BertForMaskedLMTF [fp32,seqlen512,tensorflow] (exported\_tf) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2799.129 |
2808.946 |
29.514 |
BertForMaskedLMTF [fp32,seqlen512,tensorflow] (exported\_tf) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
772.880 |
773.024 |
4.853 |
BertForMaskedLMTF [fp32,seqlen512,tensorflow] (exported\_tf) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
426.981 |
427.374 |
1.387 |
BertForMaskedLMTF [fp32,seqlen512,tensorflow] (exported\_tf) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2858.480 |
2862.826 |
48.327 |
BertForMaskedLMTF [fp32,seqlen512,tensorflow] (exported\_tf) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2804.156 |
2804.853 |
33.159 |
BertForMaskedLMTF [fp32,seqlen512,tensorflow] (exported\_tf) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
770.220 |
771.036 |
4.538 |
BertForMaskedLMTF [fp32,seqlen512,tensorflow] (exported\_tf) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
428.628 |
428.579 |
1.054 |
BertForMaskedLMTF [fp32,seqlen512,tensorflow] (exported\_tf) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2890.286 |
2898.554 |
39.229 |
DeepLabV3\_fp32 [fp32] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
31.311 |
31.253 |
0.120 |
DeepLabV3\_fp32 [fp32] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
12.940 |
12.957 |
0.088 |
DeepLabV3\_fp32 [fp32] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
10.235 |
10.229 |
0.027 |
DeepLabV3\_fp32 [fp32] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
29.248 |
29.273 |
0.211 |
DeepLabV3\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
27.318 |
27.315 |
0.079 |
DeepLabV3\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
10.776 |
10.780 |
0.040 |
DeepLabV3\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
8.351 |
8.356 |
0.021 |
DeepLabV3\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
26.247 |
26.218 |
0.137 |
EfficientNetV2STF [fp32,cnn,tensorflow] (exported\_tf) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
274.696 |
274.635 |
1.644 |
EfficientNetV2STF [fp32,cnn,tensorflow] (exported\_tf) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
86.532 |
86.572 |
0.212 |
EfficientNetV2STF [fp32,cnn,tensorflow] (exported\_tf) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
57.950 |
57.954 |
0.078 |
EfficientNetV2STF [fp32,cnn,tensorflow] (exported\_tf) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
274.436 |
274.831 |
1.393 |
EfficientNetV2STF [fp32,cnn,tensorflow] (exported\_tf) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
311.265 |
311.153 |
1.836 |
EfficientNetV2STF [fp32,cnn,tensorflow] (exported\_tf) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
96.117 |
95.938 |
0.574 |
EfficientNetV2STF [fp32,cnn,tensorflow] (exported\_tf) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
63.087 |
63.088 |
0.081 |
EfficientNetV2STF [fp32,cnn,tensorflow] (exported\_tf) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
311.440 |
311.508 |
0.917 |
EfficientNet\_int8 [int8] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
120.665 |
120.696 |
0.141 |
EfficientNet\_int8 [int8] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
38.632 |
38.632 |
0.041 |
EfficientNet\_int8 [int8] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
24.030 |
24.042 |
0.040 |
EfficientNet\_int8 [int8] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
119.573 |
119.586 |
0.083 |
EfficientNet\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
120.416 |
120.377 |
0.301 |
EfficientNet\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
38.631 |
38.631 |
0.037 |
EfficientNet\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
23.896 |
23.892 |
0.023 |
EfficientNet\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
119.180 |
119.143 |
0.120 |
MiniLML12H384Uncased [int32,seqlen128] (exported\_tf) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
95.866 |
95.720 |
0.657 |
MiniLML12H384Uncased [int32,seqlen128] (exported\_tf) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
33.389 |
33.367 |
0.152 |
MiniLML12H384Uncased [int32,seqlen128] (exported\_tf) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
23.080 |
23.088 |
0.050 |
MiniLML12H384Uncased [int32,seqlen128] (exported\_tf) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
93.150 |
92.920 |
0.848 |
MiniLML12H384Uncased [int32,seqlen128] (exported\_tf) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
94.759 |
94.789 |
0.931 |
MiniLML12H384Uncased [int32,seqlen128] (exported\_tf) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
33.450 |
33.375 |
0.195 |
MiniLML12H384Uncased [int32,seqlen128] (exported\_tf) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
23.230 |
23.221 |
0.058 |
MiniLML12H384Uncased [int32,seqlen128] (exported\_tf) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
94.505 |
94.471 |
0.511 |
MobileBertSquad\_fp16 [fp16] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
192.756 |
192.319 |
1.099 |
MobileBertSquad\_fp16 [fp16] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
71.904 |
71.876 |
0.209 |
MobileBertSquad\_fp16 [fp16] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
49.826 |
49.822 |
0.135 |
MobileBertSquad\_fp16 [fp16] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
196.917 |
196.619 |
0.926 |
MobileBertSquad\_fp16 [fp16] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
190.853 |
192.573 |
4.815 |
MobileBertSquad\_fp16 [fp16] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
72.057 |
72.083 |
0.232 |
MobileBertSquad\_fp16 [fp16] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
50.009 |
50.006 |
0.087 |
MobileBertSquad\_fp16 [fp16] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
197.325 |
196.607 |
1.776 |
MobileBertSquad\_fp32 [fp32] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
191.914 |
191.659 |
1.264 |
MobileBertSquad\_fp32 [fp32] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
73.744 |
73.629 |
0.391 |
MobileBertSquad\_fp32 [fp32] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
51.772 |
51.784 |
0.183 |
MobileBertSquad\_fp32 [fp32] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
193.179 |
193.259 |
2.133 |
MobileBertSquad\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
191.448 |
191.263 |
1.054 |
MobileBertSquad\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
74.387 |
