -> tensor<1x2x1xi8>
return %res : tensor<1x2x1xi8>
}
// -----
// CHECK-LABEL: func @pooling_nhwc_i8_max
// CHECK: linalg.pooling_nhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x1xi8>, memref<3x3xi8>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x1xi8>)
func.func @pooling_nhwc_i8_max(%input: memref<1x4x4x1xi8>, %fake: memref<3x3xi8>, %output: memref<1x2x2x1xi8>) {
linalg.pooling_nhwc_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: memref<1x4x4x1xi8>, memref<3x3xi8>)
outs(%output: memref<1x2x2x1xi8>)
return
}
// -----
// CHECK-LABEL: func @pooling_nwc_i8_max
// CHECK: linalg.pooling_nwc_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x1xi8>, memref<3xi8>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x1xi8>)
func.func @pooling_nwc_i8_max(%input: memref<1x4x1xi8>, %fake: memref<3xi8>, %output: memref<1x2x1xi8>) {
linalg.pooling_nwc_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: memref<1x4x1xi8>, memref<3xi8>)
outs(%output: memref<1x2x1xi8>)
return
}
// -----
// CHECK-LABEL: func @pooling_nhwc_i16_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x1xi16>, tensor<3x3xi16>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x1xi16>) -> tensor<1x2x2x1xi16>
func.func @pooling_nhwc_i16_max_tensor(%input: tensor<1x4x4x1xi16>) -> tensor<1x2x2x1xi16> {
%fake = tensor.empty() : tensor<3x3xi16>
%init = tensor.empty() : tensor<1x2x2x1xi16>
%cst = arith.constant 0 : i16
%fill = linalg.fill ins(%cst : i16) outs(%init : tensor<1x2x2x1xi16>) -> tensor<1x2x2x1xi16>
%res = linalg.pooling_nhwc_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: tensor<1x4x4x1xi16>, tensor<3x3xi16>)
outs(%fill: tensor<1x2x2x1xi16>) -> tensor<1x2x2x1xi16>
return %res : tensor<1x2x2x1xi16>
}
// -----
// CHECK-LABEL: func @pooling_nwc_i16_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nwc_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x1xi16>, tensor<3xi16>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x1xi16>) -> tensor<1x2x1xi16>
func.func @pooling_nwc_i16_max_tensor(%input: tensor<1x4x1xi16>) -> tensor<1x2x1xi16> {
%fake = tensor.empty() : tensor<3xi16>
%init = tensor.empty() : tensor<1x2x1xi16>
%cst = arith.constant 0 : i16
%fill = linalg.fill ins(%cst : i16) outs(%init : tensor<1x2x1xi16>) -> tensor<1x2x1xi16>
%res = linalg.pooling_nwc_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: tensor<1x4x1xi16>, tensor<3xi16>)
outs(%fill: tensor<1x2x1xi16>) -> tensor<1x2x1xi16>
return %res : tensor<1x2x1xi16>
}
// -----
// CHECK-LABEL: func @pooling_nhwc_i16_max
// CHECK: linalg.pooling_nhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x1xi16>, memref<3x3xi16>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x1xi16>)
func.func @pooling_nhwc_i16_max(%input: memref<1x4x4x1xi16>, %fake: memref<3x3xi16>, %output: memref<1x2x2x1xi16>) {
linalg.pooling_nhwc_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: memref<1x4x4x1xi16>, memref<3x3xi16>)
outs(%output: memref<1x2x2x1xi16>)
return
}
// -----
// CHECK-LABEL: func @pooling_nwc_i16_max
// CHECK: linalg.pooling_nwc_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x1xi16>, memref<3xi16>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x1xi16>)
func.func @pooling_nwc_i16_max(%input: memref<1x4x1xi16>, %fake: memref<3xi16>, %output: memref<1x2x1xi16>) {
linalg.pooling_nwc_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: memref<1x4x1xi16>, memref<3xi16>)
