YOLO Segment `sigmoid()` fix #13939

glenn-jocher · 2024-06-24T15:38:01Z

…chieve better segmentation edges. (#13912)

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

This PR simplifies mask processing logic and enhances testing in the Ultralytics repository.

📊 Key Changes

Updated CI workflow to install pytest-cov for test coverage reporting.
Modified process_mask_* functions in ultralytics/utils/ops.py to remove unnecessary sigmoid operations and adjusted the mask threshold condition.

🎯 Purpose & Impact

Enhanced Testing: 👍 Adding pytest-cov helps in measuring test coverage, leading to higher code quality and reliability.
Optimization: 🚀 Simplifying mask processing by removing redundant operations slightly improves performance and reduces computational overhead.
Accuracy Adjustment: 🎯 Adjusting the mask threshold from 0.5 to 0.0 could impact mask generation results, potentially increasing model sensitivity.

…chieve better segmentation edges. (#13912) Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

codecov · 2024-06-24T15:41:14Z

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 70.25%. Comparing base (3bb0c5a) to head (da3aa13).

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #13939   +/-   ##
=======================================
  Coverage   70.25%   70.25%           
=======================================
  Files         124      124           
  Lines       15891    15891           
=======================================
  Hits        11164    11164           
  Misses       4727     4727

Flag	Coverage Δ
Benchmarks	`35.81% <33.33%> (ø)`
GPU	`36.90% <0.00%> (ø)`
Tests	`66.48% <100.00%> (ø)`

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

glenn-jocher · 2024-06-24T18:05:54Z

Main branch results:

