When using docker-based model fixed-point tool, what is the specific format of calib.txt file in the quantization process of caffe's target detection model? about vitis-ai HOT 3 CLOSED

Hi @tensorflowt ,

According to your output, I think the path to your images are not correctly configured. Let take files in "https://github.com/Xilinx/Vitis-AI/tree/master/mpsoc/vitis-ai-tool-example" as example.

In this example, we assume "vai_q_caffe" is being run under path of "vitis-ai-tool-example".

The tool will use the prototxt file in "cf_resnet50" folder and "list.txt" in "images" folder, with configuration as below:

In this correct configuration, tool will find images under folder "images/" (make sure '/' is not missing) with name specified in "list.txt". Taking first image as example, vai_q_caffe will find "./images/ILSVRC2012_val_00000001.JPEG" and its lable “65” according to the "list.txt".

Please check if you have set the path correctly.

from vitis-ai.

Hi @tensorflowt
I wonder if you have solved this issue. If not, I will submit this issue to the engineer in charge of it. If it has been resolved, please close this issue.
Note that if we do not receive any reply from you within 2 weeks, we will assume that this issue has been resolved and will close this issue. Thank you.

from vitis-ai.

Hi @tensorflowt
Since I haven't received your reply for a long time, I will close this issue. If you still have any questions, you can reopen this issue. Thank you.

from vitis-ai.