Description I am building a runtime engine using tensorrt from a .onnx file - YoloV4. It is able to build successfully however, even when i give the workspace 3 GB (3000 in MB in the command), it prints a message while building saying Some tactics do not have sufficient workspace memory to run. In…

Hi @MostafaTheReal , Please refer to the below link for setting up the workspace. [image] Some tactics do not have sufficient workspace memory to run. Increasing workspace size may increase performance, please check verbose output DeepStream SDK Found the problem. The unit…

Hi @AakankshaS , yes i have read this thread, the flag for the trtexec command is workspace not workspace-size, which is also in MB and i have also given it 3GB as mentioned above. That discussion is about the DeepStream SDK. I am asking about TensorRT. The file they referenced to change that paramet…

Hi @MostafaTheReal , I was having the same problem when I tried to port the yolov4 onnx model to TensorRT with trtexec. Using onnx-tensorrt solved the problem for me. At least, that sort of warning message didn’t come up anymore. [Screenshot from 2020-07-22 11-30-34]

Hi @MostafaTheReal , Increasing workspace looks like the only solution. Some TensorRT algorithms require additional workspace on the GPU. Applications should therefore allow the TensorRT builder as much workspace as they can afford; at runtime TensorRT will allocate no more than this, and typically…

@AakankshaS Thanks for your feedback, I’ll update this post if I discover anything else.

@Htut Thanks for letting me know, I’ll give that a try when i have some time.

Hi, @Htut , what does the -d 16 means ?

Hi @MostafaTheReal Do you solve this issue? I’ve used 8GB with converting my model but it’s still give me same errors: Calling isShapeTensor before the entire network is constructed may result in an inaccurate result. also [Some tactics do not have sufficient workspace memory to run. Increasing…

Hi, Please check the below links, as they might answer your concerns. Developer Guide :: NVIDIA Deep Learning TensorRT Documentation This is the revision history of the NVIDIA TensorRT 8.4 Developer Guide. Developer Guide :: NVIDIA Deep Learning Te…

@NVES so what shall we pass as argument for trtexec if getting this err? [TRT] Tactic Device request: 3246MB Available: 1536MB. Device memory is insufficient to use tactic.

Trtexec Yolov4: Some tactics do not have sufficient workspace memory to run

AI & Data Science Deep Learning (Training & Inference) TensorRT

AakankshaS July 21, 2020, 6:45pm 2

Hi @MostafaTheReal,

Please refer to the below link for setting up the workspace.

Thanks!

Topic		Replies	Views
Why dla need so much workspace size? TensorRT tensorrt , jetson-inference	2	876	March 19, 2021
Error - Some tactics do not have sufficient workspace memory to run TAO Toolkit	3	2879	June 30, 2021
How to set workspace in Tensorrt Python API when converting from onnx to engine model TensorRT	4	1647	June 15, 2023
TensorRT engine cannot be built due to workspace size even if it's set higher TensorRT tensorrt	5	6876	January 28, 2022
[TRT] Some tactics do not have sufficient workspace memory to run. Increasing workspace size may increase performance, please check verbose output DeepStream SDK	14	3575	August 9, 2021
Some tactics do not have sufficient workspace memory to run. Increasing workspace size may increase performance, please check verbose output DeepStream SDK	9	17669	November 25, 2020
Trtexec error converting onnx TensorRT onnx	5	1431	August 5, 2022
Can't build TensorRT engine from ONNX due to insufficient memory TensorRT	2	605	January 19, 2024
Errors in converting yolov3 to TensorRT Jetson TX2	8	1412	July 3, 2019
AssertionError: Max workspace size for TensorRT inference should be positive, got 0 TensorRT	4	844	July 21, 2021

Trtexec Yolov4: Some tactics do not have sufficient workspace memory to run

Related topics