Anton Lokhmotov

Thanks @christ1ne. I had to look up [SKU](https://en.wikipedia.org/wiki/Stock_keeping_unit) in Wikipedia (and donate £5 for its upkeep!) The comment is wrong, but so is `*.Offline.target_qps = 4800` for Dell EMC's CPUs?

@GaryYuyjl I assume you are talking about NVIDIA's open submission with low-precision ResNet50? What's your TensorRT version? /cc @DilipSequeira @nvpohanh

FWIW, once we successfully did the opposite - copied TRT6 libraries from a board with JetPack 4.3 to a board with JetPack 4.2 (with TRT5), and easily switched between the...

Thanks @tjablin. It's fine not to disclose any particular SKU details but I was making the case for being consistent with other submissions e.g. - [Intel](https://github.com/mlperf/inference_results_v0.5/blob/master/closed/Intel/systems/clx_9282-2s_openvino-linux.json#L8) ``` "host_processor_model_name": "Intel(R) Xeon(R)...

Hi @JustinInAI, I've just sent an email to Intel colleagues in charge of OpenVINO. Hope they can help you soon.

@JustinInAI Can you copy some of the errors here please? Also, did you really mean to close this ticket?

@JustinInAI Did @attafosu's advice solve your problem please?

> But the new problem is revise some code from include to include”/boost/...” What do you mean @JustinInAI ? Can you perhaps provide a diff between the original code that...

If I remove the [`--reverse_input_channels` flag](https://github.com/ctuning/ck-openvino/blob/master/package/model-openvino-convert-from-tf/install.sh#L9) from the model conversion, I get a much higher accuracy (20.798%): ``` Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] =...

Using the [other quantized model](https://zenodo.org/record/3252084/) without `--reverse_input_channels` gives worse accuracy (20.150%): ``` Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.201 Average Precision (AP) @[ IoU=0.50...