outs(%output: memref<1x2x1xi16>)
return
}
// -----
// CHECK-LABEL: func @pooling_nhwc_i32_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x1xi32>, tensor<3x3xi32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x1xi32>) -> tensor<1x2x2x1xi32>
func.func @pooling_nhwc_i32_max_tensor(%input: tensor<1x4x4x1xi32>) -> tensor<1x2x2x1xi32> {
%fake = tensor.empty() : tensor<3x3xi32>
%init = tensor.empty() : tensor<1x2x2x1xi32>
%cst = arith.constant 0 : i32
%fill = linalg.fill ins(%cst : i32) outs(%init : tensor<1x2x2x1xi32>) -> tensor<1x2x2x1xi32>
%res = linalg.pooling_nhwc_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: tensor<1x4x4x1xi32>, tensor<3x3xi32>)
outs(%fill: tensor<1x2x2x1xi32>) -> tensor<1x2x2x1xi32>
return %res : tensor<1x2x2x1xi32>
}
// -----
// CHECK-LABEL: func @pooling_nwc_i32_max_tensor
// CHECK: %{{.+}} = linalg.pooling_nwc_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x1xi32>, tensor<3xi32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x1xi32>) -> tensor<1x2x1xi32>
func.func @pooling_nwc_i32_max_tensor(%input: tensor<1x4x1xi32>) -> tensor<1x2x1xi32> {
%fake = tensor.empty() : tensor<3xi32>
%init = tensor.empty() : tensor<1x2x1xi32>
%cst = arith.constant 0 : i32
%fill = linalg.fill ins(%cst : i32) outs(%init : tensor<1x2x1xi32>) -> tensor<1x2x1xi32>
%res = linalg.pooling_nwc_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: tensor<1x4x1xi32>, tensor<3xi32>)
outs(%fill: tensor<1x2x1xi32>) -> tensor<1x2x1xi32>
return %res : tensor<1x2x1xi32>
}
// -----
// CHECK-LABEL: func @pooling_nhwc_i32_max
// CHECK: linalg.pooling_nhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x1xi32>, memref<3x3xi32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x1xi32>)
func.func @pooling_nhwc_i32_max(%input: memref<1x4x4x1xi32>, %fake: memref<3x3xi32>, %output: memref<1x2x2x1xi32>) {
linalg.pooling_nhwc_max {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: memref<1x4x4x1xi32>, memref<3x3xi32>)
outs(%output: memref<1x2x2x1xi32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nwc_i32_max
// CHECK: linalg.pooling_nwc_max
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x1xi32>, memref<3xi32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x1xi32>)
func.func @pooling_nwc_i32_max(%input: memref<1x4x1xi32>, %fake: memref<3xi32>, %output: memref<1x2x1xi32>) {
linalg.pooling_nwc_max {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: memref<1x4x1xi32>, memref<3xi32>)
outs(%output: memref<1x2x1xi32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nhwc_min_tensor
// CHECK: %{{.+}} = linalg.pooling_nhwc_min
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x1xf32>, tensor<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
func.func @pooling_nhwc_min_tensor(%input: tensor<1x4x4x1xf32>) -> tensor<1x2x2x1xf32> {
%fake = tensor.empty() : tensor<3x3xf32>
%init = tensor.empty() : tensor<1x2x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
%res = linalg.pooling_nhwc_min {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: tensor<1x4x4x1xf32>, tensor<3x3xf32>)
outs(%fill: tensor<1x2x2x1xf32>) -> tensor<1x2x2x1xf32>
return %res : tensor<1x2x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_nwc_min_tensor
// CHECK: %{{.+}} = linalg.pooling_nwc_min
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x1xf32>, tensor<3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