root@113944910eee:/usr/src/ultralytics# yolo val model=yolov8n-seg.pt data=coco.yaml device=5 
Ultralytics YOLOv8.2.41 🚀 Python-3.10.14 torch-2.3.1 CUDA:5 (NVIDIA A100-SXM4-80GB, 81051MiB)
YOLOv8n-seg summary (fused): 195 layers, 3404320 parameters, 0 gradients, 12.6 GFLOPs
val: Scanning /usr/src/datasets/coco/labels/val2017.cache... 4952 images, 48 backgrounds, 0 corrupt: 100%|██████████| 5000/5000 [00:00<?, ?it/s]
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95)     Mask(P          R      mAP50  mAP50-95): 100%|██████████| 313/313 [13:45<00:00,  2.64s/it]
                   all       5000      36335      0.621       0.48      0.518      0.364      0.622      0.457      0.491      0.306
                person       2693      10777      0.742       0.68      0.744      0.515      0.741      0.651       0.71        0.4
               bicycle        149        314      0.629      0.385      0.445      0.245      0.585      0.334       0.36      0.135
                   car        535       1918      0.632      0.516      0.556      0.356      0.638      0.484      0.528      0.296
            motorcycle        159        367      0.702      0.567      0.657      0.406      0.687      0.518      0.576      0.286
              airplane         97        143      0.715      0.769      0.822      0.639      0.739      0.769      0.792      0.492
                   bus        189        283      0.744      0.707      0.751      0.608      0.754      0.692      0.735      0.559
                 train        157        190      0.827      0.782      0.851      0.635      0.815       0.74      0.818      0.599
                 truck        250        414      0.541      0.391      0.451        0.3      0.547      0.365      0.415      0.253
                  boat        121        424      0.545       0.34      0.369      0.197      0.528      0.292      0.332      0.144
         traffic light        191        634       0.58      0.338      0.384      0.198      0.607      0.326      0.383      0.184
          fire hydrant         86        101      0.852      0.723      0.775      0.619      0.861      0.723      0.778      0.561
             stop sign         69         75      0.706      0.627      0.672      0.603      0.717      0.608      0.672      0.573
         parking meter         37         60      0.752        0.5      0.608      0.442      0.772        0.5      0.584      0.407
                 bench        235        411      0.543      0.277      0.305      0.201      0.532      0.253      0.269      0.132
                  bird        125        427      0.588      0.365      0.409      0.271      0.605      0.356      0.397      0.219
                   cat        184        202      0.744      0.822      0.838      0.646      0.759      0.812      0.849      0.626
                   dog        177        218      0.676      0.679      0.716      0.574      0.699      0.671       0.71      0.532
                 horse        128        272      0.715      0.658      0.718      0.537      0.721      0.645      0.689      0.376
                 sheep         65        354      0.559      0.644      0.656      0.447      0.573      0.624      0.629      0.367
                   cow         87        372      0.656      0.616      0.673      0.475      0.655      0.581      0.633       0.38
              elephant         89        252      0.722      0.837      0.813      0.619      0.738      0.845      0.825      0.554
                  bear         49         71      0.793      0.845       0.85      0.679      0.769      0.817      0.831      0.654
                 zebra         85        266      0.774      0.835      0.892      0.668      0.791      0.823      0.871       0.55
               giraffe        101        232      0.848      0.843      0.892      0.695      0.836      0.819      0.858       0.54
              backpack        228        371       0.44       0.17      0.183     0.0939      0.442      0.151      0.168     0.0796
              umbrella        174        407      0.631      0.499      0.544      0.366      0.693      0.516      0.573      0.378
               handbag        292        540        0.5      0.152      0.179     0.0902      0.547      0.143      0.171     0.0815
                   tie        145        252       0.67      0.397      0.441      0.279      0.709      0.397      0.436       0.24
              suitcase        105        299      0.598      0.408      0.473      0.322      0.607      0.381      0.445      0.288
               frisbee         84        115      0.776      0.678      0.761      0.591      0.799      0.678      0.761      0.534
                  skis        120        241      0.552      0.357      0.368      0.188      0.392      0.228      0.201     0.0433
             snowboard         49         69      0.527      0.348      0.368       0.26      0.514      0.319      0.329      0.166
           sports ball        169        260      0.683      0.455      0.479      0.329      0.686      0.438       0.46      0.278
                  kite         91        327      0.583      0.538      0.571      0.378      0.577      0.483      0.524      0.275
          baseball bat         97        145       0.54      0.379      0.398      0.217      0.589      0.385      0.395      0.171
        baseball glove        100        148      0.633      0.493      0.526      0.319      0.651      0.473      0.526      0.303
            skateboard        127        179      0.656      0.665      0.662      0.463      0.622      0.592      0.596      0.273
             surfboard        149        267      0.608      0.502      0.518      0.311      0.604      0.472      0.488      0.242
         tennis racket        167        225      0.715      0.627      0.673      0.392      0.737      0.622      0.673      0.431
                bottle        379       1013      0.603      0.397      0.449      0.294      0.605      0.373      0.426      0.255
            wine glass        110        341        0.7      0.346      0.422      0.275      0.697      0.323      0.379      0.205
                   cup        390        895       0.59      0.426      0.469      0.336      0.591      0.406      0.451      0.308
                  fork        155        215      0.577      0.349      0.372      0.248      0.547      0.302      0.279      0.118
                 knife        181        325      0.474      0.163      0.177      0.104      0.471      0.154      0.144     0.0688
                 spoon        153        253      0.383      0.138      0.158     0.0877      0.403      0.125      0.124     0.0502
                  bowl        314        623       0.61      0.492       0.53      0.399       0.58      0.446      0.471      0.277
                banana        103        370       0.52      0.311      0.341      0.216       0.53      0.286      0.312      0.164
                 apple         76        236      0.399      0.205      0.209      0.148      0.413      0.191      0.199      0.133
              sandwich         98        177      0.563      0.492      0.483      0.358      0.538      0.446      0.418      0.312
                orange         85        285      0.483        0.4      0.377      0.293      0.495      0.386      0.363      0.256
              broccoli         71        312      0.513      0.369      0.376      0.207      0.567      0.378      0.383      0.182
                carrot         81        365      0.433       0.29      0.278      0.176      0.439      0.268      0.265      0.145
               hot dog         51        125      0.581      0.388      0.419      0.285      0.527      0.336      0.342      0.201
                 pizza        153        284      0.683      0.613      0.668        0.5      0.688      0.602      0.638      0.452
                 donut         62        328      0.557      0.476      0.515      0.403      0.575      0.462      0.505      0.372
                  cake        124        310      0.509      0.378      0.429      0.272      0.534      0.371      0.425      0.259
                 chair        580       1771      0.574      0.343      0.398      0.246      0.541        0.3      0.333      0.148
                 couch        195        261      0.588      0.559      0.589      0.444      0.581      0.521      0.529      0.344
          potted plant        172        342       0.52      0.354      0.367      0.211       0.49      0.316       0.31      0.132
                   bed        149        163      0.572      0.558      0.573      0.389      0.541      0.509      0.488       0.29
          dining table        501        695      0.511      0.442      0.418      0.276      0.402      0.324      0.271      0.103
                toilet        149        179      0.674      0.729       0.76      0.623      0.728      0.737      0.774      0.585
                    tv        207        288       0.67      0.646      0.688      0.528      0.683      0.639      0.683      0.491
                laptop        183        231      0.637      0.615      0.667      0.541      0.615      0.574      0.595      0.383
                 mouse         88        106      0.648      0.695      0.703      0.521      0.655      0.679      0.681      0.481
                remote        145        283      0.429      0.219      0.268      0.156      0.488      0.225      0.267      0.133
              keyboard        106        153      0.635      0.608      0.636      0.461      0.635      0.582      0.642      0.441
            cell phone        214        262      0.519      0.363      0.378      0.252      0.536       0.34      0.369       0.23
             microwave         54         55       0.57      0.554      0.611       0.49       0.57       0.53      0.615      0.452
                  oven        115        143      0.627      0.483      0.541      0.355      0.619      0.454      0.479      0.278
               toaster          8          9          1      0.214      0.565       0.38          1       0.21      0.565      0.427
                  sink        187        225      0.588      0.498      0.506      0.335      0.619      0.507      0.513      0.298
          refrigerator        101        126      0.711      0.635      0.677      0.521      0.736      0.627      0.683      0.491
                  book        230       1129      0.459      0.113      0.181     0.0872      0.389     0.0834      0.121     0.0473
                 clock        204        267      0.662      0.615      0.648      0.444      0.678      0.614      0.659      0.426
                  vase        137        274      0.547      0.428      0.444      0.302      0.543      0.401      0.414      0.263
              scissors         28         36      0.521      0.302      0.304      0.229      0.507      0.278      0.303      0.155
            teddy bear         94        190      0.703      0.558      0.623      0.458      0.713      0.553      0.611      0.409
            hair drier          9         11          1          0    0.00414    0.00218          1          0     0.0398     0.0116
            toothbrush         34         57      0.455      0.228      0.188      0.117       0.45      0.211      0.197     0.0831
Speed: 0.1ms preprocess, 1.4ms inference, 0.0ms loss, 0.9ms postprocess per image
Saving /usr/src/ultralytics/runs/segment/val3/predictions.json...

Evaluating pycocotools mAP using /usr/src/ultralytics/runs/segment/val3/predictions.json and /usr/src/datasets/coco/annotations/instances_val2017.json...
loading annotations into memory...
Done (t=0.39s)
creating index...
index created!
Loading and preparing results...
DONE (t=7.10s)
creating index...
index created!
Running per image evaluation...
Evaluate annotation type *bbox*
DONE (t=58.32s).
Accumulating evaluation results...
DONE (t=13.35s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.367
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.522
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.398
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.179
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.404
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.521
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.315
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.532
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.586
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.366
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.650
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.767
Running per image evaluation...
Evaluate annotation type *segm*
DONE (t=64.70s).
Accumulating evaluation results...
DONE (t=13.44s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.304
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.491
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.321
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.122
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.334
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.461
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.273
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.437
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.472
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.252
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.532
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.674
Results saved to /usr/src/ultralytics/runs/segment/val3
💡 Learn more at https://docs.ultralytics.com/modes/val

