Where can download model file for ssd-resnet34 and Bert-Large

Autonomous Machines Jetson & Embedded Systems Jetson Xavier NX
1

Problem found:

Found Jetson Benchmark results posted are consist of SSD-Resnet34 and Bert-Large model and there is URL shared Jetson Benchmarks | NVIDIA Developer. But there is no SSD-Resnet34 & Bert-Large model been described and where to download. Can I know where I can have the model files download?

My Environment and running with Jetson NX.

  • NVIDIA Jetson Xavier NX (Developer Kit Version)
    • Jetpack 4.6 [L4T 32.6.1]
    • NV Power Mode: MODE_20W_6CORE - Type: 8
    • jetson_stats.service: active
  • Board info:
    • Type: Xavier NX (Developer Kit Version)
    • SOC Family: tegra194 - ID:25
    • Module: P3668 - Board: P3509-000
    • Code Name: jakku
    • CUDA GPU architecture (ARCH_BIN): 7.2
    • Serial Number: 1424620042014
  • Libraries:
    • CUDA: 10.2.300
    • cuDNN: 8.2.1.32
    • TensorRT: 8.0.1.6
    • Visionworks: 1.6.0.501
    • OpenCV: 4.1.1 compiled CUDA: NO
    • VPI: ii libnvvpi1 1.1.11 arm64 NVIDIA Vision Programming Interface library
    • Vulkan: 1.2.70
  • jetson-stats:
    • Version 3.1.1
    • Works on Python 3.6.9
2

Hi,
This looks like a Jetson issue. Please refer to the below samlples in case useful.

For any further assistance, we recommend you to raise it to the respective platform from the below link

Thanks!

4

Hi NVES,
Thanks for your information. I did research the repo before and there is no SSD-Resnet34 or Bert-Large model as well. By the way, I tried and cannot use the models files from MLPerf also, so would like to know where to get the model files since there is result posted, thank you.

5

Hi,

Hope you following link will be helpful to you.

Thank you.

6

Hi Spolisetty,
Thanks your information, I found Bert models there but pending on SSD-Resnet34, thank you.

7

Hi,

Currently we are not sure about it, could you please post your concern on Jetson forum as mentioned above to get better help.

Thank you.

8

Hi spolisetty,
As above Topic I mention, Nvidia do published the SSD-Resnet34 result (https://developer.nvidia.com/embedded/jetson-benchmarks), but we cannot find the models on the repo shared. Since there is result published and should there some way we can get the ssd-resnet34 model file to trtexec workload.

image
image1483×260 7.23 KB

9

Moving this topic to the Jetson Xavier NX forum.

@boon.hai.ng, please refer to MLPerf inference results for benchmarking the models:

10

Hi dusty_nv,
Thanks for your information. Which means the scores was mixed either performance result ran from trtexec and Mlperf. It’s confuse to me and keep looking where I can get the model run in trtrexec whereby the page only attached jetson-benchmark repo (used trtexec), thank you.

11

Yes, the models that are not included in jetson_benchmarks repo were run with MLPerf.