func.func @pooling_nwc_min_tensor(%input: tensor<1x4x1xf32>) -> tensor<1x2x1xf32> {
%fake = tensor.empty() : tensor<3xf32>
%init = tensor.empty() : tensor<1x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
%res = linalg.pooling_nwc_min {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: tensor<1x4x1xf32>, tensor<3xf32>)
outs(%fill: tensor<1x2x1xf32>) -> tensor<1x2x1xf32>
return %res : tensor<1x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_nhwc_min
// CHECK: linalg.pooling_nhwc_min
// CHECK-SAME: dilations = dense<1> : tensor<2xi64>
// CHECK-SAME: strides = dense<1> : tensor<2xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x1xf32>, memref<3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x1xf32>)
func.func @pooling_nhwc_min(%input: memref<1x4x4x1xf32>, %fake: memref<3x3xf32>, %output: memref<1x2x2x1xf32>) {
linalg.pooling_nhwc_min {dilations = dense<1> : tensor<2xi64>, strides = dense<1> : tensor<2xi64>}
ins(%input, %fake: memref<1x4x4x1xf32>, memref<3x3xf32>)
outs(%output: memref<1x2x2x1xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_nwc_min
// CHECK: linalg.pooling_nwc_min
// CHECK-SAME: dilations = dense<1> : tensor<1xi64>
// CHECK-SAME: strides = dense<1> : tensor<1xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x1xf32>, memref<3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x1xf32>)
func.func @pooling_nwc_min(%input: memref<1x4x1xf32>, %fake: memref<3xf32>, %output: memref<1x2x1xf32>) {
linalg.pooling_nwc_min {dilations = dense<1> : tensor<1xi64>, strides = dense<1> : tensor<1xi64>}
ins(%input, %fake: memref<1x4x1xf32>, memref<3xf32>)
outs(%output: memref<1x2x1xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_ndhwc_sum_tensor
// CHECK: %{{.+}} = linalg.pooling_ndhwc_sum
// CHECK-SAME: dilations = dense<1> : tensor<3xi64>
// CHECK-SAME: strides = dense<1> : tensor<3xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x4x1xf32>, tensor<3x3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
func.func @pooling_ndhwc_sum_tensor(%input: tensor<1x4x4x4x1xf32>) -> tensor<1x2x2x2x1xf32> {
%fake = tensor.empty() : tensor<3x3x3xf32>
%init = tensor.empty() : tensor<1x2x2x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
%res = linalg.pooling_ndhwc_sum {dilations = dense<1> : tensor<3xi64>, strides = dense<1> : tensor<3xi64>}
ins(%input, %fake: tensor<1x4x4x4x1xf32>, tensor<3x3x3xf32>)
outs(%fill: tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
return %res : tensor<1x2x2x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_ndhwc_sum
// CHECK: linalg.pooling_ndhwc_sum
// CHECK-SAME: dilations = dense<1> : tensor<3xi64>
// CHECK-SAME: strides = dense<1> : tensor<3xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x4x1xf32>, memref<3x3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x2x1xf32>)
func.func @pooling_ndhwc_sum(%input: memref<1x4x4x4x1xf32>, %fake: memref<3x3x3xf32>, %output: memref<1x2x2x2x1xf32>) {
linalg.pooling_ndhwc_sum {dilations = dense<1> : tensor<3xi64>, strides = dense<1> : tensor<3xi64>}
ins(%input, %fake: memref<1x4x4x4x1xf32>, memref<3x3x3xf32>)
outs(%output: memref<1x2x2x2x1xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_ndhwc_max_tensor
// CHECK: %{{.+}} = linalg.pooling_ndhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<3xi64>
// CHECK-SAME: strides = dense<1> : tensor<3xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x4x1xf32>, tensor<3x3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
