Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

how the final epoic has higher mAP and the final weights not #11889

Closed
1 task done
VYRION-Ai opened this issue Jul 22, 2023 · 6 comments
Closed
1 task done

how the final epoic has higher mAP and the final weights not #11889

VYRION-Ai opened this issue Jul 22, 2023 · 6 comments
Labels
question Further information is requested

Comments

@VYRION-Ai
Copy link

Search before asking

Question

i do traning on yolov5 and i got this results , now please how the final epoic has higher mAP and the final weights not
look to this image
Untitled

`Starting training for 100 epochs...

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   0/99      22.1G    0.09157    0.03945    0.02419        209        640: 100% 162/162 [01:06<00:00,  2.44it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:19<00:00,  2.44s/it]
               all       1952       3162     0.0995      0.196     0.0628     0.0155

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   1/99      34.1G    0.07464    0.03621     0.0163        163        640: 100% 162/162 [00:58<00:00,  2.75it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.45s/it]
               all       1952       3162      0.558      0.413      0.426      0.153

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   2/99      34.1G    0.06802    0.03329    0.01293        198        640: 100% 162/162 [00:59<00:00,  2.73it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.46s/it]
               all       1952       3162      0.595      0.498      0.517      0.188

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   3/99      34.1G    0.06154    0.03088    0.01093        262        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.48s/it]
               all       1952       3162      0.677      0.601      0.608      0.262

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   4/99      34.1G    0.05764    0.02958   0.009453        215        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.46s/it]
               all       1952       3162      0.747      0.651      0.722      0.338

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   5/99      34.1G    0.05449     0.0285   0.008766        221        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.863       0.72      0.776      0.371

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   6/99      34.1G    0.05278     0.0276   0.008268        301        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.48s/it]
               all       1952       3162      0.844      0.692      0.758       0.36

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   7/99      34.1G    0.05147    0.02747   0.007919        212        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.41s/it]
               all       1952       3162      0.905      0.758      0.812      0.345

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   8/99      34.1G    0.05081    0.02725   0.007376        199        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.42s/it]
               all       1952       3162      0.896      0.753      0.815      0.382

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
   9/99      34.1G    0.04917    0.02647   0.007279        157        640: 100% 162/162 [00:57<00:00,  2.80it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.873      0.746      0.804      0.364

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  10/99      34.1G    0.04807    0.02624    0.00717        147        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.907      0.794      0.842      0.453

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  11/99      34.1G    0.04784    0.02598   0.006845        480        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162      0.884      0.784      0.844      0.432

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  12/99      34.1G    0.04713    0.02546   0.006737        198        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.904      0.782      0.842      0.448

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  13/99      34.1G    0.04629    0.02528   0.006573        190        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.42s/it]
               all       1952       3162      0.919      0.807      0.854      0.484

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  14/99      34.1G    0.04604    0.02506   0.006433        144        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.925      0.813      0.859      0.459

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  15/99      34.1G    0.04575    0.02557   0.006384        221        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.918       0.82       0.86      0.467

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  16/99      34.1G    0.04511    0.02492   0.006083        210        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.35s/it]
               all       1952       3162      0.931      0.816      0.863      0.461

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  17/99      34.1G    0.04563    0.02494   0.005966        379        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.928      0.806      0.854      0.467

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  18/99      34.1G     0.0446     0.0248   0.005933        219        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.939      0.818      0.869      0.469

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  19/99      34.1G    0.04434    0.02487   0.005759        203        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.43s/it]
               all       1952       3162      0.941      0.819      0.875      0.498

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  20/99      34.1G    0.04467    0.02424   0.005718        288        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.944      0.819      0.876      0.507

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  21/99      34.1G    0.04403    0.02435   0.005735        241        640: 100% 162/162 [00:57<00:00,  2.80it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.42s/it]
               all       1952       3162      0.918      0.826       0.87      0.489

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  22/99      34.1G    0.04384    0.02428   0.005568        302        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.41s/it]
               all       1952       3162      0.918      0.829      0.871      0.496