glenn-jocher · 2024-06-24T18:24:55Z

PR results

root@113944910eee:/usr/src/ultralytics# yolo val model=yolov8n-seg.pt data=coco.yaml device=5
Downloading https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov8n-seg.pt to 'yolov8n-seg.pt'...
100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 6.73M/6.73M [00:00<00:00, 230MB/s]
Ultralytics YOLOv8.2.42 🚀 Python-3.10.14 torch-2.3.1 CUDA:5 (NVIDIA A100-SXM4-80GB, 81051MiB)
YOLOv8n-seg summary (fused): 195 layers, 3404320 parameters, 0 gradients, 12.6 GFLOPs
val: Scanning /usr/src/datasets/coco/labels/val2017.cache... 4952 images, 48 backgrounds, 0 corrupt: 100%|██████████| 5000/5000 [00:00<?, ?it/s]
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95)     Mask(P          R      mAP50  mAP50-95): 100%|██████████| 313/313 [13:46<00:00,  2.64s/it]
                   all       5000      36335      0.621       0.48      0.518      0.364      0.622      0.456       0.49      0.303
                person       2693      10777      0.742       0.68      0.744      0.515      0.741       0.65      0.709      0.395
               bicycle        149        314      0.629      0.385      0.445      0.245      0.585      0.334       0.36      0.135
                   car        535       1918      0.632      0.516      0.556      0.356      0.638      0.484      0.526      0.291
            motorcycle        159        367      0.702      0.567      0.657      0.406      0.688      0.518      0.575      0.286
              airplane         97        143      0.715      0.769      0.822      0.639      0.739      0.769      0.792       0.48
                   bus        189        283      0.744      0.707      0.751      0.608      0.753      0.691      0.735      0.559
                 train        157        190      0.827      0.782      0.851      0.635      0.814      0.739      0.818      0.599
                 truck        250        414      0.541      0.391      0.451        0.3      0.549      0.365      0.415      0.251
                  boat        121        424      0.545       0.34      0.369      0.197      0.526       0.29      0.326      0.142
         traffic light        191        634       0.58      0.338      0.384      0.198      0.609      0.326      0.381      0.182
          fire hydrant         86        101      0.852      0.723      0.775      0.619      0.862      0.723      0.778      0.551
             stop sign         69         75      0.706      0.627      0.672      0.603      0.717      0.607      0.672      0.575
         parking meter         37         60      0.752        0.5      0.608      0.442      0.773        0.5      0.584      0.406
                 bench        235        411      0.543      0.277      0.305      0.201      0.532      0.253      0.269      0.132
                  bird        125        427      0.588      0.365      0.409      0.271      0.602      0.351      0.394      0.216
                   cat        184        202      0.744      0.822      0.838      0.646      0.764      0.812      0.849      0.626
                   dog        177        218      0.676      0.679      0.716      0.574      0.699       0.67       0.71      0.528
                 horse        128        272      0.715      0.658      0.718      0.537       0.72      0.644      0.689      0.373
                 sheep         65        354      0.559      0.644      0.656      0.447      0.571      0.621      0.627      0.362
                   cow         87        372      0.656      0.616      0.673      0.475      0.655      0.581      0.633      0.373
              elephant         89        252      0.722      0.837      0.813      0.619      0.739      0.845      0.825      0.552
                  bear         49         71      0.793      0.845       0.85      0.679       0.77      0.817      0.831      0.652
                 zebra         85        266      0.774      0.835      0.892      0.668      0.795      0.823      0.871      0.549
               giraffe        101        232      0.848      0.843      0.892      0.695      0.836      0.819      0.858      0.537
              backpack        228        371       0.44       0.17      0.183     0.0939      0.446      0.151      0.167     0.0774
              umbrella        174        407      0.631      0.499      0.544      0.366      0.689      0.513      0.571      0.373
               handbag        292        540        0.5      0.152      0.179     0.0902      0.548      0.142      0.171     0.0795
                   tie        145        252       0.67      0.397      0.441      0.279      0.708      0.393      0.436      0.234
              suitcase        105        299      0.598      0.408      0.473      0.322      0.608      0.381      0.445      0.288
               frisbee         84        115      0.776      0.678      0.761      0.591        0.8      0.678      0.761      0.521
                  skis        120        241      0.552      0.357      0.368      0.188      0.379       0.22      0.192     0.0395
             snowboard         49         69      0.527      0.348      0.368       0.26      0.515      0.319      0.324      0.164
           sports ball        169        260      0.683      0.455      0.479      0.329      0.682      0.435      0.455      0.267
                  kite         91        327      0.583      0.538      0.571      0.378      0.581      0.486      0.523      0.258
          baseball bat         97        145       0.54      0.379      0.398      0.217      0.598       0.39      0.409      0.163
        baseball glove        100        148      0.633      0.493      0.526      0.319      0.652      0.473      0.525      0.298
            skateboard        127        179      0.656      0.665      0.662      0.463      0.622      0.592      0.596      0.271
             surfboard        149        267      0.608      0.502      0.518      0.311      0.591      0.461      0.482      0.238
         tennis racket        167        225      0.715      0.627      0.673      0.392      0.736      0.621      0.673      0.424
                bottle        379       1013      0.603      0.397      0.449      0.294      0.606      0.371      0.426      0.251
            wine glass        110        341        0.7      0.346      0.422      0.275      0.698      0.323      0.376      0.203
                   cup        390        895       0.59      0.426      0.469      0.336      0.592      0.406      0.451      0.305
                  fork        155        215      0.577      0.349      0.372      0.248      0.549      0.302      0.278      0.117
                 knife        181        325      0.474      0.163      0.177      0.104       0.47      0.151      0.142     0.0667
                 spoon        153        253      0.383      0.138      0.158     0.0877      0.397      0.122      0.121     0.0486
                  bowl        314        623       0.61      0.492       0.53      0.399      0.581      0.445      0.471      0.276
                banana        103        370       0.52      0.311      0.341      0.216      0.524      0.282      0.311      0.163
                 apple         76        236      0.399      0.205      0.209      0.148      0.414      0.191      0.199      0.132
              sandwich         98        177      0.563      0.492      0.483      0.358      0.539      0.446      0.418      0.312
                orange         85        285      0.483        0.4      0.377      0.293      0.494      0.382      0.363      0.256
              broccoli         71        312      0.513      0.369      0.376      0.207      0.562      0.374      0.381      0.182
                carrot         81        365      0.433       0.29      0.278      0.176      0.441      0.266      0.264      0.143
               hot dog         51        125      0.581      0.388      0.419      0.285      0.528      0.336      0.342      0.201
                 pizza        153        284      0.683      0.613      0.668        0.5      0.689      0.602      0.638      0.451
                 donut         62        328      0.557      0.476      0.515      0.403      0.574      0.461      0.505       0.37
                  cake        124        310      0.509      0.378      0.429      0.272      0.534      0.368      0.425      0.259
                 chair        580       1771      0.574      0.343      0.398      0.246       0.54      0.299      0.332      0.147
                 couch        195        261      0.588      0.559      0.589      0.444       0.58      0.519      0.529      0.342
          potted plant        172        342       0.52      0.354      0.367      0.211       0.49      0.316       0.31      0.131
                   bed        149        163      0.572      0.558      0.573      0.389      0.541      0.509      0.488       0.29
          dining table        501        695      0.511      0.442      0.418      0.276      0.402      0.324      0.271      0.103
                toilet        149        179      0.674      0.729       0.76      0.623       0.73      0.737      0.774      0.585
                    tv        207        288       0.67      0.646      0.688      0.528      0.684      0.639      0.683       0.49
                laptop        183        231      0.637      0.615      0.667      0.541      0.615      0.571      0.595      0.383
                 mouse         88        106      0.648      0.695      0.703      0.521      0.655      0.679      0.681      0.475
                remote        145        283      0.429      0.219      0.268      0.156      0.487      0.225      0.266       0.13
              keyboard        106        153      0.635      0.608      0.636      0.461      0.636      0.582      0.642       0.44
            cell phone        214        262      0.519      0.363      0.378      0.252      0.537       0.34      0.366      0.229
             microwave         54         55       0.57      0.554      0.611       0.49      0.569      0.529      0.615       0.45
                  oven        115        143      0.627      0.483      0.541      0.355      0.617       0.45      0.479      0.278
               toaster          8          9          1      0.214      0.565       0.38          1      0.209      0.565      0.418
                  sink        187        225      0.588      0.498      0.506      0.335      0.622      0.504      0.513      0.297
          refrigerator        101        126      0.711      0.635      0.677      0.521      0.737      0.627      0.683       0.49
                  book        230       1129      0.459      0.113      0.181     0.0872      0.389     0.0824       0.12     0.0463
                 clock        204        267      0.662      0.615      0.648      0.444      0.678      0.614      0.659      0.422
                  vase        137        274      0.547      0.428      0.444      0.302      0.543      0.401      0.414      0.261
              scissors         28         36      0.521      0.302      0.304      0.229      0.516      0.278      0.303      0.156
            teddy bear         94        190      0.703      0.558      0.623      0.458      0.714      0.553      0.611      0.409
            hair drier          9         11          1          0    0.00414    0.00218          1          0     0.0398     0.0116
            toothbrush         34         57      0.455      0.228      0.188      0.117      0.451      0.211      0.197     0.0819
Speed: 0.1ms preprocess, 1.4ms inference, 0.0ms loss, 0.9ms postprocess per image
Saving /usr/src/ultralytics/runs/segment/val/predictions.json...