func.func @pooling_ndhwc_max_tensor(%input: tensor<1x4x4x4x1xf32>) -> tensor<1x2x2x2x1xf32> {
%fake = tensor.empty() : tensor<3x3x3xf32>
%init = tensor.empty() : tensor<1x2x2x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
%res = linalg.pooling_ndhwc_max {dilations = dense<1> : tensor<3xi64>, strides = dense<1> : tensor<3xi64>}
ins(%input, %fake: tensor<1x4x4x4x1xf32>, tensor<3x3x3xf32>)
outs(%fill: tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
return %res : tensor<1x2x2x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_ndhwc_max
// CHECK: linalg.pooling_ndhwc_max
// CHECK-SAME: dilations = dense<1> : tensor<3xi64>
// CHECK-SAME: strides = dense<1> : tensor<3xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x4x1xf32>, memref<3x3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x2x1xf32>)
func.func @pooling_ndhwc_max(%input: memref<1x4x4x4x1xf32>, %fake: memref<3x3x3xf32>, %output: memref<1x2x2x2x1xf32>) {
linalg.pooling_ndhwc_max {dilations = dense<1> : tensor<3xi64>, strides = dense<1> : tensor<3xi64>}
ins(%input, %fake: memref<1x4x4x4x1xf32>, memref<3x3x3xf32>)
outs(%output: memref<1x2x2x2x1xf32>)
return
}
// -----
// CHECK-LABEL: func @pooling_ndhwc_min_tensor
// CHECK: %{{.+}} = linalg.pooling_ndhwc_min
// CHECK-SAME: dilations = dense<1> : tensor<3xi64>
// CHECK-SAME: strides = dense<1> : tensor<3xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<1x4x4x4x1xf32>, tensor<3x3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
func.func @pooling_ndhwc_min_tensor(%input: tensor<1x4x4x4x1xf32>) -> tensor<1x2x2x2x1xf32> {
%fake = tensor.empty() : tensor<3x3x3xf32>
%init = tensor.empty() : tensor<1x2x2x2x1xf32>
%cst = arith.constant 0.000000e+00 : f32
%fill = linalg.fill ins(%cst : f32) outs(%init : tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
%res = linalg.pooling_ndhwc_min {dilations = dense<1> : tensor<3xi64>, strides = dense<1> : tensor<3xi64>}
ins(%input, %fake: tensor<1x4x4x4x1xf32>, tensor<3x3x3xf32>)
outs(%fill: tensor<1x2x2x2x1xf32>) -> tensor<1x2x2x2x1xf32>
return %res : tensor<1x2x2x2x1xf32>
}
// -----
// CHECK-LABEL: func @pooling_ndhwc_min
// CHECK: linalg.pooling_ndhwc_min
// CHECK-SAME: dilations = dense<1> : tensor<3xi64>
// CHECK-SAME: strides = dense<1> : tensor<3xi64>
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<1x4x4x4x1xf32>, memref<3x3x3xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<1x2x2x2x1xf32>)
func.func @pooling_ndhwc_min(%input: memref<1x4x4x4x1xf32>, %fake: memref<3x3x3xf32>, %output: memref<1x2x2x2x1xf32>) {
linalg.pooling_ndhwc_min {dilations = dense<1> : tensor<3xi64>, strides = dense<1> : tensor<3xi64>}
ins(%input, %fake: memref<1x4x4x4x1xf32>, memref<3x3x3xf32>)
outs(%output: memref<1x2x2x2x1xf32>)
return
}
// -----
#map0 = affine_map<(d0, d1, d2, d3, d4, d5, d6) -> (d0, d1 * 2, d2 * 2 + d5, d6)>
#map1 = affine_map<(d0, d1, d2, d3, d4, d5, d6) -> (d4, d5, d6, d3)>
#map2 = affine_map<(d0, d1, d2, d3, d4, d5, d6) -> (d0, d1, d2, d3)>
func.func @conv_interface_wrong_input_indexing_map(
%arg0 : tensor<?x?x?x?xf32>, %arg2 : tensor<?x?x?x?xf32>, %arg1 : tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32> {
// expected-error @+1 {{unexpected input index map for convolutions}}
%0 = "linalg.conv_2d_nhwc_hwcf"(%arg0, %arg1, %arg2) ({
^bb0(%arg3: f32, %arg4: f32, %arg5 : f32):
