yolo-tf
yolo-tf copied to clipboard
Question on training Yolo v1
Hi @ruiminshen, thanks for the project.
Could you show me how should I train for the yolo v1 (Tiny)? Do I need to change the config.ini file? (I already downloaded the datasets VOC)
Thanks
What I'm trying to do....
- Download VOC 2007/2012, and extract at: Documents/Database/VOCdevki
- python3 cache.py -c config.ini config/yolo/tiny-20.ini -v
- python train.py -c config.ini config/yolo/tiny-20.ini
Hi @ruiminshen I think is working
Btw I'm documenting my findings about Yolo here: https://leonardoaraujosantos.gitbooks.io/artificial-inteligence/content/single-shot-detectors/yolo.html
Do you accept a pull request for updating the Readme file?
@leonardoaraujosantos Sorry, I'm a little busy these days and I didn't noticed your message.
Your operation is correct. To switch between models (YOLOv1 and YOLOv2), you can change the option "model" in "config" section in the default configuration file "config.ini", and the -c argument is used to override the options by another configuration file.
When I trained it by following leonardoaraujosantos, I got this problem. WARNING:tensorflow:Error encountered when serializing LAYER_NAME_UIDS. Type is unsupported, or the types of the items don't match field type in CollectionDef. 'dict' object has no attribute 'name'
@leonardoaraujosantos @xhzcyc @ruiminshen
Do you successfully train for YOLOv2 with VOC? And could you share your trained model? I try to train yolo2 using yolo-tf sevral times, but I always failed. @leonardoaraujosantos @xhzcyc
@TaihuLight hi? can you get a accurate resuklt with yolo2?
@zhyj3038 I can get the mAP with YOLO2, but multi-scale trianing and training with own dataset are need to be finished!
@TaihuLight Hello, have you successfully implemented the test? Is the effect good?