Evaluating pycocotools mAP using /usr/src/ultralytics/runs/segment/val/predictions.json and /usr/src/datasets/coco/annotations/instances_val2017.json...
loading annotations into memory...
Done (t=0.40s)
creating index...
index created!
Loading and preparing results...
DONE (t=7.28s)
creating index...
index created!
Running per image evaluation...
Evaluate annotation type *bbox*
DONE (t=57.65s).
Accumulating evaluation results...
DONE (t=13.32s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.367
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.522
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.398
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.179
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.404
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.521
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.315
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.532
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.586
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.366
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.650
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.767
Running per image evaluation...
Evaluate annotation type *segm*
DONE (t=64.19s).
Accumulating evaluation results...
DONE (t=13.39s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.304
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.491
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.320
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.121
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.334
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.461
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.273
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.436
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.471
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.251
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.532
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.674
Results saved to /usr/src/ultralytics/runs/segment/val
💡 Learn more at https://docs.ultralytics.com/modes/val

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

glenn-jocher · 2024-06-24T19:06:04Z

@Laughing-q this fix PR seems to make a lot of sense, and the edges are much smoother (see #13912), but segment mAP drops very slightly after implementing it. What do you think?

Laughing-q · 2024-06-25T06:42:57Z

@glenn-jocher I also tested yolov8s-seg and yolov8m-seg, both of them shown slightly mAP drop as well.
yolov8m-seg:

 # main
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.405
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.633
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.432
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.216
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.456
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.582
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.327
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.521
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.556
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.351
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.621
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.741

 # segment-fix
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.404
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.634
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.431
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.213
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.456
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.581
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.326
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.520
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.555
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.347
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.620
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.740

yolov8s-seg:

 # main
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.366
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.580
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.387
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.168
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.410
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.537
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.306
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.488
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.521
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.301
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.585
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.715
 # segment-fix
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.365
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.580
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.385
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.168
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.409
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.537
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.306
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.487
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.520
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.300
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.583
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.714

But the update does show better visualization, meanwhile eliminating sigmoid operator should be slightly faster in post-processing. How about we remove it and leave a comment says using sigmoid might get better mAP and see #13939 ?

Laughing-q · 2024-06-25T06:56:05Z

@glenn-jocher Here are some visualized samples on main branch and this branch, and it looks like on the latter in some level it's fixing the "cut" issue we are having when retina_mask=False. Let's leave a comment and merge this then. :)

BTW I think we are having too many process_mask functions even though each one of them stands for different case. I'll take a look at this part when I have time and try to merge them into one function if possible, this way we have chance to simply add an option to the function(instead of three places) to determine if use sigmoid or not, since using them gets better mAP but we get better visualizations without it.

glenn-jocher · 2024-06-25T09:52:38Z

@Laughing-q yes I think on a mathematical level this PR does make sense, it's taking all the values that range from -inf to +inf and placing the threshold at 0, so there will be a better dynamic range on the op than if we sigmoid from 0 to 1 and then do the threshold at 0.5.

