Have difficulty in fixing KeyError: 'layer1-conv'

[b154012]

First of all, I would like to apologize if this post is related to the previous problem. However, I am still stuck in fixing the following errors:

I typed the command:

!python darknet2caffe.py yolov3_testing.cfg yolov3_training_final.weights yolov3.prototxt yolov3.caffemodel

The following link shows the error message:

https://drive.google.com/file/d/1lXmA2HfnQSIBoLlrOXtIkwCb1xq8fxwA/view

I have modified the darknet2caffe.py file. I changed the has_key() function to the contain() function . It can be reviewed here:

https://drive.google.com/file/d/1joxRt-vnKV3HY0CfDOmoxMrCxOC82Jl2/view?usp=sharing

I used python 2.7, caffe and pytorch (1.7) > 0.4, or 3.6.

the Makefile example file:

https://drive.google.com/file/d/1bBlv_2V3RZ45EEG-TwzuJz072bGu8FUf/view?usp=sharing

How to solve this issue?

Perhaps my cfg and weights have some bugs? However, I already tested it using Darknett and was able to predict my classes.

I desperately need this conversion so that it can be run on Raspberry Pi for real-time object detection. I could provide my custom YOLOv3 files for you if the files are fine. Perhaps I am missing something?

I highly appreciate any help or clarification.

[ChenYingpeng]

@b154012 Please see this link.

[b154012]

@b154012 Please see this link.

Hi. I have already fixed the prototxt.py file and made minor modifications of darknet2caffe.py. I am currently using Google Colab with Python 3.6, Cuda 10 and pytorch > 0.4. However, I have been doing it for hours and days and still cannot fix it. I honestly found it difficult/tricky for some reason.

Therefore, if I may make a request, I desperately need my own custom YOLO model to be converted into Caffe. This model will be used for performing real-time object detection on the Raspberry Pi B+. Yolov3 is known for its slower performance on Raspberry Pi.

I do not mind to pay you online. I have WeChat.

Would it be okay for you?
If you may, I will provide my WeChat.

The folder link can be found here:

https://drive.google.com/drive/folders/1KV2ErZfTUMXKiAfbf2vqvujwPOYaRY4L?usp=sharing.

It includes *.cfg, *.weight and classes.txt.

Thank you for your consideration. I greatly appreciate your help.

[b154012]

@b154012 Please see this link.

Hi, would it possible to help convert my data?

https://drive.google.com/drive/folders/1KV2ErZfTUMXKiAfbf2vqvujwPOYaRY4L?usp=sharing.

My WeChat id is wxid_1rn83ihu0mo722

I greatly appreciate your help and consideration. Thank you