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  23/99      34.1G    0.04402    0.02421   0.005417        284        640: 100% 162/162 [00:57<00:00,  2.80it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162      0.936      0.824      0.873       0.51

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  24/99      34.1G    0.04355    0.02419   0.005475        217        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.35s/it]
               all       1952       3162      0.948      0.823      0.874      0.517

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  25/99      34.1G    0.04267    0.02377   0.005462        194        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162      0.939      0.837      0.881      0.512

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  26/99      34.1G     0.0431    0.02382   0.005275        241        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162      0.949      0.828      0.883      0.523

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  27/99      34.1G    0.04283    0.02393   0.005166        196        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.41s/it]
               all       1952       3162      0.936      0.828      0.881      0.516

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  28/99      34.1G    0.04265    0.02398   0.005311        174        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.919      0.835      0.884      0.501

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  29/99      34.1G    0.04241    0.02361   0.005044        202        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162      0.946      0.836       0.88      0.507

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  30/99      34.1G     0.0421    0.02354   0.005081        190        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.937      0.839      0.887      0.532

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  31/99      34.1G    0.04251    0.02355   0.005008        196        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.35s/it]
               all       1952       3162      0.937       0.84      0.885      0.499

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  32/99      34.1G    0.04159     0.0234   0.004962        203        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.947      0.839      0.889       0.52

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  33/99      34.1G    0.04133    0.02342   0.004891        312        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162      0.946      0.838      0.886      0.537

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  34/99      34.1G    0.04179     0.0235     0.0049        242        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162      0.937      0.843      0.887      0.534

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  35/99      34.1G    0.04157    0.02353   0.004781        167        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.44s/it]
               all       1952       3162      0.944      0.842      0.892      0.537

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  36/99      34.1G    0.04146    0.02323   0.004801        193        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162       0.95      0.842      0.886      0.525

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  37/99      34.1G    0.04074     0.0229   0.004725        212        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162      0.956      0.838      0.889       0.53

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  38/99      34.1G    0.04123    0.02301   0.004877        174        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.42s/it]
               all       1952       3162      0.939      0.847      0.892      0.516

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  39/99      34.1G    0.04117    0.02313   0.004636        225        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.944       0.85      0.895      0.534

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  40/99      34.1G    0.04059    0.02293   0.004518        225        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.936      0.849      0.896      0.526

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  41/99      34.1G    0.04052    0.02294   0.004581        243        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162      0.949      0.848      0.898      0.544

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  42/99      34.1G    0.04042    0.02301    0.00462        206        640: 100% 162/162 [00:58<00:00,  2.75it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162      0.949      0.847      0.896      0.533

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  43/99      34.1G     0.0406    0.02292   0.004505        225        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.936      0.858        0.9       0.55

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  44/99      34.1G    0.04035    0.02286   0.004587        177        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162      0.958      0.844      0.899       0.54

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  45/99      34.1G    0.04016    0.02251   0.004494        203        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.42s/it]
               all       1952       3162      0.942      0.856        0.9      0.542

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  46/99      34.1G    0.04004    0.02245   0.004333        233        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.35s/it]
               all       1952       3162      0.942      0.858      0.901      0.546

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  47/99      34.1G    0.03981    0.02251   0.004399        185        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.35s/it]
               all       1952       3162      0.947      0.857      0.901      0.548

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  48/99      34.1G    0.03975     0.0221   0.004334        389        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.35s/it]
               all       1952       3162      0.953      0.854        0.9      0.544

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  49/99      34.1G    0.03965    0.02256   0.004295        193        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.35s/it]
               all       1952       3162      0.943      0.855      0.901       0.56

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  50/99      34.1G    0.03963    0.02226   0.004351        193        640: 100% 162/162 [00:57<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.949      0.859      0.901      0.545

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  51/99      34.1G     0.0391    0.02243   0.004272        170        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.951      0.858      0.902      0.552

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  52/99      34.1G    0.03972    0.02232    0.00406        228        640: 100% 162/162 [00:58<00:00,  2.76it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.35s/it]
               all       1952       3162      0.958      0.852      0.904      0.555

