Jetson_benchmarks got Error opening engine file

Hi All,
I run jetson_benchmark in Orin Nano 4G.
I got 0 FPS using the command

sudo python3 benchmark.py --model_name inception_v4 --csv_file_path home/orin_nano/jetson_benchmarks/benchmark_csv/orin-nano-benchmarks.csv --model_dir /home/orin_nano/jetson_benchmarks/models/ --jetson_clocks

I check the log in modules file, I got “Error opening engine file”.
I have already tried to change the inception_v4 WS_GPU to 1024 and 512 but it got the same result.
Could you give some suggestion?
Thanks
inception_v4_b32_ws512_gpu.txt (4.8 KB)

Hi,

The error indicates that it cannot find the TensorRT engine.
Have you applied the setup command below?

Thanks.

Yes, I have already follow the" Set up instructions" and " Install Requirements".

Hi,

Could you check if the download_models.py save all the required model successfully?
For example, we save the model to $(pwd)/models with the below command:

$ python3 utils/download_models.py --all --csv_file_path benchmark_csv/orin-nano-benchmarks.csv --save_dir $(pwd)/models

Within $(pwd)/models, we can find all the models are downloaded:

$ ll models/
total 210996
drwxrwxr-x 2 nvidia nvidia     4096 Sep  4 05:39 ./
drwxrwxr-x 6 nvidia nvidia     4096 Sep  4 05:37 ../
-rw-rw-r-- 1 nvidia nvidia   120973 Sep  4 05:38 inception_v4.prototxt
-rw-rw-r-- 1 nvidia nvidia    22561 Sep  4 05:39 mobilenet_v1_ssd.prototxt
-rw-rw-r-- 1 nvidia nvidia    45979 Sep  4 05:39 pose_estimation.prototxt
-rw-rw-r-- 1 nvidia nvidia    35016 Sep  4 05:39 ResNet50_224x224.prototxt
-rw-rw-r-- 1 nvidia nvidia    32058 Sep  4 05:39 ssd_resnet34_1200x1200.prototxt
-rw-rw-r-- 1 nvidia nvidia   273491 Mar 12  2020 super_resolution_bsd500-bs1.onnx
-rw-rw-r-- 1 nvidia nvidia   273491 Mar 12  2020 super_resolution_bsd500-bs2.onnx
-rw-rw-r-- 1 nvidia nvidia   273491 Mar 12  2020 super_resolution_bsd500-bs4.onnx
-rw-rw-r-- 1 nvidia nvidia   273491 Mar 12  2020 super_resolution_bsd500-bs8.onnx
-rw-rw-r-- 1 nvidia nvidia  1971389 Sep  4 05:39 unet-segmentation.uff
-rw-rw-r-- 1 nvidia nvidia     5936 Sep  4 05:38 vgg19_N2.prototxt
-rw-rw-r-- 1 nvidia nvidia 35448300 Sep 28  2022 yolov3-tiny-416-bs16.onnx
-rw-rw-r-- 1 nvidia nvidia 35448300 Sep 28  2022 yolov3-tiny-416-bs1.onnx
-rw-rw-r-- 1 nvidia nvidia 35448300 Sep 28  2022 yolov3-tiny-416-bs2.onnx
-rw-rw-r-- 1 nvidia nvidia 35448628 Sep 28  2022 yolov3-tiny-416-bs32.onnx
-rw-rw-r-- 1 nvidia nvidia 35448300 Sep 28  2022 yolov3-tiny-416-bs4.onnx
-rw-rw-r-- 1 nvidia nvidia 35447411 Sep 28  2022 yolov3-tiny-416-bs8.onnx

Thanks.

HI,
I checked $(pwd)/models,

$ ll models/
total 210996
drwxr-xr-x 2 orin_nano orin_nano     4096  九   5 17:15 ./
drwxr-xr-x 6 root      root          4096  八  31 13:42 ../
-rw-rw-r-- 1 orin_nano orin_nano   120973  九   5 17:13 inception_v4.prototxt
-rw-rw-r-- 1 orin_nano orin_nano    22561  九   5 17:15 mobilenet_v1_ssd.prototxt
-rw-rw-r-- 1 orin_nano orin_nano    45979  九   5 17:13 pose_estimation.prototxt
-rw-rw-r-- 1 orin_nano orin_nano    35016  九   5 17:15 ResNet50_224x224.prototxt
-rw-rw-r-- 1 orin_nano orin_nano    32058  九   5 17:15 ssd_resnet34_1200x1200.prototxt
-rw-rw-r-- 1 orin_nano orin_nano   273491  三  12  2020 super_resolution_bsd500-bs1.onnx
-rw-rw-r-- 1 orin_nano orin_nano   273491  三  12  2020 super_resolution_bsd500-bs2.onnx
-rw-rw-r-- 1 orin_nano orin_nano   273491  三  12  2020 super_resolution_bsd500-bs4.onnx
-rw-rw-r-- 1 orin_nano orin_nano   273491  三  12  2020 super_resolution_bsd500-bs8.onnx
-rw-rw-r-- 1 orin_nano orin_nano  1971389  九   5 17:13 unet-segmentation.uff
-rw-rw-r-- 1 orin_nano orin_nano     5936  九   5 17:13 vgg19_N2.prototxt
-rw-rw-r-- 1 orin_nano orin_nano 35448300  九  28  2022 yolov3-tiny-416-bs16.onnx
-rw-rw-r-- 1 orin_nano orin_nano 35448300  九  28  2022 yolov3-tiny-416-bs1.onnx
-rw-rw-r-- 1 orin_nano orin_nano 35448300  九  28  2022 yolov3-tiny-416-bs2.onnx
-rw-rw-r-- 1 orin_nano orin_nano 35448628  九  28  2022 yolov3-tiny-416-bs32.onnx
-rw-rw-r-- 1 orin_nano orin_nano 35448300  九  28  2022 yolov3-tiny-416-bs4.onnx
-rw-rw-r-- 1 orin_nano orin_nano 35447411  九  28  2022 yolov3-tiny-416-bs8.onnx

the result is same as yours but run $ sudo python3 benchmark.py --all --csv_file_path /home/orin_nano/jetson_benchmarks/benchmark_csv/orin-nano-benchmarks.csv --model_dir /home/orin_nano/jetson_benchmarks/models/ --jetson_clocks , it still got 0 FPS and Error in Build,the log is still the same.

Hi,

Thanks for the info.

We will check it further.
Will get back to you soon.

Hi,

We can get the inception_v4 result.
Could you double-check it again?

$ sudo python3 benchmark.py --all --csv_file_path benchmark_csv/orin-nano-benchmarks.csv --model_dir $(pwd)/models --jetson_clocks
Please close all other applications and Press Enter to continue...
Setting Jetson orin in max performance mode
Jetson clocks are Set
Running all benchmarks.. This will take at least 2 hours...
------------Executing inception_v4------------


---------------------- 89.2536105263158 0 0
--------------------------

Model Name: inception_v4 
FPS:358.53 

--------------------------
...

Thanks.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.