The mAP drops are in the Ultralytics mAP but the COCO mAP is mostly the same, and the visual results do look better, less discretization, so let's merge this.

glenn-jocher · 2024-06-25T09:54:02Z

@zxDeepDiver PR merged! Thank you for your contributions.

glenn-jocher · 2024-06-25T09:55:02Z

@Laughing-q yes the multiple mask processing functions are too many, I agree, I think there are at least two because of the order of processing the upsample, either before or after the mask processing and depending on if it's retina=True or not, but it would be nice to unify them.

Alarmod · 2024-07-06T11:56:15Z

@glenn-jocher Maybe this fix can be made optional? My previously trained neural network, focused on working with ultra-small objects, shows a lower quality...

Alarmod · 2024-07-06T12:16:26Z

pip3 install -U ultralytics==8.2.42

predict_t2_brain_512
Precision --- 0.982777
Recall --- 0.981744
F1 98.22602283 (multiplied by 100.0)

predict_t2_ischemia
Precision --- 0.997980
Recall --- 0.961361
F1 97.93283058

pip3 install -U ultralytics==8.2.43

predict_t2_brain_512
Precision --- 0.975905
Recall --- 0.985638
F1 98.07473530

predict_t2_ischemia
Precision --- 0.971908
Recall --- 0.961250
F1 96.65496198

All F1 values worse on the new version

glenn-jocher · 2024-07-06T16:09:50Z

Hi @Alarmod,

Thank you for sharing your detailed results. It appears that the recent update has impacted the performance metrics of your model, particularly the F1 scores. We understand how crucial these metrics are for your application, especially when dealing with ultra-small objects.

To address this, we can consider making the recent fix optional. This would allow users to choose whether to apply the sigmoid operation based on their specific use case.

In the meantime, you can revert to the previous version (8.2.42) to maintain your model's performance. If you could provide a minimum reproducible example of your code, it would help us investigate the issue further and ensure that any changes we make will not negatively impact your use case. You can find more details on how to create a reproducible example here.

Additionally, please ensure that you are using the latest versions of all dependencies to rule out any compatibility issues.

Thank you for your patience and understanding. We are committed to improving the performance and flexibility of our models and appreciate your feedback.

yangkai1129 · 2024-08-09T06:38:50Z

The simplification of process_mask_* functions caused boundary displacement error.

I have tested in my datasets and found a so many problems like the ones in the red box above.
When I added sigmoid back and changed the threshold back to 0.5, I got better results

pderrenger · 2024-08-09T10:24:10Z

Thank you for your detailed feedback and for providing visual examples. It appears that the recent changes to the process_mask_* functions have introduced boundary displacement errors in your results. We appreciate your thorough testing and observations.

To address this, we are considering making the sigmoid operation and threshold adjustments optional, allowing users to choose the configuration that best suits their needs. In the meantime, you can revert to the previous implementation by adding the sigmoid operation and adjusting the threshold back to 0.5 in your local setup.

We recommend checking if the issue persists with the latest package versions to ensure compatibility and performance. Your input is invaluable in helping us improve the model, and we are committed to resolving this issue promptly. Thank you for your patience and understanding.

