Why my Classification modelnet40 get so low accuracy，請問為何我的分類正確率特別低?

[firecoolman]

2023-01-19 22:17:58,794 - Model - INFO - PARAMETER ... 2023-01-19 22:17:58,794 - Model - INFO - Namespace(batch_size=24, decay_rate=0.0001, epoch=200, gpu='0', learning_rate=0.001, log_dir='pointnet2_cls_ssg2', model='pointnet2_cls_ssg', num_category=40, num_point=1024, optimizer='Adam', process_data=False, use_cpu=False, use_normals=False, use_uniform_sample=False) 2023-01-19 22:17:58,794 - Model - INFO - Load dataset ... 2023-01-19 22:37:44,447 - Model - INFO - No existing model, starting training from scratch... 2023-01-19 22:37:44,448 - Model - INFO - Start training... 2023-01-19 22:37:44,448 - Model - INFO - Epoch 1 (1/200): 2023-01-19 22:54:20,687 - Model - INFO - Train Instance Accuracy: 0.078354 2023-01-19 22:54:55,380 - Model - INFO - Test Instance Accuracy: 0.000000, Class Accuracy: 0.000000 2023-01-19 22:54:55,380 - Model - INFO - Best Instance Accuracy: 0.000000, Class Accuracy: 0.000000 2023-01-19 22:54:55,380 - Model - INFO - Save model... 2023-01-19 22:54:55,380 - Model - INFO - Saving at log\classification\pointnet2_cls_ssg2\checkpoints/best_model.pth 2023-01-19 22:54:55,427 - Model - INFO - Epoch 2 (2/200): 2023-01-19 22:57:17,406 - Model - INFO - Train Instance Accuracy: 0.086484 2023-01-19 22:57:50,877 - Model - INFO - Test Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 22:57:50,879 - Model - INFO - Best Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 22:57:50,879 - Model - INFO - Save model... 2023-01-19 22:57:50,879 - Model - INFO - Saving at log\classification\pointnet2_cls_ssg2\checkpoints/best_model.pth 2023-01-19 22:57:50,930 - Model - INFO - Epoch 3 (3/200): 2023-01-19 23:00:08,223 - Model - INFO - Train Instance Accuracy: 0.086585 2023-01-19 23:00:41,749 - Model - INFO - Test Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:00:41,749 - Model - INFO - Best Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:00:41,750 - Model - INFO - Save model... 2023-01-19 23:00:41,750 - Model - INFO - Saving at log\classification\pointnet2_cls_ssg2\checkpoints/best_model.pth 2023-01-19 23:00:41,790 - Model - INFO - Epoch 4 (4/200): 2023-01-19 23:03:01,451 - Model - INFO - Train Instance Accuracy: 0.087093 2023-01-19 23:03:34,224 - Model - INFO - Test Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:03:34,224 - Model - INFO - Best Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:03:34,224 - Model - INFO - Save model... 2023-01-19 23:03:34,224 - Model - INFO - Saving at log\classification\pointnet2_cls_ssg2\checkpoints/best_model.pth 2023-01-19 23:03:34,263 - Model - INFO - Epoch 5 (5/200): 2023-01-19 23:05:59,614 - Model - INFO - Train Instance Accuracy: 0.088821 2023-01-19 23:06:36,431 - Model - INFO - Test Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:06:36,431 - Model - INFO - Best Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:06:36,432 - Model - INFO - Save model... 2023-01-19 23:06:36,432 - Model - INFO - Saving at log\classification\pointnet2_cls_ssg2\checkpoints/best_model.pth 2023-01-19 23:06:36,467 - Model - INFO - Epoch 6 (6/200): 2023-01-19 23:09:09,582 - Model - INFO - Train Instance Accuracy: 0.087398 2023-01-19 23:09:45,079 - Model - INFO - Test Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:09:45,079 - Model - INFO - Best Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:09:45,079 - Model - INFO - Save model... 2023-01-19 23:09:45,079 - Model - INFO - Saving at log\classification\pointnet2_cls_ssg2\checkpoints/best_model.pth 2023-01-19 23:09:45,111 - Model - INFO - Epoch 7 (7/200): 2023-01-19 23:12:00,355 - Model - INFO - Train Instance Accuracy: 0.088516 2023-01-19 23:12:33,710 - Model - INFO - Test Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:12:33,710 - Model - INFO - Best Instance Accuracy: 0.040453, Class Accuracy: 0.025000 2023-01-19 23:12:33,711 - Model - INFO - Save model... 2023-01-19 23:12:33,711 - Model - INFO - Saving at log\classification\pointnet2_cls_ssg2\checkpoints/best_model.pth 2023-01-20 09:27:24,303 - Model - INFO - Epoch 99 (99/200): Train Instance Accuracy: 0.794715 Test Instance Accuracy: 0.040453, Class Accuracy: 0.025000 Best Instance Accuracy: 0.040453, Class Accuracy: 0.025000

when i train to 99 epoch ,my accuracy still very low .What did I do wrong or what do I need to improve?I am truly grateful for someone tell me.Thank you

[cesaryuan]

Hi, I have the same issue as you. Have you figured it out?

[firecoolman]

Hi, I have the same issue as you. Have you figured it out?

I choose another github to train my own data to achieve classfication 3D model https://github.com/zhulf0804/Pointnet2.PyTorch hope this github can help you

[ludashi451]

我是在Windows平台上出现了这样的情况，我在乌班图系统上没有出现这种情况

[lovepythonhm]

Hi, I have the same problem as you. Have you figured it out?

I chose another github to train my own data to achieve a classified 3D model https://github.com/zhulf0804/Pointnet2.PyTorch hope this github can help you Is this problem solved on the win system now?

[xiximelon]

same problem occurred, is it due to the windows OS?

[liu-jun-chen]

Met the same problem on Windows 11. Does anyone know where the problem is?

[USTBVEO]

same problem. I also used the windows os.

[UCASAlbert]

the same problem!

[USTBVEO]

您发给我的信件已收到。from Vien Yann/杨威

This is an automatic reply, confirming that your e-mail was received. Thank you!

[UCASAlbert]

same problem. I also used the windows os. 你好，贝壳校友。 the same problem.

[UCASAlbert]

Hi, I have the same issue as you. Have you figured it out?

I choose another github to train my own data to achieve classfication 3D model https://github.com/zhulf0804/Pointnet2.PyTorch hope this github can help you

hi, how many minutes dose the training process take? And what is your GPU?