srinidhigoud opened a new issue #8561: URL: https://github.com/apache/tvm/issues/8561
This is a follow up to the PR #8174 by @masahi. Some of the tf2-tvm object detection (OD) models performed poorer than just tf2 framework. Following are the numbers observed. TF-TVM latency is on models that are post processed with combined-nms Model | TF-TVM Latency(ms) | TF latency (ms) | Boxes Shape | Scores Shape -|-|-|-|- ssd_mobilenet_v1_fpn_640x640_1_nms | 377.20 | 203.31 | (1, 51150, 1, 4) | (1, 51150, 90) efficientdet_d0_1_nms | 333.01 | 140.92 | (1, 49104, 1, 4) | (1, 49104, 90) GPU activities for *ssd_mobilenet_v1_fpn_640x640_1_nms* ``` "GPU activities",71.961571,2.649234,10,264.923361,256.370119,274.001845,"fused_vision_all_class_non_max_suppression_kernel2" "GPU activities",9.830572,0.361908,160,2.261926,0.870473,7.882956,"volta_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1" "GPU activities",2.893616,0.106527,70,1.521816,0.068355,7.010179,"trt_volta_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1" ``` GPU activities for *efficientdet_d0_1_nms* ``` "GPU activities",79.611974,2.520980,10,252.097991,246.456348,261.662513,"fused_vision_all_class_non_max_suppression_kernel2" "GPU activities",3.295807,0.104364,160,0.652278,0.038498,5.779056,"void cuReduceLayer::nonTailReduceExcludeW<int=32, nvinfer1::ReduceOp, float, float>(float*, cuReduceLayer::nonTailReduceExcludeW<int=32, nvinfer1::ReduceOp, float, float> const *, cuReduceLayer::LaunchParams)" "GPU activities",1.695395,0.053686,40,1.342152,1.012731,1.530201,"void thrust::cuda_cub::cub::DeviceRadixSortDownsweepKernel<thrust::cuda_cub::cub::DeviceRadixSortPolicy<float, long, int>::Policy700, bool=0, bool=1, float, long, int>(thrust::cuda_cub::cub::DeviceRadixSortPolicy<float, long, int>::Policy700 const *, thrust::cuda_cub::cub::DeviceRadixSortDownsweepKernel<thrust::cuda_cub::cub::DeviceRadixSortPolicy<float, long, int>::Policy700, bool=0, bool=1, float, long, int>*, bool=0 const *, thrust::cuda_cub::cub::DeviceRadixSortDownsweepKernel<thrust::cuda_cub::cub::DeviceRadixSortPolicy<float, long, int>::Policy700, bool=0, bool=1, float, long, int>**, bool=1*, thrust::cuda_cub::cub::DeviceRadixSortDownsweepKernel<thrust::cuda_cub::cub::DeviceRadixSortPolicy<float, long, int>::Policy700, bool=0, bool=1, float, long, int>**, int, int, thrust::cuda_cub::cub::GridEvenShare<thrust::cuda_cub::cub::DeviceRadixSortDownsweepKernel<thrust::cuda_cub::cub::DeviceRadixSortPolicy<float, lon g, int>::Policy700, bool=0, bool=1, float, long, int>**>)" ``` Did you encounter these input shapes before and notice the same performance issues, do you think there is anything different we can do to optimize this input shape case? @masahi -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