* Display Val images per class (ultralytics#12645) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Docs: Update HUB Teams page (ultralytics#13215) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix ARM64 Docker building (ultralytics#13217) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix Classifier last layer indexing (ultralytics#13219) * Update Exporter for `tf_keras` install (ultralytics#13231) * Install `cmake` with `onnxsim` (ultralytics#13222) * Add multiple lines graph support in `analytics` 8.2.26 (ultralytics#13214) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add Prettier for YAML formatting to Ultralytics Actions (ultralytics#13236) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Docs Solutions to Navigation Bar (ultralytics#13249) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Richard Abrich <richard.abrich@gmail.com> * Add Data Collection and Annotation Docs Page and Preprocessing Annotated Data Docs Page (ultralytics#13253) * Update `tensorstore` link for ARM64 Docker (ultralytics#13264) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add YOLOv10 model description in docs home page (ultralytics#13265) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix `xyxyxyxy2xywhr` for Numpy inputs (ultralytics#13273) * `ultralytics 8.2.27` replace `onnxsim` with `onnxslim` (ultralytics#12989) Co-authored-by: inisis <desmond.yao@buaa.edu.cn> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: inisis <46103969+inisis@users.noreply.github.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Use new `ultralytics-thop` package (ultralytics#13282) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Revert `ultralytics-thop` to optional (ultralytics#13290) * Update HUB Integrations page in Docs (ultralytics#13292) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Docs spelling and grammar fixes (ultralytics#13307) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: RainRat <rainrat78@yahoo.ca> * Update HUB SDK Docs (ultralytics#13309) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * import `torchvision` during Predict warmup (ultralytics#13313) * `ultralytics 8.2.28` improved NMS speeds (ultralytics#13315) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix for ValueError (expected type INT8) (ultralytics#13341) * Suppress YouTube Test `ConnectionError` (ultralytics#13357) * Add ops.py comment for `nm` = number of masks (ultralytics#13353) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix Docs export table from `onnxsim` to `onnxslim` (ultralytics#13324) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add area chart in `analytics` (ultralytics#13391) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Clean up Docs pages (ultralytics#13370) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Search for model metadata with TensorFlow GraphDef (ultralytics#13389) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Edge TPU export table fix (ultralytics#13420) * Add https://youtu.be/pWYiene9lYw & https://youtu.be/Unt4Lfid7aY to docs (ultralytics#13380) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update Minimum Reproducible Example (MRE) Docs page (ultralytics#13443) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update Python file headers (ultralytics#13445) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix Paddle 2.6.0 Dockerfile install bug (ultralytics#13447) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.29` new fractional AutoBatch feature (ultralytics#13446) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Burhan <62214284+Burhan-Q@users.noreply.github.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Refactor Python code (ultralytics#13448) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Code Refactor for Speed and Readability (ultralytics#13450) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.30` automated tags and release notes (ultralytics#13164) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix release publish action (ultralytics#13462) * Update publish.yml to handle existing tag (ultralytics#13463) * Update publish.yml (ultralytics#13464) * Docs Prettier reformat (ultralytics#13483) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Update CI to `pull_request_target` (ultralytics#13495) * Add NVIDIA Jetpack4 and Jetpack5 Docker Images (ultralytics#13100) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Lakshantha <lakshantha@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.31` NVIDIA Jetpack4 and Jetpack5 Dockerfile Images (ultralytics#13496) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix `jp` to `jetpack` (ultralytics#13499) * Add https://youtu.be/ydGdibB5Mds to docs (ultralytics#13555) * `ultralytics 8.2.32` Apple MPS device Autobatch handling (ultralytics#13568) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.33` pin `numpy<2.0.0` for compatibility (ultralytics#13661) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.34` bump `ultralytics-thop>=2.0.0` (ultralytics#13662) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Code Refactor `ruff check --fix --extend-select I` (ultralytics#13672) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Clean up unused `Silence` module (ultralytics#13674) * `ultralytics 8.2.35` add YOLOv9t/s/m models (ultralytics#13504) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update YOLOv9 YAMLs (ultralytics#13756) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix TensorRT Doc Overview URL (ultralytics#13759) * YOLOv9 model docs page cleanup (ultralytics#13757) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Apply new Ruff actions to Python codeblocks (ultralytics#13783) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * ROS Quickstart Guide (ultralytics#13553) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update Ruff formatting actions (ultralytics#13787) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * https://youtu.be/biIW5Z6GYl0 to docs (ultralytics#13797) * Replace https://youtu.be/biIW5Z6GYl0 with https://www.youtube.com/embed/biIW5Z6GYl0 (ultralytics#13804) * Deprecate `Silence` module in favor of `nn.Identity` (ultralytics#13785) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Laughing-q <1185102784@qq.com> * `ultralytics 8.2.36` update Ultralytics color palette (ultralytics#13808) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update CLA Action (ultralytics#13831) * Update OpenCV image read error message (ultralytics#13822) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * ROS quickstart Docs page cleanup (ultralytics#13835) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Annotator `txt_color` updates (ultralytics#13842) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.37` update `temporary_modules` and Remove YOLOv9e `Silence` module (ultralytics#13819) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.38` official YOLOv10 support (ultralytics#13113) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Laughing-q <1185102784@qq.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com> * Code refactor https://ultralytics.com/actions (ultralytics#13844) * ROS Quickstart, fixed code formatting (ultralytics#13855) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Replace `+=` with faster list `.append()` (ultralytics#13849) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix ambiguous variable names (ultralytics#13864) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Alex Pasquali <alexpasquali98@gmail.com> * Add https://youtu.be/tq3FU_QczxE to docs (ultralytics#13867) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update yolov8-p6.yaml with model parameters and GFLOPs (ultralytics#13862) * Fix HUB link https://ultralytics.com/hub (ultralytics#13884) Co-authored-by: Mughees Ahmad <itsmughees@gmail.com> * `ultralytics 8.2.39` update `onnxslim>=0.1.31` (ultralytics#13883) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Kayzwer <68285002+Kayzwer@users.noreply.github.com> * `ultralytics 8.2.40` refactor HUB code into callbacks (ultralytics#13896) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Adjust `numpy<2.0.0` compatibility (ultralytics#13906) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix HUB session with DDP training (ultralytics#13103) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Burhan <62214284+Burhan-Q@users.noreply.github.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Compress ROS Guide Images (ultralytics#13914) * `ultralytics 8.2.41` fix HUB unzip subdirectory bug (ultralytics#13913) * Add YOLOv8 OpenVINO C++ Inference example (ultralytics#13839) Co-authored-by: Muhammad Amir Abdurrozaq <m.amir.hs19@gmail.