search

foolwood / SiamMask

[CVPR2019] Fast Online Object Tracking and Segmentation: A Unifying Approach
http://www.robots.ox.ac.uk/~qwang/SiamMask
MIT License
3.47k stars 817 forks source link

tool/train_siammask: val dataset in val_loader didn't use to evaluate model ? #134

Open kaizen0890 opened 4 years ago

kaizen0890 commented 4 years ago

I read train function in train_siammask then saw that have no evaluation step on val dataset. That's why mean_iou, iou_5, iou_7 are not displayed during training time. Can anyone explain for me?

ghost commented 4 years ago

i have too problem with IoU during train, and whan i train refine model, i have mask_loss = 0

[2020-06-12 04:51:55,126-rk0-train_siammask.py#282] Epoch: [1][1800/2500] lr: 0.001000 batch_time: 0.100528 (0.104396) data_time: 0.000167 (0.000895) rpn_cls_loss: 0.020901 (0.088931) rpn_loc_loss: 0.144215 (0.243506) rpn_mask_loss: 0.000000 (0.000000) siammask_loss: 0.193958 (0.381138) mask_iou_mean: 0.000000 (0.000000) mask_iou_at_5: 0.000000 (0.000000) mask_iou_at_7: 0.000000 (0.000000) INFO:global:Epoch: [1][1800/2500] lr: 0.001000 batch_time: 0.100528 (0.104396)data_time: 0.000167 (0.000895) rpn_cls_loss: 0.020901 (0.088931) rpn_loc_loss: 0.144215 (0.243506) rpn_mask_loss: 0.000000 (0.000000) siammask_loss: 0.193958 (0.381138) mask_iou_mean: 0.000000 (0.000000) mask_iou_at_5: 0.000000 (0.000000) mask_iou_at_7: 0.000000 (0.000000) [2020-06-12 04:51:55,126-rk0-log_helper.py# 97] Progress: 1800 / 50000 [3%], Speed: 0.104 s/iter, ETA 0:01:23 (D:H:M)

INFO:global:Progress: 1800 / 50000 [3%], Speed: 0.104 s/iter, ETA 0:01:23 (D:H:M)

[2020-06-12 04:51:56,066-rk0-train_siammask.py#282] Epoch: [1][1810/2500] lr: 0.001000 batch_time: 0.102480 (0.104330) data_time: 0.000134 (0.000891) rpn_cls_loss: 0.018088 (0.088552) rpn_loc_loss: 0.141182 (0.243092) rpn_mask_loss: 0.000000 (0.000000) siammask_loss: 0.187506 (0.380262) mask_iou_mean: 0.000000 (0.000000) mask_iou_at_5: 0.000000 (0.000000) mask_iou_at_7: 0.000000 (0.000000) INFO:global:Epoch: [1][1810/2500] lr: 0.001000 batch_time: 0.102480 (0.104330)data_time: 0.000134 (0.000891) rpn_cls_loss: 0.018088 (0.088552) rpn_loc_loss: 0.141182 (0.243092) rpn_mask_loss: 0.000000 (0.000000) siammask_loss: 0.187506 (0.380262) mask_iou_mean: 0.000000 (0.000000) mask_iou_at_5: 0.000000 (0.000000) mask_iou_at_7: 0.000000 (0.000000) [2020-06-12 04:51:56,067-rk0-log_helper.py# 97] Progress: 1810 / 50000 [3%], Speed: 0.104 s/iter, ETA 0:01:23 (D:H:M)

INFO:global:Progress: 1810 / 50000 [3%], Speed: 0.104 s/iter, ETA 0:01:23 (D:H:M)

[2020-06-12 04:51:57,066-rk0-train_siammask.py#282] Epoch: [1][1820/2500] lr: 0.001000 batch_time: 0.086918 (0.104298) data_time: 0.000094 (0.000887) rpn_cls_loss: 0.023221 (0.088168) rpn_loc_loss: 0.147909 (0.242639) rpn_mask_loss: 0.000000 (0.000000) siammask_loss: 0.200712 (0.379335) mask_iou_mean: 0.000000 (0.000000) mask_iou_at_5: 0.000000 (0.000000) mask_iou_at_7: 0.000000 (0.000000) INFO:global:Epoch: [1][1820/2500] lr: 0.001000 batch_time: 0.086918 (0.104298)data_time: 0.000094 (0.000887) rpn_cls_loss: 0.023221 (0.088168) rpn_loc_loss: 0.147909 (0.242639) rpn_mask_loss: 0.000000 (0.000000) siammask_loss: 0.200712 (0.379335) mask_iou_mean: 0.000000 (0.000000) mask_iou_at_5: 0.000000 (0.000000) mask_iou_at_7: 0.000000 (0.000000) [2020-06-12 04:51:57,066-rk0-log_helper.py# 97] Progress: 1820 / 50000 [3%], Speed: 0.104 s/iter, ETA 0:01:23 (D:H:M)

INFO:global:Progress: 1820 / 50000 [3%], Speed: 0.104 s/iter, ETA 0:01:23 (D:H:M)