GCC Code Coverage Report
|
Coverage |
Exec |
Excl |
Total |
| Lines: |
91.7% |
12614 |
0 |
13753 |
| Functions: |
95.1% |
967 |
3 |
1020 |
| Branches: |
55.0% |
6381 |
0 |
11597 |
called 330 times
100.0%
-%
100.0%
called 330 times
100.0%
-%
100.0%
not called
14.3%
0.0%
0.0%
called 80369 times
100.0%
100.0%
100.0%
called 80170 times
100.0%
50.0%
75.0%
called 11687 times
100.0%
75.0%
100.0%
called 80613 times
100.0%
62.5%
80.0%
called 65704493 times
100.0%
62.5%
87.0%
called 65805288 times
100.0%
57.1%
90.0%
called 1602 times
70.0%
32.1%
52.0%
called 164 times
100.0%
75.0%
100.0%
called 333 times
100.0%
-%
100.0%
called 464 times
100.0%
-%
100.0%
called 154 times
100.0%
50.0%
100.0%
called 1602 times
100.0%
-%
100.0%
called 154 times
23.5%
2.2%
7.0%
called 169 times
100.0%
75.0%
83.0%
not called
0.0%
0.0%
0.0%
called 39 times
66.7%
50.0%
50.0%
called 28 times
100.0%
-%
100.0%
called 49 times
46.7%
22.7%
37.0%
called 22 times
100.0%
-%
100.0%
called 18223 times
100.0%
-%
100.0%
called 6630 times
100.0%
-%
100.0%
called 3294 times
16.7%
-%
100.0%
called 65921843 times
28.6%
-%
100.0%
called 65931057 times
100.0%
100.0%
100.0%
called 16 times
66.7%
21.4%
100.0%
called 16 times
100.0%
58.3%
100.0%
called 163750 times
100.0%
-%
100.0%
called 874 times
100.0%
-%
100.0%
called 542 times
100.0%
-%
100.0%
called 8022 times
100.0%
75.0%
66.0%
called 3701 times
100.0%
75.0%
66.0%
called 64864 times
100.0%
-%
100.0%
called 1663 times
100.0%
64.3%
81.0%
called 62 times
100.0%
87.5%
90.0%
called 18 times
80.0%
75.0%
81.0%
called 155 times
80.0%
75.0%
81.0%
called 4427 times
100.0%
100.0%
100.0%
called 1 time
100.0%
-%
100.0%
called 21643 times
100.0%
75.0%
90.0%
called 13 times
100.0%
-%
100.0%
called 5 times
100.0%
-%
100.0%
called 295 times
100.0%
-%
100.0%
called 6 times
100.0%
-%
100.0%
called 730 times
100.0%
-%
100.0%
called 5 times
80.0%
50.0%
66.0%
called 577 times
100.0%
-%
100.0%
called 11 times
90.9%
50.0%
71.0%
called 427 times
100.0%
-%
100.0%
called 668 times
100.0%
-%
100.0%
called 14684410 times
100.0%
-%
100.0%
called 51006659 times
100.0%
-%
100.0%
called 282 times
100.0%
-%
100.0%
called 332 times
100.0%
-%
100.0%
called 50383 times
100.0%
-%
100.0%
called 50404 times
100.0%
-%
100.0%
called 1 time
100.0%
-%
100.0%
called 5874 times
100.0%
-%
100.0%
called 7985 times
100.0%
-%
100.0%
called 1024 times
100.0%
-%
100.0%
called 47 times
100.0%
100.0%
100.0%
called 1672 times
100.0%
75.0%
80.0%
called 3955087 times
100.0%
-%
100.0%
called 12961814 times
100.0%
-%
100.0%
called 46841205 times
100.0%
-%
100.0%
called 142981 times
100.0%
75.0%
100.0%
called 58960 times
100.0%
95.0%
94.0%
called 50365431 times
100.0%
-%
100.0%
called 5819 times
48.0%
24.1%
40.0%
called 60 times
100.0%
68.8%
84.0%
called 101 times
100.0%
68.8%
84.0%
called 223 times
100.0%
68.8%
84.0%
called 1422 times
100.0%
-%
100.0%
called 172 times
100.0%
-%
66.0%
called 249 times
100.0%
-%
100.0%
called 3 times
100.0%
-%
100.0%
called 11438 times
100.0%
50.0%
57.0%
called 165166 times
100.0%
50.0%
100.0%
called 50 times
100.0%
50.0%
66.0%
called 6 times
85.7%
50.0%
66.0%
called 44 times
85.7%
50.0%
66.0%
called 36 times
85.7%
50.0%
66.0%
called 363 times
100.0%
50.0%
66.0%
called 3 times
100.0%
50.0%
66.0%
called 15 times
100.0%
50.0%
66.0%
called 17 times
100.0%
50.0%
66.0%
brotensor::attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:287)
called 14 times
100.0%
50.0%
66.0%
brotensor::attention_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:270)
called 14 times
100.0%
50.0%
66.0%
called 10 times
100.0%
50.0%
66.0%
called 2916 times
64.7%
30.8%
48.0%
called 2 times
100.0%
50.0%
66.0%
brotensor::batch_norm_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:2664)
called 5 times
90.9%
50.0%
66.0%
brotensor::batch_norm_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor&, int, int, int, int, float, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:2628)
called 9 times
92.9%
50.0%
66.0%
brotensor::batch_norm_inference(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, float, brotensor::Tensor&) (src/ops.cpp:2649)
called 11 times
87.5%
50.0%
66.0%
called 11 times
90.9%
50.0%
66.0%
called 2859433 times
100.0%
-%
100.0%
brotensor::bias_act_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, float, float, float, brotensor::Tensor&, brotensor::Tensor*) (src/ops.cpp:2809)
called 42 times
100.0%
75.0%
80.0%
called 42 times
100.0%
50.0%
66.0%
called 10 times
100.0%
50.0%
66.0%
called 10 times
100.0%
50.0%
66.0%
called 8 times
100.0%
50.0%
66.0%
called 24 times
100.0%
50.0%
66.0%
brotensor::causal_conv1d(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, brotensor::Tensor&, brotensor::Tensor&) (include/brotensor/ops/conv1d.h:151)
called 4 times
100.0%
-%
100.0%
called 26 times
87.5%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 36 times
100.0%
50.0%
66.0%
called 3 times
85.7%
50.0%
66.0%
called 3 times
100.0%
50.0%
66.0%
called 3 times
85.7%
50.0%
66.0%
called 54 times
100.0%
50.0%
66.0%
called 5 times
87.5%
50.0%
66.0%
called 8 times
100.0%
-%
100.0%
called 8 times
100.0%
-%
100.0%
called 16 times
87.5%
25.0%
33.0%
called 24 times
87.5%
25.0%
33.0%
called 20 times
90.9%
50.0%
70.0%
called 9 times
87.5%
25.0%
33.0%
brotensor::conv1d(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (include/brotensor/ops/conv1d.h:50)
called 7 times
100.0%
-%
100.0%
called 3 times
100.0%
-%
100.0%
brotensor::conv1d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (include/brotensor/ops/conv1d.h:71)
called 3 times
100.0%
-%
100.0%
brotensor::conv1d_backward_weight(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (include/brotensor/ops/conv1d.h:91)
called 3 times
100.0%
-%
100.0%
called 16 times
100.0%
50.0%
66.0%
brotensor::conv2d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (include/brotensor/ops/conv.h:76)
not called
0.0%
-%
0.0%
brotensor::conv2d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:719)
called 28 times
100.0%
50.0%
66.0%
brotensor::conv2d_backward_weight(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (include/brotensor/ops/conv.h:111)
not called
0.0%
-%
0.0%
brotensor::conv2d_backward_weight(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:733)
called 26 times
100.0%
50.0%
66.0%
brotensor::conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (include/brotensor/ops/conv.h:41)
not called
0.0%
-%
0.0%
brotensor::conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:705)
called 44 times
100.0%
50.0%
66.0%
brotensor::conv2d_int8w_fp16_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:2028)
not called
0.0%
0.0%
0.0%
brotensor::conv3d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:775)
called 37 times
100.0%
50.0%
66.0%
brotensor::conv3d_int8w_fp16_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:791)
called 6 times
100.0%
100.0%
100.0%
called 8 times
85.7%
50.0%
66.0%
brotensor::conv_transpose1d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:2335)
called 12 times
88.9%
50.0%
66.0%
brotensor::conv_transpose1d_backward_weight(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:2349)
called 10 times
88.9%
50.0%
66.0%
brotensor::conv_transpose1d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:2322)
called 144 times
87.5%
50.0%
66.0%
called 3 times
85.7%
50.0%
66.0%
brotensor::conv_transpose2d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:1095)
called 8 times
90.0%
50.0%
66.0%
brotensor::conv_transpose2d_backward_weight(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:1112)
called 7 times
90.0%
50.0%
66.0%
brotensor::conv_transpose2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:1077)
called 82 times
90.0%
50.0%
66.0%
called 10 times
85.7%
50.0%
66.0%
called 8 times
100.0%
50.0%
66.0%
called 8 times
87.5%
33.3%
40.0%
called 3 times
100.0%
50.0%
66.0%
called 32 times
100.0%
50.0%
66.0%
brotensor::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:1305)
called 10 times
92.9%
50.0%
66.0%
called 16 times
100.0%