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Ultralytics TensorRT10 update (ultralytics#13933) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Dockerfile FROM `pytorch/pytorch:2.3.1-cuda12.1-cudnn8-runtime` (ultralytics#13937) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add CLI commands for `predict` and `train` YOLOv10 models. (ultralytics#13940) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.42` NVIDIA TensorRT 10 default (ultralytics#13943) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: DeepDiver <zhaoxu1015@gmail.com> Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com> * YOLO Segment `sigmoid()` fix (ultralytics#13939) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: DeepDiver <zhaoxu1015@gmail.com> Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com> * Add warning for inference end2end models with `augment` arg (ultralytics#13958) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Adjust box labels on right image side (ultralytics#13959) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.43` enable `classes` filter for end2end models (ultralytics#13971) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Added a `max_size` parameter to the `plot_images` function (ultralytics#14002) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add Tips for Model Training Docs Page (ultralytics#14011) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.44` Increase Predict dataloader robustness (ultralytics#14005) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.45` Fix YOLOv8 `augment` inference (ultralytics#14017) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.46` fix OBB Results `xyxy` attribute (ultralytics#14020) * Add https://youtu.be/eX5ad6udQ9Q to docs (ultralytics#14077) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix deprecation warning (ultralytics#14091) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add Insights on Model Evaluation and Fine-Tuning Docs Page (ultralytics#14085) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Strip `dfl_loss` from `BboxLoss` (ultralytics#14041) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.46` Results, DFL and AIGym fixes (ultralytics#14074) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: AAOMM <52826299+Chayanonjackal@users.noreply.github.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> Co-authored-by: zzzer <48149018+zhixuwei@users.noreply.github.com> Co-authored-by: Abirami Vina <abirami.vina@gmail.com> * `ultralytics 8.2.47` YOLOv8 zero-shot action recognition example (ultralytics#13935) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.48` strip model `criterion` on save (ultralytics#14106) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Default strip_optimizer() to `use_dill=False` (ultralytics#14107) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Ultralytics Code Refactor https://ultralytics.com/actions (ultralytics#14109) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Update NVIDIA Jetson DeepStream Guide with YOLOv8 and Jetson Orin Support (ultralytics#14059) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> * Update FAQ.md (ultralytics#14134) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Update Results and CFG docstrings (ultralytics#14139) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add Docs models pages FAQs (ultralytics#14167) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add Model Testing Guide and Best Practices for Model Deployment Guide (ultralytics#14105) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: RizwanMunawar <chr043416@gmail.com> Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> * Add https://youtu.be/mUybgOlSxxA to docs (ultralytics#14195) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add FAQ sections to Modes and Tasks (ultralytics#14181) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Abirami Vina <abirami.vina@gmail.com> Co-authored-by: RizwanMunawar <chr043416@gmail.com> Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> * `ultralytics 8.2.49` fix classification `setup_model` (ultralytics#14199) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add FAQs to Docs Datasets and Help sections (ultralytics#14211) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix mkdocs.yml raw image URLs (ultralytics#14213) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Burhan <62214284+Burhan-Q@users.noreply.github.com> * Fix Action Recognition Example with `torch>=2.0` (ultralytics#14232) * Add Discourse at https://community.ultralytics.com (ultralytics#14231) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.50` new Streamlit live inference Solution (ultralytics#14210) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: RizwanMunawar <chr043416@gmail.com> Co-authored-by: Kayzwer <68285002+Kayzwer@users.noreply.github.com> * Pin `tensorrt<=10.1.0` to fix `libnvinfer_builder_resource_win.so.10.2.0` error (ultralytics#14239) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Dockerfile install `tensorrt-cu12==10.1.0` (ultralytics#14240) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update Pose docs with keypoint explanations (ultralytics#14248) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.51` update YOLOv9 `GITHUB_ASSETS_NAMES` (ultralytics#14261) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Allow OpenVINO export from CUDA (ultralytics#14256) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix `end2end` attribute in `init_criterion` (ultralytics#14267) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add Maintaining Your Computer Vision Models Docs Page (ultralytics#14304) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `allow_empty=True` for Classify dataset class directories (ultralytics#14301) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix Annotator PIL Image size (width, height) order (ultralytics#14227) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.52` fix CenterCrop transforms for PIL Image inputs (ultralytics#14308) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Lucas Buligon Antunes <lukasbuligonantunes@gmail.com> * BaseTrainer with `find_unused_parameters=True` when using DistributedDataParallel() (ultralytics#14323) * Ultralytics Asset URL Update (ultralytics#14345) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.53` Heatmaps fix for empty images (ultralytics#14329) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: RizwanMunawar <chr043416@gmail.com> * `ultralytics 8.2.54` resolve YouTube bug with switch to `pytubefix` (ultralytics#14354) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add https://youtu.be/fLBbyhPbWzY to docs (ultralytics#14356) * `ultralytics 8.2.55` adaptive `tflite_support` logic (ultralytics#14385) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> Co-authored-by: Nguyễn Anh Bình <sometimesocrazy@gmail.com> Co-authored-by: Johnny <johnnynuca14@gmail.com> Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.56` Streamlit tracking app (ultralytics#14269) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> Co-authored-by: Nguyễn Anh Bình <sometimesocrazy@gmail.com> Co-authored-by: Johnny <johnnynuca14@gmail.com> * Fix `TORCHVISION_0_18` for `allow_empty=True` (ultralytics#14415) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add https://youtu.be/gX5kSRD56Gs to docs (ultralytics#14417) * Fix `model.save()` method to FP16 (ultralytics#14418) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.57` new Solutions Tests and Docs (ultralytics#14408) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Scope `tkinter` and `streamlit` in Solutions tests (ultralytics#14426) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix `_process_batch()` docstrings (ultralytics#14454) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.58` FastSAM code refactor (ultralytics#14450) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix `model` parameter in Pose, Segment dataset train examples (ultralytics#14505) * Warn on `save_hybrid=True` (ultralytics#14484) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Remove redundant assignment (ultralytics#14498) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update `FastSAM` and `SAM` docs (ultralytics#14499) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add Kaggle Integrations Docs Page (ultralytics#14487) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix `Enable Tracking` Button and Optimize FPS in Streamlit Application (ultralytics#14508) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.59` use `Results.save_txt` for validation (ultralytics#14496) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Allow `test-dev2017.txt` for val (ultralytics#14519) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add https://youtu.be/isc06_9qnM0 to docs (ultralytics#14525) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix NCNN multiple-volumes PNNX download bug (ultralytics#14533) * `ultralytics 8.2.60` refactor `process_mask_upsample` (ultralytics#14474) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix Multi-GPU trained model export (ultralytics#14551) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix Docs pretty `__init__.py` URLs (ultralytics#14550) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add https://youtu.be/ziJR01lKnio to docs (ultralytics#14554) * Add Custom CLIP Model Download Path (ultralytics#14517) Co-authored-by: wangsrGit119 <1215618342@email.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.61` fix `num_threads` for CPU training (ultralytics#14544) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Pin `pytubefix==6.3.4` for YouTube fix (ultralytics#14571) * Replace `enumerate` + index with `zip()` (ultralytics#14574) * Engine Model and Results Docs improvements (ultralytics#14564) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Felipe Parodi <fparodi@pennmedicine.upenn.edu> * Add Streamlit Inference Python `model` arg (ultralytics#14563) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.62` add Explorer CLI `model` and `data` args (ultralytics#14581) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Mohammed Yasin <32206511+Y-T-G@users.noreply.github.