mirrors/mmdeploy
1
0
Fork 0
mirror of https://github.com/open-mmlab/mmdeploy.git synced 2025-01-14 08:09:43 +08:00
Code Issues Projects Releases Wiki Activity
mmdeploy/docs/en/03-benchmark/benchmark_tvm.md
q.yao 7cb4b9b18a
[Enhancement] Support tvm (#1216) 
* finish framework

* add autotvm and auto-scheduler tuner

* add python deploy api

* add SDK net(WIP

* add sdk support

* support det, support vm

* fix vm sdk

* support two stage detector

* add instance seg support

* add docstring

* update docs and ut

* add quantize

* update doc

* update docs

* synchronize stream

* support dlpack

* remove submodule

* fix stride

* add alignment

* support dlpack

* remove submodule

* replace exclusive_scan

* add backend check

* add build script

* fix comment

* add ci

* fix ci

* ci fix2

* update build script

* update ci

* add pytest

* update sed command

* update sed again

* add xgboost

* remove tvm ut

* update ansor runner

* add stream sync

* fix topk

* sync default stream

* fix tvm net

* fix window
2022-12-12 21:19:40 +08:00

5.2 KiB
Raw Blame History

Test on TVM

Supported Models

Model Codebase Model config
RetinaNet MMDetection config
Faster R-CNN MMDetection config
YOLOv3 MMDetection config
YOLOX MMDetection config
Mask R-CNN MMDetection config
SSD MMDetection config
ResNet MMClassification config
ResNeXt MMClassification config
SE-ResNet MMClassification config
MobileNetV2 MMClassification config
ShuffleNetV1 MMClassification config
ShuffleNetV2 MMClassification config
VisionTransformer MMClassification config
FCN MMSegmentation config
PSPNet MMSegmentation config
DeepLabV3 MMSegmentation config
DeepLabV3+ MMSegmentation config
UNet MMSegmentation config

The table above list the models that we have tested. Models not listed on the table might still be able to converted. Please have a try.

Test

  • Ubuntu 20.04
  • tvm 0.9.0
mmcls metric PyTorch TVM
ResNet-18 top-1 69.90 69.90
ResNeXt-50 top-1 77.90 77.90
ShuffleNet V2 top-1 69.55 69.55
MobileNet V2 top-1 71.86 71.86
mmdet(*) metric PyTorch TVM
SSD box AP 25.5 25.5

*: We only test model on ssd since dynamic shape is not supported for now.

mmseg metric PyTorch TVM
FCN mIoU 72.25 72.36
PSPNet mIoU 78.55 77.90