search

experiencor / keras-yolo2

Easy training on custom dataset. Various backends (MobileNet and SqueezeNet) supported. A YOLO demo to detect raccoon run entirely in brower is accessible at https://git.io/vF7vI (not on Windows).
MIT License
1.73k stars 787 forks source link

Precessing Augmentation Error #406

Open thuongtrannc opened 5 years ago

thuongtrannc commented 5 years ago

I met the error when I run the example of Blood Cell Detection

_ValueError: could not convert string to float: 'constant' In processing.py file

ValueError Traceback (most recent call last)

in ----> 1 image = batches[0][0][0][0] 2 plt.imshow(image.astype('uint8')) ~/Th-Folder/5.YOLO/keras-yolo2-master/preprocessing.py in __getitem__(self, idx) 174 for train_instance in self.images[l_bound:r_bound]: 175 # augment input image and fix object's position and size --> 176 img, all_objs = self.aug_image(train_instance, jitter=self.jitter) 177 178 # construct output from object's x, y, w, h ~/Th-Folder/5.YOLO/keras-yolo2-master/preprocessing.py in aug_image(self, train_instance, jitter) 276 if flip > 0.5: image = cv2.flip(image, 1) 277 --> 278 image = self.aug_pipe.augment_image(image) 279 280 # resize the image to standard size Does anyone meet this error? Thank you for any help!