74.494 |
0.398 |
MobileBertSquad\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
51.720 |
51.696 |
0.291 |
MobileBertSquad\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
193.287 |
193.801 |
2.399 |
MobileBertSquad\_int8 [int8] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
485.377 |
485.341 |
0.539 |
MobileBertSquad\_int8 [int8] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
152.489 |
152.432 |
0.164 |
MobileBertSquad\_int8 [int8] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
96.122 |
96.138 |
0.106 |
MobileBertSquad\_int8 [int8] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
484.560 |
484.474 |
0.354 |
MobileBertSquad\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
484.604 |
484.693 |
0.390 |
MobileBertSquad\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
153.036 |
153.044 |
0.150 |
MobileBertSquad\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
95.999 |
96.000 |
0.092 |
MobileBertSquad\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
484.488 |
484.472 |
0.510 |
MobileNetV1\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
30.545 |
30.557 |
0.405 |
MobileNetV1\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
12.161 |
12.147 |
0.082 |
MobileNetV1\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
10.747 |
10.767 |
0.046 |
MobileNetV1\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
30.159 |
30.147 |
0.259 |
MobileNetV1\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
26.047 |
26.002 |
0.212 |
MobileNetV1\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
9.896 |
9.894 |
0.113 |
MobileNetV1\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
8.447 |
8.444 |
0.043 |
MobileNetV1\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
25.906 |
25.847 |
0.251 |
MobileNetV2\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
14.363 |
14.289 |
0.216 |
MobileNetV2\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
7.155 |
7.156 |
0.028 |
MobileNetV2\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
6.153 |
6.142 |
0.064 |
MobileNetV2\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
14.116 |
14.076 |
0.113 |
MobileNetV2\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
11.476 |
11.473 |
0.121 |
MobileNetV2\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
5.024 |
5.027 |
0.029 |
MobileNetV2\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
4.308 |
4.304 |
0.017 |
MobileNetV2\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
11.433 |
11.405 |
0.132 |
MobileNetV3Small\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
4.418 |
4.406 |
0.068 |
MobileNetV3Small\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
3.267 |
3.269 |
0.028 |
MobileNetV3Small\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
3.394 |
3.394 |
0.016 |
MobileNetV3Small\_fp32 [fp32,imagenet] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
4.219 |
4.215 |
0.029 |
MobileNetV3Small\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
4.105 |
4.109 |
0.017 |
MobileNetV3Small\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2.795 |
2.796 |
0.012 |
MobileNetV3Small\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2.896 |
2.896 |
0.011 |
MobileNetV3Small\_fp32 [fp32,imagenet] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
4.049 |
4.039 |
0.023 |
MobileSSD\_fp32 [fp32] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
42.197 |
42.138 |
0.289 |
MobileSSD\_fp32 [fp32] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
16.950 |
16.839 |
0.308 |
MobileSSD\_fp32 [fp32] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
12.882 |
12.884 |
0.031 |
MobileSSD\_fp32 [fp32] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
42.123 |
42.071 |
0.364 |
MobileSSD\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
34.158 |
34.150 |
0.158 |
MobileSSD\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
12.505 |
12.497 |
0.061 |
MobileSSD\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
9.201 |
9.204 |
0.040 |
MobileSSD\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
33.954 |
33.914 |
0.159 |
PersonDetect\_int8 [int8] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2.565 |
2.565 |
0.002 |
PersonDetect\_int8 [int8] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
1.662 |
1.661 |
0.007 |
PersonDetect\_int8 [int8] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
1.613 |
1.610 |
0.008 |
PersonDetect\_int8 [int8] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2.474 |
2.474 |
0.001 |
PersonDetect\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2.567 |
2.568 |
0.002 |
PersonDetect\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
1.656 |
1.657 |
0.004 |
PersonDetect\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
1.617 |
1.616 |
0.007 |
PersonDetect\_int8 [int8] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
2.471 |
2.470 |
0.001 |
PoseNet\_fp32 [fp32] (exported\_tflite) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
21.639 |
21.586 |
0.260 |
PoseNet\_fp32 [fp32] (exported\_tflite) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
8.397 |
8.401 |
0.040 |
PoseNet\_fp32 [fp32] (exported\_tflite) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
6.487 |
6.481 |
0.027 |
PoseNet\_fp32 [fp32] (exported\_tflite) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
21.293 |
21.228 |
0.283 |
PoseNet\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
19.912 |
19.899 |
0.062 |
PoseNet\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
7.265 |
7.231 |
0.076 |
PoseNet\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
5.406 |
5.404 |
0.018 |
PoseNet\_fp32 [fp32] (exported\_tflite) [experimental-flags,fuse-padding][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
19.764 |
19.709 |
0.234 |
Resnet50TF [fp32] (exported\_tf) [default-flags][1-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
194.149 |
194.378 |
2.211 |
Resnet50TF [fp32] (exported\_tf) [default-flags][4-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
58.114 |
58.171 |
0.272 |
Resnet50TF [fp32] (exported\_tf) [default-flags][8-thread,full-inference,default-flags] with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
34.055 |
34.015 |
0.159 |
Resnet50TF [fp32] (exported\_tf) [default-flags][full-inference,default-flags] with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) |
193.892 |
193.741 |
1.848 |