50.0%
66.0%
brotensor::cross_attention_forward_train(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:1286)
called 10 times
92.3%
50.0%
66.0%
brotensor::cross_attention_forward_with_attn(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor const*, int, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:1220)
called 10 times
88.9%
50.0%
66.0%
called 14 times
100.0%
50.0%
66.0%
called 73 times
100.0%
100.0%
100.0%
brotensor::deform_conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:756)
called 12 times
100.0%
50.0%
66.0%
called 2 times
100.0%
100.0%
100.0%
called 1 time
85.7%
50.0%
66.0%
called 4 times
85.7%
50.0%
66.0%
called 65831 times
100.0%
-%
100.0%
called 211440 times
100.0%
-%
100.0%
called 14 times
100.0%
-%
100.0%
called 64973 times
100.0%
80.0%
100.0%
called 862 times
95.0%
90.0%
91.0%
called 261144 times
100.0%
-%
100.0%
called 4 times
100.0%
50.0%
81.0%
called 1 time
100.0%
50.0%
68.0%
called 2 times
100.0%
50.0%
80.0%
called 160 times
100.0%
-%
100.0%
called 103 times
100.0%
-%
100.0%
called 71151 times
100.0%
100.0%
100.0%
called 185363 times
100.0%
100.0%
100.0%
called 12838 times
100.0%
-%
100.0%
called 4814 times
100.0%
100.0%
100.0%
called 1228 times
100.0%
50.0%
60.0%
called 8 times
100.0%
87.5%
93.0%
called 6958 times
100.0%
100.0%
100.0%
brotensor::detail::cpu::(anonymous namespace)::apply_axis(float const*, float*, float const*, float const*, int const*, int, int, int, int, int, int, int) (src/cpu/rope_mrope.cpp:53)
called 18 times
95.0%
75.0%
77.0%
called 4262 times
100.0%
100.0%
100.0%
brotensor::detail::cpu::(anonymous namespace)::attention_core(float const*, float const*, float const*, float const*, int, int, int, int, bool, float*, float*, int, int, int, int, float*) (src/cpu/flash_attention.cpp:154)
called 106 times
100.0%
100.0%
75.0%
brotensor::detail::cpu::(anonymous namespace)::attention_core(float const*, float const*, float const*, float const*, int, int, int, int, bool, float*, float*, int, int, int, int, float*)::$_0::operator()(unsigned long) const (src/cpu/flash_attention.cpp:221)
called 5412 times
100.0%
50.0%
100.0%
brotensor::detail::cpu::(anonymous namespace)::attention_core(float const*, float const*, float const*, float const*, int, int, int, int, bool, float*, float*, int, int, int, int, float*)::$_1::operator()(int, float*) const (src/cpu/flash_attention.cpp:167)
called 5628 times
91.5%
78.9%
81.0%
brotensor::detail::cpu::(anonymous namespace)::attention_core_backward(float const*, float const*, float const*, float const*, float const*, int, int, int, int, bool, float*, float*, float*, float*, float*, float*) (src/cpu/flash_attention.cpp:252)
called 50 times
100.0%
100.0%
73.0%
brotensor::detail::cpu::(anonymous namespace)::attention_core_backward(float const*, float const*, float const*, float const*, float const*, int, int, int, int, bool, float*, float*, float*, float*, float*, float*)::$_0::operator()(unsigned long) const (src/cpu/flash_attention.cpp:368)
called 69 times
100.0%
50.0%
100.0%
brotensor::detail::cpu::(anonymous namespace)::attention_core_backward(float const*, float const*, float const*, float const*, float const*, int, int, int, int, bool, float*, float*, float*, float*, float*, float*)::$_1::operator()(int, float*, float*, float*) const (src/cpu/flash_attention.cpp:272)
called 146 times
100.0%
90.9%
97.0%
called 137564 times
100.0%
100.0%
100.0%
called 60 times
75.0%
18.8%
15.0%
called 157 times
62.5%
15.0%
15.0%
called 15 times
87.5%
41.7%
50.0%
called 82 times
90.9%
27.5%
35.0%
called 13 times
91.7%
55.0%
56.0%
called 61 times
86.7%
50.0%
52.0%
called 18 times
50.0%
9.1%
10.0%
called 22 times
75.0%
8.3%
23.0%
called 143 times
60.0%
7.1%
11.0%
called 712 times
60.0%
7.1%
11.0%
called 86 times
75.0%
6.2%
9.0%
called 135 times
60.0%
5.6%
8.0%
called 10 times
75.0%
16.7%
22.0%
called 48 times
75.0%
6.2%
9.0%
called 112 times
75.0%
6.2%
9.0%
called 17 times
75.0%
16.7%
22.0%
called 12 times
75.0%
6.2%
9.0%
called 82 times
60.0%
5.6%
8.0%
called 170 times
75.0%
6.2%
9.0%
called 79 times
75.0%
6.2%
9.0%
called 5 times
75.0%
16.7%
22.0%
called 116 times
60.0%
5.6%
8.0%
called 28 times
75.0%
6.2%
9.0%
called 92 times
60.0%
5.6%
8.0%
called 61 times
75.0%
16.7%
22.0%
called 60 times
75.0%
16.7%
22.0%
called 41 times
75.0%
6.2%
9.0%
called 157 times
60.0%
5.6%
8.0%
called 13 times
75.0%
16.7%
22.0%
called 163 times
60.0%
5.6%
8.0%
called 12 times
75.0%
16.7%
22.0%
called 22 times
75.0%
6.2%
9.0%
called 5 times
75.0%
16.7%
22.0%
called 76 times
60.0%
5.6%
8.0%
called 15 times
75.0%
16.7%
22.0%
called 6419 times
81.8%
30.0%
34.0%
brotensor::detail::cpu::(anonymous namespace)::check_geometry(char const*, int, int, int, int, int, int, int, int, int, int) (src/cpu/conv_transpose2d.cpp:62)
called 81 times
70.0%
18.8%
38.0%
called 151 times
75.0%
25.0%
40.0%
called 192 times
60.0%
8.3%
21.0%
called 25 times
60.0%
8.3%
21.0%
called 81 times
75.0%
25.0%
40.0%
called 17 times
100.0%
75.0%
80.0%
called 61 times
75.0%
6.2%
9.0%
called 79 times
62.5%
8.8%
8.0%
called 10 times
66.7%
10.0%
15.0%
called 16 times
63.6%
35.7%
35.0%
called 206010 times
100.0%
100.0%
100.0%
called 2602 times
100.0%
100.0%
100.0%
called 692 times
100.0%
100.0%
100.0%
called 702 times
100.0%
50.0%
71.0%
called 151 times
100.0%
-%
100.0%
called 162 times
100.0%
-%
100.0%
brotensor::detail::cpu::(anonymous namespace)::cross_attention_forward_core(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/cross_attention.cpp:84)
called 13 times
100.0%
23.9%
56.0%
brotensor::detail::cpu::(anonymous namespace)::cross_attention_forward_core(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/cross_attention.cpp:143)
called 48 times
100.0%
93.8%
97.0%
brotensor::detail::cpu::(anonymous namespace)::cross_attention_forward_core(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_1::operator()(unsigned long) const (src/cpu/cross_attention.cpp:178)
called 44 times
100.0%
87.5%
100.0%
brotensor::detail::cpu::(anonymous namespace)::cross_attention_forward_core(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_2::operator()(unsigned long) const (src/cpu/cross_attention.cpp:197)
called 76 times
100.0%
92.9%
100.0%
called 820 times
100.0%
100.0%
100.0%
called 54 times
100.0%
-%
50.0%
called 211 times
100.0%
87.5%
100.0%
called 176 times
100.0%
66.7%
100.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
called 2 times
100.0%
50.0%
68.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
called 2 times
100.0%
50.0%
68.0%
not called
0.0%
0.0%
0.0%
called 6 times
100.0%
50.0%
68.0%
called 3 times
100.0%
50.0%
68.0%
not called
0.0%
0.0%
0.0%
called 2 times
100.0%
50.0%
68.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
called 1 time
100.0%
50.0%
68.0%
called 513 times
100.0%
-%
100.0%
called 577 times
100.0%
-%
100.0%
called 548 times
100.0%
-%
100.0%
called 513 times
100.0%
-%
100.0%
called 513 times
100.0%
-%
100.0%
called 1089 times
100.0%
-%
100.0%
called 583 times
100.0%
-%
100.0%
called 513 times
100.0%
-%
100.0%
called 794 times
80.0%
33.3%
50.0%
called 107 times
100.0%
-%
50.0%
called 908 times
100.0%
87.5%
100.0%
called 36 times
100.0%
-%
50.0%
called 291 times
100.0%
83.3%
100.0%
called 1263 times
100.0%
90.0%
100.0%
called 8060200 times
100.0%
-%
100.0%
called 1006960 times
100.0%
-%
100.0%
called 4418 times
85.7%
22.2%
29.0%
called 384 times
100.0%
-%
100.0%
called 75 times
100.0%
-%
100.0%
called 12 times
100.0%
-%
100.0%
called 36 times
100.0%
-%
100.0%
called 7236 times
100.0%
87.5%
89.0%
called 7909 times
100.0%
93.8%
94.0%
called 318027 times
100.0%
100.0%
100.0%
called 50348 times
100.0%
100.0%
100.0%
called 403010 times
100.0%
100.0%
100.0%
called 152754 times
100.0%
-%
100.0%
called 4030100 times
100.0%
-%
100.0%
called 503480 times
100.0%
-%
100.0%
called 250256 times
100.0%
-%
100.0%
called 50348 times
100.0%
-%
100.0%
called 9952 times
100.0%
100.0%
100.0%
called 513 times
100.0%
-%
100.0%
called 577 times
100.0%
-%
100.0%
