search

youchangxin / YOLOv4_tensorflow2

YOLOv4 Implemented in Tensorflow 2.1，which supports training on custom dataset.
MIT License
25 stars 5 forks source link

the loss is nan #2

Closed zcdliuwei closed 4 years ago

zcdliuwei commented 4 years ago
image
youchangxin commented 4 years ago

I will fix the Bug as soon as possible, I'm testing the new coed to save it .

zcdliuwei commented 4 years ago

I want to ask for your advice. I want to identify the bottles in the trash can. The real image is about this:

20200611112928_B_Q

20200611112934_B_Q

20200611113139_B_Q

The most complex part of this scene is the dense shielding and stacking of bottles. Our purpose is to tell the customer who threw the bottle what kind of bottle he threw in, and it's better to return the location information of the bottle at the bottom of the trash can. At present, I use Yolo series for testing, and the effect is always unsatisfactory.

I want to know does you have any suggestions to improve the detection and classification ability in this scenario?

Much obliged for your assistance.

msergenergen commented 3 years ago

Hi, Firstly, thanks for your repo creation. Did you fix this bug? I still get nan, you can see output following output.

=>EPOCH 0 STEP 241 lr: 0.000068 ciou_loss: 9.67 conf_loss: 11385.40 prob_loss: 84.94 total_loss: 11480.01 =>EPOCH 0 STEP 242 lr: 0.000068 ciou_loss: 1.82 conf_loss: 11522.95 prob_loss: 16.12 total_loss: 11540.89 =>EPOCH 0 STEP 243 lr: 0.000069 ciou_loss: 3.79 conf_loss: 11352.79 prob_loss: 33.10 total_loss: 11389.68 =>EPOCH 0 STEP 244 lr: 0.000069 ciou_loss: 1.77 conf_loss: 11479.06 prob_loss: 17.01 total_loss: 11497.84 =>EPOCH 0 STEP 245 lr: 0.000069 ciou_loss: 1.86 conf_loss: 11647.79 prob_loss: 15.82 total_loss: 11665.48 =>EPOCH 0 STEP 246 lr: 0.000069 ciou_loss: 1.88 conf_loss: 11362.14 prob_loss: 16.45 total_loss: 11380.47 =>EPOCH 0 STEP 247 lr: 0.000070 ciou_loss: nan conf_loss: nan prob_loss: nan total_loss: nan =>EPOCH 0 STEP 248 lr: 0.000070 ciou_loss: nan conf_loss: nan prob_loss: nan total_loss: nan =>EPOCH 0 STEP 249 lr: 0.000070 ciou_loss: nan conf_loss: nan prob_loss: nan total_loss: nan =>EPOCH 0 STEP 250 lr: 0.000071 ciou_loss: nan conf_loss: nan prob_loss: nan total_loss: nan =>EPOCH 0 STEP 251 lr: 0.000071 ciou_loss: nan conf_loss: nan prob_loss: nan total_loss: nan =>EPOCH 0 STEP 252 lr: 0.000071 ciou_loss: nan conf_loss: nan prob_loss: nan total_loss: nan =>EPOCH 0 STEP 253 lr: 0.000071 ciou_loss: nan conf_loss: nan prob_loss: nan total_loss: nan