Dimensions of inputs should match

[ycdhqzhiai]

训练yolov5报错

Yolov5_tf/core/yolov5.py

Line 215 in ef9b214

y38 = tf.concat([y38, y19_upsample], axis=-1)

[vvtian]

Yes, I have the same problem

[ycdhqzhiai]

@vvtian 可以看看https://github.com/ycdhqzhiai/Yolov5_Tensorflow

[jm5948]

Yes, I have the same problem

have you solved this problem??

[avBuffer]

Hey All, Thanks, I've fixed the error.

[Linaom1214]

Hey All, Thanks, I've fixed the error.

in the new version ,this bug is seemed not fix

[Tracy-git]

i think i found the error.
in train.py line 203 _, summary, train_step_loss, global_step_val = self.sess.run( [train_op, self.write_op, self.loss, self.global_step], feed_dict={self.input_data: train_data[0], self.label_sbbox: train_data[3], self.label_mbbox: train_data[2], self.label_lbbox: train_data[1], self.true_sbboxes: train_data[4], self.true_mbboxes: train_data[5], self.true_lbboxes: train_data[6], self.trainable: True,})

the label_sbbox's index in train_data is 3,but author write 1, and the label_lbbox is 1 but author write3
this is the evidence

but i don't know why i use yolov3 or yolov4 as the net_type , the error above is not happend