called 184512 times
100.0%
83.3%
88.0%
called 11578 times
80.0%
7.1%
15.0%
called 25 times
75.0%
16.7%
22.0%
called 701 times
75.0%
16.7%
22.0%
called 227 times
75.0%
16.7%
22.0%
called 591 times
75.0%
16.7%
22.0%
called 1230 times
66.7%
16.7%
18.0%
called 12838 times
66.7%
16.7%
18.0%
called 5 times
75.0%
16.7%
22.0%
called 569 times
100.0%
-%
100.0%
called 64 times
100.0%
-%
100.0%
called 40 times
100.0%
-%
100.0%
brotensor::detail::cpu::(anonymous namespace)::run_scan(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, brotensor::Tensor&, brotensor::Tensor&, char const*) (src/cpu/gated_delta_rule.cpp:59)
called 16 times
88.9%
23.1%
30.0%
called 96 times
100.0%
100.0%
100.0%
called 108162 times
100.0%
-%
100.0%
called 6913 times
100.0%
-%
100.0%
called 46977 times
100.0%
-%
100.0%
called 583 times
100.0%
-%
100.0%
called 583 times
100.0%
-%
100.0%
called 96 times
100.0%
-%
100.0%
called 8 times
37.5%
16.7%
25.0%
called 555764 times
100.0%
-%
100.0%
called 13152 times
100.0%
87.5%
89.0%
brotensor::detail::cpu::(anonymous namespace)::upfirdn2d_run(brotensor::Tensor const&, int, int, int, int, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&, char const*) (src/cpu/upfirdn2d.cpp:41)
called 38 times
92.5%
39.8%
43.0%
called 56 times
100.0%
62.5%
100.0%
called 56 times
100.0%
50.0%
100.0%
called 2998 times
75.0%
66.7%
71.0%
called 1 time
100.0%
-%
100.0%
called 2998 times
100.0%
92.9%
100.0%
called 28 times
100.0%
75.0%
100.0%
called 56 times
100.0%
100.0%
100.0%
called 56 times
100.0%
50.0%
100.0%
called 25 times
100.0%
62.5%
83.0%
called 4 times
92.7%
55.9%
68.0%
called 38 times
92.3%
62.5%
70.0%
called 22 times
100.0%
100.0%
100.0%
called 298 times
100.0%
75.0%
77.0%
not called
0.0%
0.0%
0.0%
called 8 times
100.0%
100.0%
100.0%
called 185 times
100.0%
100.0%
100.0%
called 7 times
95.8%
54.5%
75.0%
brotensor::detail::cpu::attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/ops_impl.cpp:562)
called 7 times
93.3%
50.0%
78.0%
brotensor::detail::cpu::attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/ops_impl.cpp:609)
called 62 times
100.0%
90.9%
100.0%
brotensor::detail::cpu::attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_1::operator()(unsigned long) const (src/cpu/ops_impl.cpp:646)
called 64 times
100.0%
85.7%
100.0%
brotensor::detail::cpu::attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_2::operator()(unsigned long) const (src/cpu/ops_impl.cpp:676)
called 52 times
100.0%
88.9%
100.0%
brotensor::detail::cpu::attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_3::operator()(unsigned long) const (src/cpu/ops_impl.cpp:695)
called 63 times
100.0%
90.0%
100.0%
brotensor::detail::cpu::attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_4::operator()(unsigned long) const (src/cpu/ops_impl.cpp:752)
called 66 times
100.0%
75.0%
100.0%
brotensor::detail::cpu::attention_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/ops_impl.cpp:443)
called 7 times
100.0%
50.0%
68.0%
brotensor::detail::cpu::attention_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/ops_impl.cpp:476)
called 66 times
100.0%
83.3%
100.0%
brotensor::detail::cpu::attention_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_1::operator()(unsigned long) const (src/cpu/ops_impl.cpp:500)
called 59 times
100.0%
92.9%
97.0%
brotensor::detail::cpu::attention_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_2::operator()(unsigned long) const (src/cpu/ops_impl.cpp:531)
called 64 times
100.0%
83.3%
100.0%
brotensor::detail::cpu::attention_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_3::operator()(unsigned long) const (src/cpu/ops_impl.cpp:545)
called 61 times
100.0%
91.7%
100.0%
called 5 times
88.6%
35.7%
57.0%
called 1 time
35.7%
25.0%
31.0%
brotensor::detail::cpu::batch_norm_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/batch_norm.cpp:224)
called 3 times
91.2%
27.3%
44.0%
brotensor::detail::cpu::batch_norm_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/batch_norm.cpp:291)
called 16 times
100.0%
90.0%
100.0%
brotensor::detail::cpu::batch_norm_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor&, int, int, int, int, float, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/batch_norm.cpp:67)
called 7 times
94.4%
29.2%
60.0%
brotensor::detail::cpu::batch_norm_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor&, int, int, int, int, float, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/batch_norm.cpp:123)
called 29 times
100.0%
92.9%
100.0%
brotensor::detail::cpu::batch_norm_inference(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, float, brotensor::Tensor&) (src/cpu/batch_norm.cpp:172)
called 6 times
100.0%
30.0%
58.0%
brotensor::detail::cpu::batch_norm_inference(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, float, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/batch_norm.cpp:209)
called 35 times
100.0%
83.3%
100.0%
called 7 times
88.0%
71.1%
89.0%
brotensor::detail::cpu::bias_act_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, float, float, float, brotensor::Tensor&, brotensor::Tensor*) (src/cpu/bias_act.cpp:96)
called 30 times
90.7%
64.8%
71.0%
called 30 times
93.5%
65.8%
76.0%
called 5 times
95.2%
43.8%
65.0%
called 5 times
88.9%
40.0%
72.0%
called 4 times
90.9%
37.5%
77.0%
called 14 times
93.1%
67.6%
82.0%
called 22 times
85.3%
50.0%
71.0%
called 2442 times
100.0%
100.0%
100.0%
called 6 times
100.0%
87.5%
90.0%
called 35 times
94.4%
64.3%
73.0%
called 2 times
83.9%
55.0%
64.0%
called 2 times
94.1%
64.3%
73.0%
called 2 times
95.2%
64.3%
71.0%
called 52 times
90.0%
50.0%
57.0%
called 3 times
87.1%
35.0%
43.0%
called 7 times
96.3%
60.7%
79.0%
called 10 times
91.7%
47.1%
64.0%
called 10 times
85.3%
46.9%
65.0%
not called
0.0%
0.0%
0.0%
called 31 times
90.0%
54.5%
68.0%
brotensor::detail::cpu::conv2d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv2d.cpp:186)
called 61 times
88.9%
44.4%
54.0%
brotensor::detail::cpu::conv2d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/conv2d.cpp:220)
called 79 times
100.0%
96.7%
100.0%
brotensor::detail::cpu::conv2d_backward_weight(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv2d.cpp:259)
called 60 times
94.0%
72.2%
75.0%
brotensor::detail::cpu::conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv2d.cpp:53)
called 71 times
90.6%
50.0%
58.0%
brotensor::detail::cpu::conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/conv2d.cpp:110)
called 100 times
100.0%
83.3%
100.0%
brotensor::detail::cpu::conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const::'lambda'(int, int)::operator()(int, int) const (src/cpu/conv2d.cpp:119)
called 34905 times
100.0%
100.0%
100.0%
brotensor::detail::cpu::conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const::'lambda0'(int, int)::operator()(int, int) const (src/cpu/conv2d.cpp:142)
called 73251 times
100.0%
100.0%
100.0%
brotensor::detail::cpu::conv3d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv3d.cpp:51)
called 25 times
94.9%
47.9%
58.0%
brotensor::detail::cpu::conv3d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/conv3d.cpp:115)
called 35 times
100.0%
83.3%
100.0%
brotensor::detail::cpu::conv3d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const::'lambda'(int, int, int)::operator()(int, int, int) const (src/cpu/conv3d.cpp:124)
called 5930 times
100.0%
100.0%
100.0%
brotensor::detail::cpu::conv3d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const::'lambda0'(int, int, int)::operator()(int, int, int) const (src/cpu/conv3d.cpp:155)
called 4336 times
100.0%
97.1%
100.0%
called 5 times
90.5%
55.0%
66.0%
brotensor::detail::cpu::conv_transpose1d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv1d.cpp:187)
called 7 times
87.0%
27.3%
42.0%
brotensor::detail::cpu::conv_transpose1d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/conv1d.cpp:222)
called 12 times
100.0%
92.9%
100.0%
brotensor::detail::cpu::conv_transpose1d_backward_weight(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv1d.cpp:249)
