datlife / yolov2

YOLOv2 Implementation in TF/Keras. Allowing to experiment on different feature detectors (MobileNet, Darknet-19). Paper: https://arxiv.org/abs/1612.08242
49 stars 13 forks source link

Unable to load weights file #2

Closed oishi89 closed 7 years ago

oishi89 commented 7 years ago

Hi there I got an issue related to load_weights during training my custom data. I was using the weight file downloaded from https://drive.google.com/drive/folders/0ByoFGh573uhzWGx0aXNNNkY5Qms

screen shot 2017-09-21 at 8 45 54 pm
datlife commented 7 years ago

May I see how you set up the cfg.py?

oishi89 commented 7 years ago

Hi please take a look ` FEATURE_EXTRACTOR = 'mobilenet'

CATEGORIES = "/Users/ln160c/Documents/MLWorkspace/yolov2-master/dataset/my_new_dataset/categories.txt ANCHORS = "/Users/ln160c/Documents/MLWorkspace/yolov2-master/dataset/my_new_dataset/anchors.txt"

ENABLE_HIERARCHICAL_TREE = False HIERARCHICAL_TREE_PATH = "/home/dat/Documents/yolov2/dataset/combined_lisa/lisa.tree"

IMG_INPUT_SIZE = 608

N_CLASSES = 1

N_ANCHORS = 5

SHRINK_FACTOR = 32 `

oishi89 commented 7 years ago

My bad. I changed the FEATURE_EXTRACTOR = 'mobilenet' but I was using yolo-coco.weight