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  53/99      34.1G    0.03966    0.02222   0.004258        167        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.954      0.855      0.902      0.554

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  54/99      34.1G    0.03875    0.02196   0.004195        217        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.943       0.86      0.902       0.55

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  55/99      34.1G    0.03902    0.02235   0.004138        236        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.949      0.852        0.9      0.555

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  56/99      34.1G    0.03857    0.02195   0.004058        155        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162      0.952      0.856      0.903       0.56

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  57/99      34.1G    0.03882    0.02212   0.003942        164        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.949      0.861      0.905      0.556

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  58/99      34.1G     0.0385    0.02196   0.003887        164        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.33s/it]
               all       1952       3162      0.948       0.86      0.903      0.555

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  59/99      34.1G     0.0385    0.02231   0.003937        225        640: 100% 162/162 [00:58<00:00,  2.76it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.39s/it]
               all       1952       3162       0.94      0.864      0.904       0.56

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  60/99      34.1G    0.03871    0.02192    0.00403        187        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.943      0.864      0.904      0.562

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  61/99      34.1G     0.0383    0.02203   0.003795        167        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.946      0.868      0.909      0.561

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  62/99      34.1G    0.03776    0.02164   0.003861        240        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162      0.956       0.86      0.909      0.564

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  63/99      34.1G    0.03841    0.02162   0.003867        202        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.949      0.862      0.908      0.564

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  64/99      34.1G     0.0381    0.02179   0.003725        156        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162      0.945      0.868      0.909      0.562

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  65/99      34.1G    0.03808    0.02165   0.003747        214        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.42s/it]
               all       1952       3162       0.95      0.859      0.908      0.567

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  66/99      34.1G     0.0382    0.02156   0.003626        145        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.948      0.864      0.909      0.563

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  67/99      34.1G    0.03786    0.02167   0.003751        154        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.941      0.869      0.908      0.564

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  68/99      34.1G    0.03747    0.02133   0.003729        179        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.943      0.865      0.907      0.569

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  69/99      34.1G    0.03779     0.0217   0.003571        192        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.40s/it]
               all       1952       3162      0.944      0.867      0.907      0.569

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  70/99      34.1G    0.03702    0.02148   0.003697        229        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.945      0.867      0.908      0.566

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  71/99      34.1G    0.03704    0.02153   0.003551        197        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162      0.944      0.873      0.908      0.565

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  72/99      34.1G    0.03747    0.02145   0.003543        220        640: 100% 162/162 [00:58<00:00,  2.76it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.33s/it]
               all       1952       3162      0.947      0.872      0.909      0.566

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  73/99      34.1G    0.03687     0.0211   0.003677        197        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.36s/it]
               all       1952       3162      0.948      0.872      0.909      0.567

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  74/99      34.1G    0.03695    0.02137   0.003518        154        640: 100% 162/162 [00:58<00:00,  2.78it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.947      0.871       0.91      0.567

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  75/99      34.1G    0.03692    0.02128   0.003451        160        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.41s/it]
               all       1952       3162      0.952      0.866       0.91      0.567

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  76/99      34.1G    0.03669    0.02096   0.003503        179        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.944      0.871       0.91      0.567

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  77/99      34.1G     0.0367    0.02115   0.003459        269        640: 100% 162/162 [00:58<00:00,  2.77it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.37s/it]
               all       1952       3162      0.944      0.873       0.91      0.568

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  78/99      34.1G    0.03599    0.02103   0.003307        200        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.33s/it]
               all       1952       3162      0.949      0.874      0.911      0.567

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  79/99      34.1G    0.03636    0.02095   0.003255        186        640: 100% 162/162 [00:58<00:00,  2.76it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.951      0.874      0.912      0.567

Stopping training early as no improvement observed in last 10 epochs. Best results observed at epoch 69, best model saved as best.pt.
To update EarlyStopping(patience=10) pass a new patience value, i.e. python train.py --patience 300 or use --patience 0 to disable EarlyStopping.


