Can u share yolov5 trained model and artifacts sk-tda4vm (sdk 8.6) for test purpose
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Can u share yolov5 trained model and artifacts sk-tda4vm (sdk 8.6) for test purpose
Hi,
You can check pretrained model here : https://github.com/TexasInstruments/edgeai-yolov5/tree/master/pretrained_models/models/detection/coco/edgeai-yolov5
Compiled model artifacts for yoloV5 are not available for 8.6 release, but you generate them using edgeai-benchmark repo, refer link here , https://github.com/TexasInstruments/edgeai-benchmark/blob/master/docs/custom_models.md#compile-custom-models
Regards,
Pratik
I tried to compile but i got failed to run on board. Can u share model and aritifacts for sdk 08.05.00.03 (22 Dec 2022) or any version if possible.
Hi,
You can find compiled model artifacts for earlier releases in above mentioned link.
Adding folder link for your reference.
https://github.com/TexasInstruments/edgeai-yolov5/tree/master/pretrained_models/modelartifacts/8bits
Regards,
Pratik
Let me reiterate to the above mentioned steps again for you better understanding , as its look like posted response is not clear here.
You can find compiled model artifacts for earlier releases in above mentioned link.
Adding folder link for your reference.
https://github.com/TexasInstruments/edgeai-yolov5/tree/master/pretrained_models/modelartifacts/8bits
If you click on link files for above link, you can find version mentioned there.
You can choose any model based on your requirement.
Here is the example,
In this link sdk version is 8.2
Regards,
Pratik
In 8.2 sdk i got an error when running ./app_edgeai.py ../configs/object_detection.yaml
[ERROR] no element "tiovxdlcolorconvert"
[GST STR] appsrc format=GST_FORMAT_TIME block=true num-buffers=1 name=background_0 ! tiovxdlcolorconvert ! video/x-raw,format=NV12, width=1920, height=1080 ! filesink location=/tmp/background_0
I think only SDK works above 8.4
Can u share model and artifacts for 8.5 if possibile. I urgently need for test evaluation this week
Kind Regards
N
In 8.2 sdk i got an error when running ./app_edgeai.py ../configs/object_detection.yaml
Could you explain steps you have followed after you flashed 8.2 linux sdk till you invoke above run command ?
Can u share model and artifacts for 8.5 if possibile.
Reiterating,
Let me reiterate to the above mentioned steps again for you better understanding , as its look like posted response is not clear here.
You can find compiled model artifacts for earlier releases in above mentioned link.
Adding folder link for your reference.
https://github.com/TexasInstruments/edgeai-yolov5/tree/master/pretrained_models/modelartifacts/8bits
Please check the all model artifacts available here, you can choose based on your desired SDK if its available in above link.
Regards,
Pratik
Only model and artifacts are available for 8.2 .When i deployed on 8.2 sdk image iam getting this error
In 8.2 sdk i got an error when running ./app_edgeai.py ../configs/object_detection.yaml
[ERROR] no element "tiovxdlcolorconvert"
[GST STR] appsrc format=GST_FORMAT_TIME block=true num-buffers=1 name=background_0 ! tiovxdlcolorconvert ! video/x-raw,format=NV12, width=1920, height=1080 ! filesink location=/tmp/background_0
When i tried to download the 8.5 version model from https://github.com/TexasInstruments/edgeai-yolov5/tree/master/pretrained_models/modelartifacts/8bits
Iam unable to download the file- http://software-dl.ti.com/jacinto7/esd/modelzoo/gplv3/08_05_00_01/edgeai-yolov5/pretrained_models/modelartifacts/8bits/od-8110_onnxrt_weights_yolov5s6_384_ti_lite_32p8_51p2_onnx.tar.gz
Hi,
I have created internal change request for updating the links for 8.5 sdk.
Adding JIRA link for TI's internal purpose :
https://jira.itg.ti.com/browse/EDGEAI_ALGO-39
We will update you once this fix is being added.
Regarding this issue,
In 8.2 sdk i got an error when running ./app_edgeai.py ../configs/object_detection.yaml
[ERROR] no element "tiovxdlcolorconvert"
[GST STR] appsrc format=GST_FORMAT_TIME block=true num-buffers=1 name=background_0 ! tiovxdlcolorconvert ! video/x-raw,format=NV12, width=1920, height=1080 ! filesink location=/tmp/background_0
We already have open thread under discussion, please refer here https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1237336/sk-tda4vm-run-inference-on-sdk-v-08-02-00-02-26-apr-2022
Closing this thread as of now, will update once we have fix added on 8.5 compiled model artifacts.
Regards,
Pratik