called 6 times
93.2%
58.3%
66.0%
brotensor::detail::cpu::conv_transpose1d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv1d.cpp:61)
called 138 times
86.2%
36.8%
48.0%
brotensor::detail::cpu::conv_transpose1d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/conv1d.cpp:106)
called 273 times
100.0%
71.4%
100.0%
brotensor::detail::cpu::conv_transpose1d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_1::operator()(unsigned long) const (src/cpu/conv1d.cpp:132)
called 273 times
100.0%
75.0%
79.0%
brotensor::detail::cpu::conv_transpose1d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_1::operator()(unsigned long) const::'lambda'(int)::operator()(int) const (src/cpu/conv1d.cpp:141)
called 1109 times
100.0%
90.0%
92.0%
brotensor::detail::cpu::conv_transpose1d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_1::operator()(unsigned long) const::'lambda0'(int)::operator()(int) const (src/cpu/conv1d.cpp:159)
called 5271 times
95.8%
75.0%
90.0%
called 2 times
90.5%
54.5%
67.0%
brotensor::detail::cpu::conv_transpose2d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv_transpose2d.cpp:251)
called 5 times
89.3%
27.3%
43.0%
brotensor::detail::cpu::conv_transpose2d_backward_input(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/conv_transpose2d.cpp:295)
called 6 times
100.0%
95.5%
100.0%
brotensor::detail::cpu::conv_transpose2d_backward_weight(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv_transpose2d.cpp:335)
called 4 times
94.9%
65.9%
71.0%
brotensor::detail::cpu::conv_transpose2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/conv_transpose2d.cpp:85)
called 72 times
90.3%
50.0%
56.0%
brotensor::detail::cpu::conv_transpose2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/conv_transpose2d.cpp:137)
called 76 times
100.0%
63.6%
100.0%
brotensor::detail::cpu::conv_transpose2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_1::operator()(unsigned long) const (src/cpu/conv_transpose2d.cpp:175)
called 76 times
100.0%
75.0%
100.0%
brotensor::detail::cpu::conv_transpose2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_1::operator()(unsigned long) const::'lambda'(int, int)::operator()(int, int) const (src/cpu/conv_transpose2d.cpp:184)
called 17963 times
100.0%
100.0%
100.0%
brotensor::detail::cpu::conv_transpose2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_1::operator()(unsigned long) const::'lambda0'(int, int)::operator()(int, int) const (src/cpu/conv_transpose2d.cpp:210)
called 175842 times
100.0%
100.0%
100.0%
called 5 times
98.3%
60.9%
72.0%
called 4 times
100.0%
-%
100.0%
called 4 times
100.0%
100.0%
100.0%
called 2 times
100.0%
40.0%
72.0%
called 29 times
100.0%
70.0%
90.0%
called 71892 times
100.0%
50.0%
66.0%
called 165 times
100.0%
-%
100.0%
called 71892 times
100.0%
50.0%
100.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
called 10515 times
100.0%
50.0%
100.0%
called 2914 times
100.0%
-%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/cross_attention.cpp:385)
called 5 times
94.7%
50.0%
58.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/cross_attention.cpp:450)
called 39 times
100.0%
85.7%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_1::operator()(unsigned long) const (src/cpu/cross_attention.cpp:468)
called 214 times
100.0%
85.7%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_2::operator()(unsigned long) const (src/cpu/cross_attention.cpp:484)
called 27 times
100.0%
87.5%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_3::operator()(unsigned long) const (src/cpu/cross_attention.cpp:516)
called 27 times
100.0%
90.9%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_4::operator()(unsigned long) const (src/cpu/cross_attention.cpp:540)
called 27 times
100.0%
87.5%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_5::operator()(unsigned long) const (src/cpu/cross_attention.cpp:574)
called 217 times
100.0%
75.0%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_6::operator()(unsigned long) const (src/cpu/cross_attention.cpp:589)
called 218 times
100.0%
75.0%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_7::operator()(unsigned long) const (src/cpu/cross_attention.cpp:608)
called 37 times
100.0%
80.0%
100.0%
brotensor::detail::cpu::cross_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_8::operator()(unsigned long) const (src/cpu/cross_attention.cpp:627)
called 55 times
100.0%
80.0%
100.0%
called 8 times
100.0%
50.0%
50.0%
brotensor::detail::cpu::cross_attention_forward_train(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/cross_attention.cpp:236)
called 5 times
100.0%
-%
100.0%
brotensor::detail::cpu::cross_attention_forward_with_attn(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor const*, int, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/cross_attention.cpp:256)
called 5 times
100.0%
42.1%
58.0%
brotensor::detail::cpu::cross_attention_forward_with_attn(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor const*, int, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/cross_attention.cpp:306)
called 20 times
100.0%
89.3%
97.0%
brotensor::detail::cpu::cross_attention_forward_with_attn(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor const*, int, brotensor::Tensor&, brotensor::Tensor&)::$_1::operator()(unsigned long) const (src/cpu/cross_attention.cpp:338)
called 27 times
100.0%
75.0%
100.0%
brotensor::detail::cpu::cross_attention_forward_with_attn(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor const*, int, brotensor::Tensor&, brotensor::Tensor&)::$_2::operator()(unsigned long) const (src/cpu/cross_attention.cpp:351)
called 22 times
100.0%
80.0%
100.0%
brotensor::detail::cpu::cross_attention_forward_with_attn(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor const*, int, brotensor::Tensor&, brotensor::Tensor&)::$_3::operator()(unsigned long) const (src/cpu/cross_attention.cpp:370)
called 28 times
100.0%
75.0%
100.0%
called 7 times
92.9%
37.5%
57.0%
brotensor::detail::cpu::deform_conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/deform_conv2d.cpp:60)
called 6 times
86.2%
44.1%
52.0%
brotensor::detail::cpu::deform_conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/deform_conv2d.cpp:124)
called 7 times
100.0%
92.9%
95.0%
called 7 times
96.3%
54.5%
74.0%
called 6 times
96.0%
54.5%
74.0%
called 6 times
90.3%
30.0%
53.0%
called 7 times
100.0%
50.0%
78.0%
called 134 times
100.0%
50.0%
78.0%
called 10 times
100.0%
83.3%
90.0%
called 8 times
94.7%
62.5%
75.0%
called 127 times
100.0%
66.7%
100.0%
called 6 times
89.5%
35.7%
55.0%
called 5 times
100.0%
70.0%
90.0%
called 45 times
100.0%
40.0%
72.0%
called 335 times
93.3%
55.6%
70.0%
called 11699857 times
100.0%
-%
100.0%
called 1090 times
100.0%
70.8%
80.0%
called 1090 times
100.0%
80.0%
85.0%
called 2118945 times
100.0%
100.0%
80.0%
called 138959 times
100.0%
87.5%
88.0%
not called
0.0%
0.0%
0.0%
called 4523136 times
89.7%
88.9%
82.0%
called 138953 times
100.0%
100.0%
100.0%
called 1970990 times
100.0%
100.0%
100.0%
called 364 times
100.0%
100.0%
100.0%
called 151 times
100.0%
100.0%
85.0%
called 22554550 times
100.0%
-%
100.0%
called 22284181 times
100.0%
-%
100.0%
called 2118772 times
80.0%
75.0%
87.0%
called 364 times
100.0%
100.0%
100.0%
called 2119864 times
100.0%
-%
100.0%
brotensor::detail::cpu::flash_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, bool, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/flash_attention.cpp:701)
called 8 times
81.8%
37.5%
43.0%
called 31 times
90.4%
51.5%
65.0%
called 7 times
90.2%
53.8%
64.0%
called 34 times
80.0%
40.0%
46.0%
called 9 times
74.1%
33.3%
37.0%
called 4 times
93.8%
58.3%
65.0%
brotensor::detail::cpu::flash_attention_q_with_kv_cached_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor&) (src/cpu/flash_attention.cpp:618)
called 6 times
85.7%
48.0%
61.0%
brotensor::detail::cpu::flash_attention_qkvo_backward(brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*) (src/cpu/flash_attention.cpp:859)
