Closed pfultz2 closed 2 weeks ago
Test | Batch | Rate new a037e2 |
Rate old 45b59b |
Diff | Compare | |
---|---|---|---|---|---|---|
torchvision-resnet50 | 64 | 3,258.99 | 3,258.81 | 0.01% | :white_check_mark: | |
torchvision-resnet50_fp16 | 64 | 7,003.93 | 6,987.27 | 0.24% | :white_check_mark: | |
torchvision-densenet121 | 32 | 2,437.24 | 2,434.79 | 0.10% | :white_check_mark: | |
torchvision-densenet121_fp16 | 32 | 4,101.99 | 4,075.87 | 0.64% | :white_check_mark: | |
torchvision-inceptionv3 | 32 | 1,638.61 | 1,637.98 | 0.04% | :white_check_mark: | |
torchvision-inceptionv3_fp16 | 32 | 2,762.10 | 2,764.27 | -0.08% | :white_check_mark: | |
cadene-inceptionv4 | 16 | 776.64 | 777.02 | -0.05% | :white_check_mark: | |
cadene-resnext64x4 | 16 | 810.72 | 812.24 | -0.19% | :white_check_mark: | |
slim-mobilenet | 64 | 7,532.20 | 7,540.43 | -0.11% | :white_check_mark: | |
slim-nasnetalarge | 64 | 211.49 | 211.44 | 0.02% | :white_check_mark: | |
slim-resnet50v2 | 64 | 3,504.28 | 3,507.16 | -0.08% | :white_check_mark: | |
bert-mrpc-onnx | 8 | 1,146.36 | 1,149.75 | -0.30% | :white_check_mark: | |
bert-mrpc-tf | 1 | 466.04 | 462.59 | 0.75% | :white_check_mark: | |
pytorch-examples-wlang-gru | 1 | 431.37 | 419.12 | 2.92% | :white_check_mark: | |
pytorch-examples-wlang-lstm | 1 | 377.46 | 448.31 | -15.80% | :red_circle: | |
torchvision-resnet50_1 | 1 | 781.10 | 780.35 | 0.10% | :white_check_mark: | |
cadene-dpn92_1 | 1 | 407.36 | 403.86 | 0.87% | :white_check_mark: | |
cadene-resnext101_1 | 1 | 383.45 | 384.99 | -0.40% | :white_check_mark: | |
onnx-taau-downsample | 1 | 343.25 | 343.49 | -0.07% | :white_check_mark: | |
dlrm-criteoterabyte | 1 | 33.34 | 33.36 | -0.05% | :white_check_mark: | |
dlrm-criteoterabyte_fp16 | 1 | 52.74 | 52.73 | 0.03% | :white_check_mark: | |
agentmodel | 1 | 8,466.76 | 8,319.74 | 1.77% | :white_check_mark: | |
unet_fp16 | 2 | 58.85 | 58.90 | -0.08% | :white_check_mark: | |
resnet50v1_fp16 | 1 | 1,020.59 | 941.35 | 8.42% | :high_brightness: | |
resnet50v1_int8 | 1 | 1,002.09 | 995.25 | 0.69% | :white_check_mark: | |
bert_base_cased_fp16 | 64 | 1,170.63 | 1,169.24 | 0.12% | :white_check_mark: | |
bert_large_uncased_fp16 | 32 | 363.56 | 363.49 | 0.02% | :white_check_mark: | |
bert_large_fp16 | 1 | 198.37 | 198.88 | -0.26% | :white_check_mark: | |
distilgpt2_fp16 | 16 | 2,204.73 | 2,201.72 | 0.14% | :white_check_mark: | |
yolov5s | 1 | 541.73 | 533.10 | 1.62% | :white_check_mark: | |
tinyllama | 1 | 43.42 | 43.73 | -0.71% | :white_check_mark: | |
vicuna-fastchat | 1 | 175.58 | 173.66 | 1.10% | :white_check_mark: | |
whisper-tiny-encoder | 1 | 418.69 | 418.99 | -0.07% | :white_check_mark: | |
whisper-tiny-decoder | 1 | 433.64 | 430.97 | 0.62% | :white_check_mark: |
This build is not recommended to merge :red_circle:
:red_circle:bert_large_uncased_fp16: FAILED: MIGraphX is not within tolerance - check verbose output
should this get backported?
I dont think so.
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 92.16%. Comparing base (
6faf926
) to head (a037e2e
). Report is 3 commits behind head on develop.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
@pfultz2 should this get backported?