TDA4VH-Q1: Questions about TIDL output model

Hi,

Could you please elaborate the questions in more depth ?

We need more info like what is use case they are trying ? are they using TIDLRT or OSRT flow ? Which sdk they are using ? 

Please try to elaborate the questions in more depth that will help here

We are using TIDLRT.SDK is  PROCESSOR-SDK-AM69A.

Can you elaborate what do you mean by output here ?

Are you talking about compiled model artifacts binaries or some post processing output dump or anything else ?

Shine said:

2. How to transfer the output Models from Main Domain to MCU Domain?

What you trying to achieve exactly here in case ? and by the term output model what are you referring to ? 

Can you reconfirm you are using TIDLRT flow and not OSRT on linux sdk you mentioned above??

https://software-dl.ti.com/jacinto7/esd/processor-sdk-rtos-jacinto7/08_05_00_11/exports/docs/tidl_j721e_08_05_00_16/ti_dl/docs/user_guide_html/md_tidl_model_import.html

This is documentation for TIDLRT which is supported as part of RTOS SDK.

This is OSRT documentation : https://github.com/TexasInstruments/edgeai-tidl-tools

We are using OSRT.At present, I have deployed the model on EVM, and the fitted lane line curve can be seen on the display. This curve has several parameters, such as an offset distance of 5, and I want to pass this variable to the peripheral. What should I do.

Hi Pratik,

Thank you very much. 

Customer is waiting for our reply. Any suggestions are appreciated.

Thanks
Regards
Shine

Okay so as I see, you are using OSRT flow (edgeai-tidl-tools) to compile and execute the model.

However there are couple of open loop questions whose answers are not clear to me.

Pasting it here again.

1. Which SDK version you are using for OSRT tools ? 

2. As per posted question i can see part number listed as TDA4VH EVM, still not sure are you using AM69A Linux SDK on the same or RTOS SDK ?

3. 

yong xuan said:

his curve has several parameters, such as an offset distance of 5, and I want to pass this variable to the peripheral. What should I do.

As pointed out your model is up and running on EVM, can you throw some light in understanding above statement you have ?