called 9 times
66.7%
50.0%
68.0%
brotensor::detail::cpu::flash_attention_qkvo_backward(brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*)::$_0::operator()(brotensor::Tensor const*, brotensor::Tensor const*) const (src/cpu/flash_attention.cpp:888)
called 36 times
93.6%
55.3%
100.0%
brotensor::detail::cpu::flash_attention_qkvo_forward(brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor&) (src/cpu/flash_attention.cpp:653)
called 15 times
93.3%
54.4%
68.0%
brotensor::detail::cpu::flash_attention_varlen_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int const*, int const*, int, int, int, int, int, bool, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/flash_attention.cpp:745)
called 17 times
86.1%
43.0%
49.0%
brotensor::detail::cpu::flash_attention_varlen_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int const*, int const*, int, int, int, int, int, bool, brotensor::Tensor&) (src/cpu/flash_attention.cpp:507)
called 17 times
77.6%
38.9%
44.0%
not called
0.0%
0.0%
0.0%
called 4 times
100.0%
66.7%
84.0%
called 5 times
91.1%
47.4%
67.0%
brotensor::detail::cpu::gated_delta_rule_chunked(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/gated_delta_rule.cpp:205)
called 7 times
100.0%
-%
100.0%
brotensor::detail::cpu::gated_delta_rule_step(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/gated_delta_rule.cpp:218)
called 9 times
100.0%
-%
100.0%
called 7 times
100.0%
66.7%
91.0%
called 3 times
100.0%
-%
100.0%
called 3 times
100.0%
-%
100.0%
called 4 times
100.0%
-%
100.0%
called 5 times
100.0%
-%
100.0%
called 3 times
100.0%
50.0%
80.0%
called 3 times
100.0%
50.0%
80.0%
called 4 times
100.0%
50.0%
80.0%
called 5 times
100.0%
50.0%
80.0%
brotensor::detail::cpu::group_norm_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/group_norm.cpp:112)
called 25 times
81.2%
36.1%
50.0%
brotensor::detail::cpu::group_norm_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/group_norm.cpp:176)
called 33 times
100.0%
90.9%
100.0%
called 41 times
85.7%
27.3%
43.0%
called 57 times
100.0%
90.0%
100.0%
called 21 times
94.4%
59.1%
74.0%
called 4 times
93.9%
50.0%
64.0%
called 6 times
92.3%
38.2%
60.0%
called 12 times
100.0%
-%
100.0%
called 7 times
100.0%
66.1%
90.0%
called 62 times
100.0%
-%
100.0%
brotensor::detail::cpu::interp2d_forward_impl(brotensor::Tensor const&, int, int, int, int, int, int, int, bool, brotensor::Tensor&, char const*) (src/cpu/interp2d.cpp:112)
called 74 times
100.0%
74.5%
97.0%
called 364 times
90.6%
57.1%
68.0%
called 7 times
96.7%
75.0%
80.0%
called 5111 times
97.2%
72.9%
80.0%
called 11 times
97.1%
74.0%
80.0%
called 14 times
82.6%
30.0%
39.0%
called 6 times
97.2%
54.5%
75.0%
called 67 times
100.0%
61.1%
87.0%
called 5 times
93.3%
50.0%
65.0%
called 6 times
93.3%
71.4%
89.0%
called 8 times
87.5%
50.0%
78.0%
called 123 times
100.0%
90.0%
100.0%
called 3 times
88.2%
65.0%
87.0%
called 8 times
85.7%
21.4%
54.0%
called 125 times
100.0%
87.5%
100.0%
called 22 times
92.3%
30.0%
58.0%
called 34 times
100.0%
87.5%
100.0%
called 4 times
100.0%
50.0%
78.0%
called 88 times
100.0%
50.0%
78.0%
called 20 times
100.0%
76.9%
82.0%
called 3 times
94.4%
81.8%
94.0%
called 881 times
100.0%
71.4%
80.0%
called 8 times
72.1%
50.0%
60.0%
called 3 times
100.0%
40.0%
72.0%
called 49 times
100.0%
70.0%
90.0%
brotensor::detail::cpu::lstm_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*) (src/cpu/lstm.cpp:149)
called 407 times
95.1%
61.1%
69.0%
brotensor::detail::cpu::lstm_forward_train(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, int, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*) (src/cpu/lstm.cpp:42)
called 901 times
96.7%
85.0%
86.0%
called 4 times
100.0%
75.0%
89.0%
called 6 times
100.0%
75.0%
92.0%
called 14 times
75.0%
33.3%
47.0%
called 285 times
100.0%
87.5%
100.0%
brotensor::detail::cpu::matmul_abt(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, int, int, int, int, long long, long long, long long, brotensor::Tensor const*, int) (src/cpu/matmul.cpp:66)
called 51 times
100.0%
90.6%
90.0%
brotensor::detail::cpu::matmul_abt(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, int, int, int, int, long long, long long, long long, brotensor::Tensor const*, int)::$_0::operator()(unsigned short) const (src/cpu/matmul.cpp:90)
called 6680 times
100.0%
100.0%
100.0%
brotensor::detail::cpu::matmul_abt(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, int, int, int, int, long long, long long, long long, brotensor::Tensor const*, int)::$_1::operator()(float) const (src/cpu/matmul.cpp:94)
called 640 times
100.0%
100.0%
100.0%
brotensor::detail::cpu::matmul_abt(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, int, int, int, int, long long, long long, long long, brotensor::Tensor const*, int)::$_2::operator()(float) const (src/cpu/matmul.cpp:98)
called 640 times
100.0%
100.0%
100.0%
called 8 times
76.2%
30.8%
41.0%
called 79 times
100.0%
83.3%
100.0%
called 118 times
100.0%
83.3%
100.0%
called 5 times
88.4%
45.5%
58.0%
brotensor::detail::cpu::max_pool2d_forward(brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/pool2d.cpp:173)
called 13 times
95.2%
51.7%
61.0%
brotensor::detail::cpu::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*) (src/cpu/ops_impl.cpp:915)
called 19 times
93.8%
50.0%
73.0%
brotensor::detail::cpu::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*)::$_0::operator()(unsigned long) const (src/cpu/ops_impl.cpp:966)
called 178 times
100.0%
91.2%
100.0%
brotensor::detail::cpu::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*)::$_1::operator()(unsigned long) const (src/cpu/ops_impl.cpp:1016)
called 69 times
100.0%
87.5%
100.0%
brotensor::detail::cpu::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*)::$_2::operator()(unsigned long) const (src/cpu/ops_impl.cpp:1047)
called 67 times
100.0%
90.0%
100.0%
brotensor::detail::cpu::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*)::$_3::operator()(unsigned long) const (src/cpu/ops_impl.cpp:1069)
called 68 times
100.0%
70.6%
100.0%
brotensor::detail::cpu::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*)::$_4::operator()(unsigned long) const (src/cpu/ops_impl.cpp:1105)
called 10 times
100.0%
75.0%
87.0%
brotensor::detail::cpu::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*)::$_5::operator()(unsigned long) const (src/cpu/ops_impl.cpp:1128)
called 68 times
100.0%
80.0%
100.0%
brotensor::detail::cpu::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*)::$_6::operator()(unsigned long) const (src/cpu/ops_impl.cpp:1154)
called 181 times
100.0%
80.0%
100.0%
brotensor::detail::cpu::mha_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/ops_impl.cpp:769)
called 26 times
100.0%
69.2%
87.0%
brotensor::detail::cpu::mha_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_0::operator()(unsigned long) const (src/cpu/ops_impl.cpp:813)
called 107 times
100.0%
92.9%
100.0%
brotensor::detail::cpu::mha_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_1::operator()(unsigned long) const (src/cpu/ops_impl.cpp:845)
called 106 times
100.0%
93.3%
97.0%
brotensor::detail::cpu::mha_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_2::operator()(unsigned long) const (src/cpu/ops_impl.cpp:879)
called 101 times
100.0%
87.5%
100.0%
brotensor::detail::cpu::mha_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&)::$_3::operator()(unsigned long) const (src/cpu/ops_impl.cpp:897)
called 242 times
100.0%
92.9%
100.0%
called 6 times
95.7%
43.8%
63.0%
brotensor::detail::cpu::modulated_conv2d_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/modulated_conv2d.cpp:140)
called 12 times
97.8%
62.2%
75.0%
brotensor::detail::cpu::modulated_conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/modulated_conv2d.cpp:76)
called 8 times
92.2%
48.2%
62.0%
called 3 times
100.0%
-%
100.0%
called 3 times
84.6%
60.0%
81.0%
called 3 times
100.0%
75.0%
87.0%
called 3 times
81.0%
56.2%
80.0%
called 147 times
88.9%
50.0%
56.0%
called 11 times