%1 = "arith.mulf"(%arg3, %arg4) : (f32, f32) -> f32
%2 = "arith.addf"(%arg5, %1) : (f32, f32) -> f32
"linalg.yield"(%2) : (f32) -> ()
}) {dilations = dense<1> : tensor<2xi64>, linalg.memoized_indexing_maps = [#map0, #map1, #map2], operand_segment_sizes = array<i32: 2, 1>, strides = dense<2> : tensor<2xi64>} : (tensor<?x?x?x?xf32>, tensor<?x?x?x?xf32>, tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
return %0 : tensor<?x?x?x?xf32>
}
// -----
#map0 = affine_map<(d0, d1, d2, d3, d4, d5, d6) -> (d0, d1 + d4, d2 + d5, d6)>
#map1 = affine_map<(d0, d1, d2, d3, d4, d5, d6) -> (d4, d5, d6, d3, d5 + 1)>
#map2 = affine_map<(d0, d1, d2, d3, d4, d5, d6) -> (d0, d1, d2, d3)>
func.func @conv_interface_wrong_num_operands(
%arg0 : tensor<?x?x?x?xf32>, %arg1 : tensor<?x?x?x?x?xf32>, %arg2 : tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32> {
// expected-error @+1 {{expected output/filter indexing maps to be projected permutations}}
%0 = "linalg.conv_2d_nhwc_hwcf"(%arg0, %arg1, %arg2) ({
^bb0(%arg3: f32, %arg4: f32, %arg5 : f32):
%1 = "arith.mulf"(%arg3, %arg4) : (f32, f32) -> f32
%2 = "arith.addf"(%arg5, %1) : (f32, f32) -> f32
"linalg.yield"(%2) : (f32) -> ()
}) {dilations = dense<1> : tensor<2xi64>, linalg.memoized_indexing_maps = [#map0, #map1, #map2], operand_segment_sizes = array<i32: 2, 1>, strides = dense<1> : tensor<2xi64>} : (tensor<?x?x?x?xf32>, tensor<?x?x?x?x?xf32>, tensor<?x?x?x?xf32>) -> tensor<?x?x?x?xf32>
return %0 : tensor<?x?x?x?xf32>
}
// -----
func.func @batch_reduce_matmul(%arg0: tensor<8x128x256xf32>, %arg1: tensor<8x256x512xf32>, %arg2: tensor<128x512xf32>) -> tensor<128x512xf32> {
// CHECK: %{{.+}} = linalg.batch_reduce_matmul
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : tensor<8x128x256xf32>, tensor<8x256x512xf32>)
// CHECK-SAME: outs(%{{.+}} : tensor<128x512xf32>) -> tensor<128x512xf32>
%0 = linalg.batch_reduce_matmul ins(%arg0, %arg1 : tensor<8x128x256xf32>, tensor<8x256x512xf32>) outs(%arg2: tensor<128x512xf32>) -> tensor<128x512xf32>
return %0: tensor<128x512xf32>
}
// -----
func.func @batch_reduce_matmul(%arg0: memref<?x?x?xf32>, %arg1: memref<?x?x?xf32>, %arg2: memref<?x?xf32>) {
// CHECK: linalg.batch_reduce_matmul
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<?x?x?xf32>, memref<?x?x?xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<?x?xf32>)
linalg.batch_reduce_matmul ins(%arg0, %arg1 : memref<?x?x?xf32>, memref<?x?x?xf32>) outs(%arg2: memref<?x?xf32>)
return
}
// -----
// CHECK-LABEL: func @matmul_transpose_b
// CHECK: linalg.matmul_transpose_b
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<3x5xf32>, memref<7x5xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<3x7xf32>)
func.func @matmul_transpose_b(%arg0: memref<3x5xf32>, %arg1: memref<7x5xf32>, %arg2: memref<3x7xf32>) {
linalg.matmul_transpose_b ins(%arg0, %arg1 : memref<3x5xf32>, memref<7x5xf32>) outs(%arg2: memref<3x7xf32>)
return
}
// -----
// CHECK-LABEL: func @batchmatmul_transpose_b
// CHECK: linalg.batch_matmul_transpose_b
// CHECK-SAME: ins(%{{.+}}, %{{.+}} : memref<2x3x5xf32>, memref<2x7x5xf32>)
// CHECK-SAME: outs(%{{.+}} : memref<2x3x7xf32>)
func.func @batchmatmul_transpose_b(%arg0: memref<2x3x5xf32>, %arg1: memref<2x7x5xf32>, %arg2: memref<2x3x7xf32>) {
linalg.batch_matmul_transpose_b ins(%arg0, %arg1 : memref<2x3x5xf32>, memref<2x7x5xf32>) outs(%arg2: memref<2x3x7xf32>)
return
}
|
