Variable FeatureExtractor/darknet_19/base/Conv2D_1/weights does not exist, or was not created with tf.get_variables(). Did you mean to set reuse= tf.AUTO_REUSE in VarScope?

[thanhtu19392]

I had a problem Variable FeatureExtractor/darknet_19/base/Conv2D_1/weights does not exist, or was not created with tf.get_variables(). Did you mean to set reuse= tf.AUTO_REUSE in VarScope? when running. It turns out to end at this line:

https://github.com/GustavZ/yolo_for_tf_od_api/blob/cd312586678724802d9c3e78c7cdc2b5a027a049/darknet.py#L79

[rky0930]

No need to change the line. I work well in my environment. Did you change any config ? or What ckpt are you using ?

[thanhtu19392]

@rky0930 : Thanks for you reply. i haven't changed anything in config file except num_classes. Whatever i tried, i still failed at the same error. I used YOLOv2 coco checkpoint on your repo. Even if i comment out the line of checkpoint, i encountered the same prob

[rky0930]

@thanhtu19392 Could you try again with yolo_v2 in my repo ? That's the full version of the yolo_v2. That use the same source code in this repo but Object detecion api code is included. Please try again and let me know the result.