100.0%
56.2%
85.0%
called 6 times
90.0%
47.6%
58.0%
called 154 times
89.7%
45.0%
56.0%
brotensor::detail::cpu::pad2d_backward(brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/pad2d.cpp:134)
called 7 times
97.2%
64.3%
79.0%
brotensor::detail::cpu::pad2d_forward(brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/pad2d.cpp:86)
called 52 times
97.2%
75.0%
85.0%
called 2997 times
100.0%
-%
100.0%
called 6 times
93.0%
53.1%
65.0%
called 4 times
100.0%
56.2%
85.0%
called 52 times
100.0%
56.2%
85.0%
called 29 times
97.3%
71.9%
76.0%
called 3 times
100.0%
50.0%
80.0%
called 4 times
100.0%
50.0%
80.0%
called 4 times
91.7%
62.5%
64.0%
called 4 times
100.0%
100.0%
100.0%
called 6 times
100.0%
100.0%
100.0%
called 2 times
100.0%
66.7%
73.0%
called 3 times
82.4%
27.8%
40.0%
called 1841 times
100.0%
100.0%
100.0%
called 6 times
100.0%
100.0%
100.0%
called 3 times
100.0%
100.0%
100.0%
called 18 times
100.0%
100.0%
100.0%
not called
0.0%
0.0%
0.0%
called 11 times
100.0%
71.4%
88.0%
called 146 times
100.0%
70.8%
90.0%
brotensor::detail::cpu::resblock_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, int, int, int, int, int, int, float, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*) (src/cpu/resblock.cpp:185)
called 8 times
84.0%
51.8%
62.0%
brotensor::detail::cpu::resblock_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, int, int, int, int, int, int, float, brotensor::Tensor&) (src/cpu/resblock.cpp:121)
called 8 times
93.9%
48.1%
60.0%
called 332 times
95.8%
66.7%
76.0%
called 7 times
89.7%
57.1%
68.0%
called 7 times
84.2%
30.0%
54.0%
called 36 times
100.0%
85.7%
100.0%
called 9 times
83.3%
31.2%
47.0%
called 44 times
100.0%
83.3%
100.0%
called 8 times
88.6%
40.0%
55.0%
called 2 times
88.6%
40.0%
55.0%
brotensor::detail::cpu::rope_apply_mrope(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int const*, int const*, int const*, int, int, int, int, int, brotensor::Tensor&) (src/cpu/rope_mrope.cpp:82)
called 7 times
89.7%
45.0%
55.0%
called 2 times
84.6%
36.8%
50.0%
called 4 times
91.9%
50.0%
62.0%
called 9 times
91.9%
50.0%
62.0%
called 2 times
100.0%
40.0%
72.0%
called 4 times
100.0%
40.0%
72.0%
called 5 times
92.0%
41.7%
63.0%
called 4 times
100.0%
40.0%
72.0%
called 29 times
100.0%
70.0%
90.0%
called 50256 times
98.0%
70.0%
84.0%
called 161898 times
100.0%
77.3%
100.0%
called 7 times
100.0%
59.1%
66.0%
called 132 times
70.6%
58.3%
100.0%
called 280 times
100.0%
100.0%
100.0%
called 2 times
100.0%
38.9%
47.0%
called 6 times
96.0%
45.8%
65.0%
brotensor::detail::cpu::self_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/self_attention.cpp:69)
called 7 times
100.0%
-%
100.0%
brotensor::detail::cpu::self_attention_bias_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, brotensor::Tensor const*, int, float, brotensor::Tensor&) (src/cpu/self_attention_bias.cpp:38)
called 21 times
85.0%
37.5%
76.0%
brotensor::detail::cpu::self_attention_bias_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, brotensor::Tensor const*, int, float, brotensor::Tensor&)::$_0::operator()(brotensor::Tensor const*, char const*) const (src/cpu/self_attention_bias.cpp:74)
called 84 times
92.0%
28.1%
27.0%
brotensor::detail::cpu::self_attention_bias_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, brotensor::Tensor const*, int, float, brotensor::Tensor&)::$_1::operator()(float const*, float const*, std::__1::vector<float, std::__1::allocator<float>>&) const (src/cpu/self_attention_bias.cpp:102)
called 63 times
100.0%
91.4%
100.0%
brotensor::detail::cpu::self_attention_decomposed_rel_pos_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, float, brotensor::Tensor&) (src/cpu/self_attention_decomposed_rel_pos.cpp:54)
called 66 times
75.0%
25.0%
70.0%
brotensor::detail::cpu::self_attention_decomposed_rel_pos_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, float, brotensor::Tensor&)::$_0::operator()(brotensor::Tensor const*, char const*) const (src/cpu/self_attention_decomposed_rel_pos.cpp:91)
called 264 times
92.3%
37.5%
27.0%
brotensor::detail::cpu::self_attention_decomposed_rel_pos_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, float, brotensor::Tensor&)::$_1::operator()(float const*, float const*, std::__1::vector<float, std::__1::allocator<float>>&) const (src/cpu/self_attention_decomposed_rel_pos.cpp:122)
called 198 times
100.0%
86.2%
100.0%
brotensor::detail::cpu::self_attention_decomposed_rel_pos_windowed_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, float, brotensor::Tensor&) (src/cpu/self_attention_decomposed_rel_pos.cpp:213)
called 10 times
96.4%
53.1%
64.0%
called 3 times
100.0%
50.0%
50.0%
brotensor::detail::cpu::self_attention_forward_train(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/cpu/self_attention.cpp:50)
called 9 times
100.0%
-%
100.0%
called 11 times
100.0%
56.2%
85.0%
not called
0.0%
0.0%
0.0%
called 6 times
100.0%
100.0%
100.0%
called 20 times
100.0%
100.0%
100.0%
called 19 times
100.0%
50.0%
80.0%
called 37 times
100.0%
75.0%
90.0%
called 2 times
100.0%
40.0%
72.0%
called 29 times
100.0%
40.0%
72.0%
called 4 times
97.0%
53.8%
73.0%
called 6 times
96.6%
55.0%
74.0%
called 8 times
85.9%
57.6%
62.0%
called 120 times
88.9%
52.5%
63.0%
called 11 times
100.0%
100.0%
100.0%
called 24 times
100.0%
94.4%
95.0%
called 27 times
100.0%
83.3%
96.0%
called 3 times
100.0%
-%
100.0%
called 5 times
90.5%
75.0%
88.0%
called 6 times
90.2%
73.2%
89.0%
called 10 times
100.0%
92.9%
94.0%
called 23 times
95.0%
60.0%
74.0%
not called
0.0%
0.0%
0.0%
called 105 times
100.0%
-%
100.0%
called 1286 times
90.9%
50.0%
56.0%
called 16337 times
100.0%
100.0%
84.0%
called 11 times
96.1%
67.6%
78.0%
called 6 times
100.0%
50.0%
76.0%
called 6 times
100.0%
50.0%
76.0%
called 5 times
95.7%
50.0%
68.0%
called 5 times
94.7%
50.0%
68.0%
called 6 times
100.0%
100.0%
100.0%
called 3 times
100.0%
100.0%
100.0%
called 21 times
100.0%
100.0%
100.0%
called 3 times
100.0%
100.0%
100.0%
called 4 times
72.2%
30.0%
45.0%
called 4 times
93.3%
53.8%
68.0%
called 12 times
100.0%
70.8%
87.0%
called 105 times
100.0%
-%
100.0%
brotensor::detail::cpu::unfold2d_forward(brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/cpu/unfold2d.cpp:59)
called 5 times
90.4%
50.0%
60.0%
brotensor::detail::cpu::upfirdn2d_backward(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&) (src/cpu/upfirdn2d.cpp:129)
called 18 times
92.9%
25.0%
37.0%
brotensor::detail::cpu::upfirdn2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&) (src/cpu/upfirdn2d.cpp:118)
called 20 times
100.0%
-%
100.0%
called 8 times
100.0%
62.5%
83.0%
called 6 times
100.0%
64.7%
84.0%
called 8 times
100.0%
61.1%
87.0%
called 6 times
100.0%
68.2%
90.0%
called 4 times
100.0%
66.7%
84.0%
called 6 times
90.5%
44.1%
66.0%
called 9 times
97.6%
61.5%
79.0%
called 5 times
97.6%
61.5%
79.0%
called 3 times
100.0%
100.0%
100.0%
called 1720 times
100.0%
-%
100.0%
called 53982 times
100.0%
-%
100.0%
called 7162 times
100.0%
-%
100.0%
called 1678 times
100.0%
-%
100.0%
called 89 times
100.0%
-%
100.0%
called 107 times
100.0%
-%
100.0%
called 77 times
100.0%
-%
100.0%
called 158 times
100.0%
-%
100.0%
called 705 times
100.0%
75.0%
85.0%
called 158 times
100.0%
100.0%
100.0%
called 9540 times
63.6%
57.1%
62.0%
called 65922 times
100.0%
100.0%
100.0%
called 319 times
100.0%
-%
100.0%
called 319 times
100.0%
-%
100.0%
called 4 times
100.0%
50.0%
81.0%
not called
0.0%
0.0%
0.0%
not called
0.0%
0.0%
0.0%
called 27 times
58.3%
70.0%
75.0%
not called
0.0%
0.0%
0.0%
called 14 times
100.0%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 407 times
100.0%
100.0%
100.0%
called 413 times
100.0%
100.0%
33.0%
called 123920 times
100.0%
100.0%
26.0%
called 123650 times
90.9%
87.5%
42.0%
called 123892 times
100.0%
80.0%
76.0%
called 1 time
100.0%
-%
100.0%
called 9 times
100.0%
50.0%
66.0%
called 1 time
100.0%
-%
100.0%
called 137 times
100.0%
50.0%
66.0%
called 25 times
100.0%
50.0%
66.0%
called 24 times