80 epochs completed in 1.558 hours.
Optimizer stripped from /content/drive/MyDrive/Traing/Yolo5Ours/_101/exp/weights/last.pt, 16.9MB
Optimizer stripped from /content/drive/MyDrive/Traing/Yolo5Ours/_101/exp/weights/best.pt, 16.9MB

Validating /content/drive/MyDrive/Traing/Yolo5Ours/_101/exp/weights/best.pt...
Fusing layers... 
noor10 summary: 133 layers, 8264511 parameters, 0 gradients, 28.2 GFLOPs
                 Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:18<00:00,  2.29s/it]
                   all       1952       3162      0.944      0.867      0.908      0.569
                  mask       1952       1367      0.979      0.973       0.98      0.635
               no-mask       1952       1795      0.909       0.76      0.835      0.504
```'


### Additional

_No response_
@VYRION-Ai VYRION-Ai added the question Further information is requested label Jul 22, 2023
@VYRION-Ai
Copy link
Author

VYRION-Ai commented Jul 22, 2023 •
edited
Loading

@glenn-jocher @AyushExel @Borda Can you please help me with that?

@glenn-jocher
Copy link
Member

@VYRION-Ai sure, I'd be happy to help. Could you please provide more information about the issue you're facing? This will help us better understand and assist you with resolving it. Thank you.

@VYRION-Ai
Copy link
Author

VYRION-Ai commented Jul 22, 2023 •
edited
Loading

@glenn-jocher i do training on my dataset , as you in the image the final Epochs has higher map50 , R ,and P

         Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
           all       1952       3162      0.951      0.874      0.912      0.567

, But the final weight not

             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:18<00:00,  2.29s/it]
               all       1952       3162      0.944      0.867      0.908      0.569

255330225-b1922397-8831-43a9-be10-96c94fdf80a6

@glenn-jocher
Copy link
Member

@VYRION-Ai hello,

Based on the information you provided, it seems that in your training, the final Epochs showed higher metric values for map50, R, and P compared to the final weight. The table you shared demonstrates this difference.

Although you mentioned an image link, I'm unable to access it in this text-based format. However, if you have any specific questions about this issue or need further assistance, feel free to ask. We're here to help!

Thank you.

@VYRION-Ai
Copy link
Author

VYRION-Ai commented Jul 22, 2023 •
edited
Loading

@glenn-jocher how the final Epochs showed higher metric values for map50, R, and P compared to the final weight?
this is the final epochs


 Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  78/99      34.1G    0.03599    0.02103   0.003307        200        640: 100% 162/162 [00:58<00:00,  2.79it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:10<00:00,  1.33s/it]
               all       1952       3162      0.949      0.874      0.911      0.567

  Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  79/99      34.1G    0.03636    0.02095   0.003255        186        640: 100% 162/162 [00:58<00:00,  2.76it/s]
             Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 8/8 [00:11<00:00,  1.38s/it]
               all       1952       3162      0.951      0.874      0.912      0.567

and this the normalize final best.pt val
Validating /content/drive/MyDrive/Traing/Yolo5Ours/_101/exp/weights/best.pt... Fusing layers... noor10 summary: 133 layers, 8264511 parameters, 0 gradients, 28.2 GFLOPs Class Images Instances P R mAP50 mAP50-95: 100% 8/8 [00:18<00:00, 2.29s/it] all 1952 3162 0.944 0.867 0.908 0.569 mask 1952 1367 0.979 0.973 0.98 0.635 no-mask 1952 1795 0.909 0.76 0.835 0.504

is the final Epochs should be consider the best map i got ?

@glenn-jocher
Copy link
Member

@VYRION-Ai based on the information you provided, it appears that the final Epochs achieved higher metric values for map50, R, and P compared to the final weight. The table you shared illustrates this difference.

In terms of determining the best map, it is generally advisable to consider the metric values from the final Epochs, as they reflect the model's performance after completing the entire training process. However, it's important to note that selecting the best map or determining model performance requires careful evaluation and consideration of various factors, such as dataset, training setup, and specific requirements.

If you have any further questions or need additional assistance, please feel free to ask.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
question Further information is requested
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@VYRION-Ai @glenn-jocher