100.0%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 7 times
100.0%
50.0%
66.0%
called 48 times
100.0%
50.0%
66.0%
called 339 times
100.0%
50.0%
66.0%
brotensor::filtered_lrelu_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, float, float, float, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor*) (src/filtered_lrelu.cpp:86)
called 7 times
53.8%
25.0%
50.0%
brotensor::filtered_lrelu_backward_composite(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, float, float, float, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor*) (src/filtered_lrelu.cpp:143)
called 7 times
75.0%
37.5%
35.0%
brotensor::filtered_lrelu_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, float, float, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/filtered_lrelu.cpp:65)
called 7 times
50.0%
50.0%
75.0%
brotensor::filtered_lrelu_forward_composite(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, int, int, int, int, int, int, int, int, int, int, float, float, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/filtered_lrelu.cpp:109)
called 7 times
94.4%
41.7%
53.0%
brotensor::flash_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, bool, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:1551)
called 16 times
100.0%
50.0%
66.0%
called 1 time
100.0%
-%
100.0%
called 51 times
100.0%
50.0%
66.0%
called 12 times
88.9%
50.0%
66.0%
called 65 times
100.0%
50.0%
66.0%
called 15 times
85.7%
50.0%
66.0%
called 8 times
87.5%
50.0%
66.0%
not called
0.0%
0.0%
0.0%
brotensor::flash_attention_q_with_kv_cached_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor&) (src/ops.cpp:1576)
called 12 times
87.5%
50.0%
66.0%
brotensor::flash_attention_q_with_kv_cached_int8w_fp16(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor&) (src/ops.cpp:2154)
not called
0.0%
0.0%
0.0%
brotensor::flash_attention_qkvo_backward(brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor*) (src/ops.cpp:1517)
called 18 times
95.2%
91.7%
92.0%
brotensor::flash_attention_qkvo_forward(brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor&) (src/ops.cpp:1501)
called 24 times
88.9%
50.0%
66.0%
brotensor::flash_attention_qkvo_int8w_fp16(brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, float const*, int, bool, brotensor::Tensor&) (src/ops.cpp:2175)
not called
0.0%
0.0%
0.0%
brotensor::flash_attention_varlen_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int const*, int const*, int, int, int, int, int, bool, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:1481)
called 27 times
92.3%
50.0%
66.0%
brotensor::flash_attention_varlen_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int const*, int const*, int, int, int, int, int, bool, brotensor::Tensor&) (src/ops.cpp:1466)
called 27 times
88.9%
50.0%
66.0%
not called
0.0%
0.0%
0.0%
called 5231094 times
89.5%
87.5%
92.0%
called 3445496 times
100.0%
100.0%
100.0%
called 6170996 times
96.6%
78.6%
87.0%
called 6 times
85.7%
50.0%
66.0%
called 8 times
87.5%
50.0%
66.0%
brotensor::gated_delta_rule_chunked(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:2598)
called 11 times
88.9%
50.0%
66.0%
brotensor::gated_delta_rule_step(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:2612)
called 12 times
88.9%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 8 times
100.0%
50.0%
66.0%
called 10 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 8 times
100.0%
50.0%
66.0%
called 10 times
100.0%
50.0%
66.0%
called 56 times
100.0%
-%
100.0%
called 977 times
100.0%
75.0%
83.0%
called 7 times
100.0%
83.3%
85.0%
called 4 times
100.0%
-%
100.0%
called 3 times
100.0%
-%
100.0%
called 3 times
100.0%
-%
100.0%
called 12 times
100.0%
-%
100.0%
called 3 times
100.0%
-%
100.0%
called 3 times
100.0%
-%
100.0%
called 160 times
100.0%
-%
66.0%
called 6 times
100.0%
-%
100.0%
called 327 times
100.0%
-%
100.0%
called 460 times
100.0%
-%
100.0%
called 25 times
100.0%
-%
100.0%
called 14 times
100.0%
50.0%
75.0%
called 35 times
100.0%
50.0%
70.0%
called 81 times
100.0%
100.0%
100.0%
called 28 times
77.8%
33.3%
39.0%
called 76 times
95.0%
59.5%
85.0%
called 95 times
100.0%
65.0%
83.0%
called 29 times
100.0%
70.0%
88.0%
called 52 times
100.0%
-%
100.0%
called 2 times
100.0%
-%
100.0%
called 6 times
100.0%
-%
100.0%
called 23 times
100.0%
100.0%
100.0%
called 44 times
100.0%
100.0%
100.0%
called 20 times
100.0%
62.5%
75.0%
called 20 times
100.0%
62.5%
75.0%
called 1 time
100.0%
-%
100.0%
called 28 times
97.2%
69.4%
77.0%
called 172 times
100.0%
-%
100.0%
called 168 times
100.0%
-%
100.0%
called 340 times
96.4%
66.7%
100.0%
called 1 time
100.0%
50.0%
100.0%
called 28 times
100.0%
-%
100.0%
called 9 times
100.0%
-%
100.0%
called 1 time
100.0%
-%
100.0%
called 5 times
100.0%
-%
100.0%
called 54 times
100.0%
-%
100.0%
called 166 times
100.0%
-%
100.0%
called 158 times
100.0%
-%
100.0%
called 324 times
100.0%
-%
100.0%
called 190 times
100.0%
-%
100.0%
called 180 times
100.0%
-%
100.0%
called 370 times
100.0%
-%
100.0%
called 20 times
100.0%
81.8%
86.0%
called 16 times
100.0%
64.7%
71.0%
brotensor::group_norm_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:821)
called 18 times
100.0%
50.0%
66.0%
called 18 times
100.0%
50.0%
66.0%
called 23 times
100.0%
50.0%
66.0%
called 6 times
85.7%
50.0%
66.0%
called 9 times
85.7%
50.0%
66.0%
called 156 times
70.6%
36.1%
54.0%
called 18 times
85.7%
50.0%
66.0%
called 13 times
100.0%
50.0%
66.0%
called 76 times
100.0%
50.0%
66.0%
called 367 times
100.0%
50.0%
66.0%
called 9 times
85.7%
50.0%
66.0%
called 430 times
100.0%
-%
100.0%
called 5114 times
100.0%
50.0%
66.0%
called 14 times
87.5%
50.0%
66.0%
called 28 times
100.0%
50.0%
66.0%
called 11 times
100.0%
50.0%
66.0%
called 73 times
100.0%
50.0%
66.0%
called 10 times
85.7%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 16 times
90.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 16 times
90.9%
50.0%
66.0%
called 26 times
85.7%
50.0%
66.0%
not called
0.0%
0.0%
0.0%
called 6 times
85.7%
50.0%
66.0%
called 94 times
85.7%
50.0%
66.0%
called 30 times
100.0%
50.0%
66.0%
called 9 times
100.0%
50.0%
66.0%
called 1264 times
100.0%
50.0%
66.0%
called 28 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 8 times
85.7%
50.0%
66.0%
not called
0.0%
0.0%
0.0%
called 3 times
85.7%
50.0%
66.0%
called 3 times
85.7%
50.0%
66.0%
called 3 times
85.7%
50.0%
66.0%
called 34 times
85.7%
50.0%
66.0%
called 34 times
85.7%
50.0%
66.0%
called 34 times
85.7%
50.0%
66.0%
called 5 times
100.0%
50.0%
66.0%
called 52 times
100.0%
50.0%
66.0%
brotensor::lora_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*) (include/brotensor/ops/lora.h:61)
called 6 times
100.0%
52.8%
57.0%
called 270 times
100.0%
56.2%
71.0%
brotensor::lstm_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*) (src/ops.cpp:1764)
called 413 times
100.0%
70.0%
90.0%
brotensor::lstm_forward_train(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, int, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*) (src/ops.cpp:1749)
called 907 times
100.0%
83.3%
85.0%
called 14 times
100.0%
50.0%
66.0%
called 18 times
100.0%
50.0%
66.0%
called 60 times
100.0%
50.0%
66.0%
brotensor::matmul_abt(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor&, int, int, int, int, long long, long long, long long, brotensor::Tensor const*, int) (src/ops.cpp:1730)
called 62 times
100.0%
50.0%
66.0%
called 16 times
100.0%
50.0%
66.0%
called 1 time
50.0%
50.0%
66.0%
called 8 times
85.7%
50.0%
66.0%
brotensor::max_pool2d_forward(brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:1035)
called 19 times
88.9%
50.0%
66.0%
brotensor::mha_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*) (src/ops.cpp:330)
called 24 times
100.0%
90.0%
90.0%
brotensor::mha_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:309)
called 26 times
100.0%
50.0%
66.0%
brotensor::mha_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (include/brotensor/ops/attention.h:86)
called 16 times
100.0%
-%
100.0%
called 12 times
100.0%
50.0%
66.0%
brotensor::modulated_conv2d_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:2862)
called 16 times
90.0%
50.0%
66.0%
brotensor::modulated_conv2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:2847)
called 12 times
88.9%
50.0%
66.0%
called 3 times
100.0%
50.0%
66.0%
called 9 times
100.0%
50.0%
66.0%
called 9 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 159 times
100.0%
50.0%
66.0%
called 19 times
100.0%
50.0%
66.0%
called 9 times
85.7%
50.0%
66.0%
called 157 times
85.7%
50.0%
66.0%
brotensor::pad2d_backward(brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:957)
called 11 times
100.0%
50.0%
66.0%
brotensor::pad2d_forward(brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:949)
called 61 times
100.0%
50.0%
66.0%
called 12 times
85.7%
50.0%
66.0%
called 7 times
100.0%
50.0%
66.0%
called 55 times
100.0%
50.0%
66.0%
called 47 times
85.7%
50.0%
66.0%
called 20 times
50.0%
33.3%
41.0%
called 825 times
100.0%
68.8%
80.0%
called 6 times
100.0%
50.0%
66.0%
called 8 times
100.0%
50.0%
66.0%
called 4 times
100.0%
50.0%
66.0%
called 4 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 2 times
100.0%
50.0%
66.0%
called 6 times
85.7%
50.0%
66.0%
called 11 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 93 times
100.0%
50.0%
66.0%
called 3 times
100.0%
50.0%
66.0%
called 16 times
85.7%
50.0%
66.0%
called 152 times
85.7%
50.0%
66.0%
brotensor::resblock_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, int, int, int, int, int, int, float, brotensor::Tensor const&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor*, brotensor::Tensor*, brotensor::Tensor*) (src/ops.cpp:1680)
called 16 times
100.0%
75.0%
92.0%
brotensor::resblock_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, int, int, int, int, int, int, float, brotensor::Tensor&) (src/ops.cpp:1636)
called 16 times
100.0%
50.0%
66.0%
brotensor::resblock_forward_int8w_fp16(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, int, int, int, int, int, int, float, brotensor::Tensor&) (src/ops.cpp:1655)
not called
0.0%
0.0%
0.0%
called 335 times
100.0%
50.0%
66.0%
called 9 times
100.0%
50.0%
66.0%
called 14 times
100.0%
50.0%
66.0%
called 18 times
100.0%
50.0%
66.0%
called 26 times
100.0%
50.0%
66.0%
called 4 times
85.7%
50.0%
66.0%
brotensor::rope_apply_mrope(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, int const*, int const*, int const*, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:1828)
called 11 times
90.0%
50.0%
66.0%
called 2 times
85.7%
50.0%
66.0%
called 8 times
100.0%
50.0%
66.0%
called 14 times
100.0%
50.0%
66.0%
called 3 times
100.0%
50.0%
66.0%
called 7 times
100.0%
50.0%
66.0%
called 11 times
85.7%
50.0%
66.0%
called 7 times
100.0%
50.0%
66.0%
called 32 times
100.0%
50.0%
66.0%
called 64 times
100.0%
-%
100.0%
called 64 times
100.0%
-%
100.0%
called 54 times
100.0%
-%
100.0%
not called
0.0%
0.0%
0.0%
called 629 times
100.0%
100.0%
100.0%
called 392 times
100.0%
62.5%
75.0%
called 8 times
100.0%
50.0%
70.0%
not called
0.0%
0.0%
0.0%
called 91 times
100.0%
80.0%
84.0%
called 180 times
57.8%
36.8%
44.0%
called 8 times
100.0%
72.2%
88.0%
called 1145 times
100.0%
90.0%
100.0%
called 19 times
90.9%
70.0%
69.0%
called 19 times
100.0%
100.0%
100.0%
called 27 times
77.8%
26.7%
30.0%
called 32 times
100.0%
67.2%
88.0%
called 30 times
100.0%
70.0%
88.0%
called 48 times
100.0%
-%
100.0%
called 2 times
100.0%
-%
100.0%
called 36 times
100.0%
100.0%
100.0%
called 5 times
100.0%
57.1%
72.0%
called 26 times
100.0%
68.8%
77.0%
called 2 times
100.0%
50.0%
100.0%
called 27 times
100.0%
-%
100.0%
called 14 times
100.0%
-%
100.0%
called 7 times
100.0%
-%
100.0%
called 50 times
100.0%
-%
100.0%
called 82 times
100.0%
-%
100.0%
called 84 times
100.0%
-%
100.0%
called 43 times
90.0%
83.3%
80.0%
called 166 times
100.0%
-%
100.0%
called 56 times
100.0%
-%
100.0%
called 28 times
100.0%
-%
100.0%
called 60 times
100.0%
-%
100.0%
called 144 times
100.0%
-%
100.0%
called 13 times
100.0%
100.0%
100.0%
called 90 times
100.0%
100.0%
100.0%
called 32 times
100.0%
100.0%
100.0%
called 9 times
100.0%
63.9%
74.0%
called 21 times
61.5%
47.1%
68.0%
called 62 times
100.0%
75.7%
100.0%
called 7 times
100.0%
-%
100.0%
called 4 times
100.0%
57.7%
70.0%
called 5 times
100.0%
62.5%
74.0%
called 13 times
98.3%
57.8%
69.0%
not called
72.7%
25.0%
0.0%
called 50266 times
100.0%
50.0%
66.0%
called 13 times
87.5%
50.0%
66.0%
called 298 times
100.0%
50.0%
66.0%
called 4 times
100.0%
50.0%
66.0%
called 9 times
85.7%
50.0%
66.0%
brotensor::self_attention_backward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:1254)
called 14 times
92.3%
50.0%
66.0%
brotensor::self_attention_bias_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, brotensor::Tensor const*, float const*, brotensor::Tensor const*, int, float, brotensor::Tensor&) (src/ops.cpp:1352)
called 39 times
87.5%
50.0%
66.0%
called 33 times
100.0%
-%
100.0%
brotensor::self_attention_bias_int8w_fp16(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, brotensor::Tensor const*, int, float, brotensor::Tensor&) (src/ops.cpp:1416)
called 5 times
90.9%
50.0%
66.0%
brotensor::self_attention_decomposed_rel_pos_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, float, brotensor::Tensor&) (src/ops.cpp:1377)
called 23 times
81.8%
50.0%
66.0%
brotensor::self_attention_decomposed_rel_pos_windowed_forward(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const*, brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, float, brotensor::Tensor&) (src/ops.cpp:1396)
called 15 times
81.8%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
brotensor::self_attention_forward_train(brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, brotensor::Tensor const&, float const*, int, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&, brotensor::Tensor&) (src/ops.cpp:1235)
called 16 times
92.3%
50.0%
66.0%
called 19 times
100.0%
50.0%
66.0%
called 2 times
50.0%
8.3%
16.0%
called 5 times
100.0%
50.0%
66.0%
called 11 times
100.0%
50.0%
66.0%
called 26 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 10 times
100.0%
50.0%
66.0%
called 3 times
100.0%
50.0%
66.0%
called 32 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 11 times
90.0%
75.0%
80.0%
called 124 times
85.7%
50.0%
66.0%
called 21 times
100.0%
50.0%
66.0%
called 35 times
100.0%
50.0%
66.0%
called 44 times
100.0%
50.0%
66.0%
called 3 times
100.0%
50.0%
66.0%
called 16 times
100.0%
50.0%
66.0%
called 12 times
90.9%
50.0%
66.0%
called 7 times
100.0%
50.0%
66.0%
called 34 times
85.7%
50.0%
66.0%
called 9 times
90.9%
56.2%
75.0%
called 1290 times
100.0%
50.0%
66.0%
called 14 times
87.5%
50.0%
66.0%
called 14 times
100.0%
50.0%
66.0%
called 15 times
100.0%
50.0%
66.0%
called 10 times
100.0%
50.0%
66.0%
called 10 times
100.0%
50.0%
66.0%
called 11 times
44.4%
8.3%
16.0%
called 2914 times
100.0%
66.7%
80.0%
called 11 times
100.0%
50.0%
66.0%
called 6 times
100.0%
50.0%
66.0%
called 97 times
100.0%
50.0%
66.0%
called 9 times
100.0%
50.0%
66.0%
called 9 times
100.0%
50.0%
66.0%
called 8 times
100.0%
50.0%
66.0%
called 17 times
100.0%
50.0%
66.0%
brotensor::unfold2d_forward(brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, brotensor::Tensor&) (src/ops.cpp:981)
called 10 times
100.0%
50.0%
66.0%
brotensor::upfirdn2d_backward(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&) (src/ops.cpp:2833)
called 21 times
100.0%
50.0%
66.0%
brotensor::upfirdn2d_forward(brotensor::Tensor const&, brotensor::Tensor const&, int, int, int, int, int, int, int, int, int, int, int, int, int, int, bool, float, brotensor::Tensor&) (src/ops.cpp:2821)
called 25 times
100.0%
50.0%
66.0%
called 15 times
100.0%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 15 times
100.0%
50.0%
66.0%
called 12 times
100.0%
50.0%
66.0%
called 6 times
85.7%
50.0%
66.0%
called 9 times
87.5%
50.0%
66.0%
called 14 times
85.7%
50.0%
66.0%
called 8 times
85.7%
50.0%
66.0%
called 3 times
100.0%
50.0%
66.0%
called 6 times
95.7%
31.8%
46.0%
called 9 times
94.7%
31.8%
46.0%