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix Docs plaintext link scan (ultralytics#14583) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> * Plaintext negative lookbehind scan (ultralytics#14601) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Replace enumerate with zip in models/yolo (ultralytics#14599) * `ultralytics 8.2.63` refactor `FastSAMPredictor` (ultralytics#14582) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com> Co-authored-by: Laughing-q <1185102784@qq.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update format.yml for PAT (ultralytics#14608) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix OBB Docs page commas error (ultralytics#14609) * Add NAS autodownload (ultralytics#14627) * Windows `torch==2.4.0` Segment `augment=True` failed test bypass (ultralytics#14637) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix torchvision InterpolationMode warnings (ultralytics#14632) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> * Patch `torch.load(..., weights_only=False)` to reduce warnings (ultralytics#14638) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Attempt to fix NAS models inference (ultralytics#14630) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> * Fix Streamlit Inference model suffix bug (ultralytics#14621) Co-authored-by: malopez <miguelangel.lopez@solute.es> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix `torch.amp.autocast('cuda')` warnings (ultralytics#14633) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> * `ultralytics 8.2.64` YOLOv10 SavedModel, TFlite, and GraphDef export (ultralytics#14572) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * New AGLU activation module (ultralytics#14644) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fix `torch.amp` has no attribute `GradScaler` (ultralytics#14647) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add warning when `cache_ram` works with classify (ultralytics#14650) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Allow `agnostic_nms` option for validation (ultralytics#14675) * `ultralytics 8.2.65` fix YouTube throttling bug (ultralytics#14684) * Remove duplicate `make_divisible` function (ultralytics#14690) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add https://youtu.be/_gRqR-miFPE to docs (ultralytics#14698) * Updates `save_period` to include first epoch (ultralytics#14700) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add compatible `tensorstore` versions for `aarch64` (ultralytics#14697) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.66` HUB model autodownload (ultralytics#14702) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `SETTINGS` type and key checks (ultralytics#14703) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add JetPack6 Docker for NVIDIA Jetson Orin Series (ultralytics#14707) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.67` new NVIDIA Jetson Orin Jetpack 6 Docker image (ultralytics#14740) * Simplify Dockerfile `WORKDIR` (ultralytics#14750) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.68` new HUB GCP region latency tests (ultralytics#14753) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add JupyterLab Integrations Docs Page (ultralytics#14755) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fixed OpenVINO Docs formatting (ultralytics#14773) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Add IBM Watsonx Integrations Docs Page (ultralytics#14785) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Update Ultralytics issue templates (ultralytics#14718) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Burhan <62214284+Burhan-Q@users.noreply.github.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Eliminate `set()` to avoid `set()` + `list()` op (ultralytics#14745) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.69` FastSAM prompt inference refactor (ultralytics#14724) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update comet from init() to login() (ultralytics#14793) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * New Meta Segment Anything Model 2 (SAM2) Docs page (ultralytics#14794) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * SAM2 mkdocs.yml fix (ultralytics#14796) * `py-cpuinfo` Exception context manager fix (ultralytics#14814) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add https://youtu.be/_gRqR-miFPE & https://youtu.be/CfbHwPG01cE to docs (ultralytics#14817) * `ultralytics 8.2.70` Segment Anything Model 2 (SAM 2) (ultralytics#14813) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Exported model batch size validation fix (ultralytics#14845) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix `model` in Pose, Segment datasets train FAQ sections (ultralytics#14877) * Fixed `box_label` docstrings (ultralytics#14866) Co-authored-by: dearMOMO <34231503+dearMOMO@users.noreply.github.com> * Add missing CLI `yolo` commands for `TASK` and `MODE` in Docs - Quickstart and CLI Guide (ultralytics#14882) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update SAM 2 docs (ultralytics#14864) * Increase Dockerfile build `max-parallel` (ultralytics#14892) * Fix incorrect CLI commands in Datasets Docs (ultralytics#14889) * Add Segment masks to YOLO-Seg labels converter (ultralytics#14557) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.71` Multinode DDP training (ultralytics#14879) Co-authored-by: Haris Rehman <haris.rehman.cowlar@gmail.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Delete Docker Cache before building Image (ultralytics#14894) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * ROS quickstart meta description (ultralytics#14932) * Fix the docstring of xywhr2xyxyxyxy (ultralytics#14934) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fixed `circle_label` and `text_label` docstrings (ultralytics#14909) * Fix `is_url()` and `check_disk_space()` docstrings in downloads.py (ultralytics#14923) * Update MLP module for RTDETR backward compatibility (ultralytics#14901) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Corrects CONTRIBUTING.md relative link to CLA.md (ultralytics#14857) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.72` SAM 2 multiple-`bboxes` support (ultralytics#14928) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.73` Meta SAM2 Refactor (ultralytics#14867) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Ignore Vimeo 401 'unauthorized' errors (ultralytics#14980) * fix example for plotting Ray Tune history (ultralytics#14970) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update NVIDIA Jetson Docs with JetPack 6 (ultralytics#14939) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> * Fix OpenVINO Export Docs (ultralytics#14918) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.74` add `fuse_score=True` BoT-SORT and ByteTrack arg (ultralytics#14965) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Fix Docker git permissions (ultralytics#14995) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Dedicated Inference API Docs (ultralytics#14992) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update HUB Inference API Docs (ultralytics#15035) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Sergiu Waxmann <waxmann.sergiu@me.com> * Add `allow_background_images=True` in split_dota.py (ultralytics#15037) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * `ultralytics 8.2.75` new Docs author profiles (ultralytics#15050) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update Docs CSS (ultralytics#15062) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Use macros for Docs tables de-duplication (ultralytics#14990) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Delete `/macros` dir from Docs site (ultralytics#15068) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Delete macros from sitemap.xml (ultralytics#15105) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Add https://youtu.be/EeDd5P4eS6A to docs (ultralytics#15107) * Optimized SAHI video inference (ultralytics#15183) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update `convert_segment_masks_to_yolo_seg` to support custom datasets (ultralytics#15176) * `ultralytics 8.2.76` Autobackend TensorRT/Triton Segment metadata fix (ultralytics#15185) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update Contributing guidelines (ultralytics#15373) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Fixed multiscale preprocess_batch (ultralytics#15392) * Improve trainer DDP device handling (ultralytics#15383) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update Conda CI (ultralytics#15443) * Update Tracker docstrings (ultralytics#15469) Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * `ultralytics 8.2.77` new `color_mode=instance` plot arg (ultralytics#15034) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update yolov8_sahi.py (ultralytics#15470) Co-authored-by: UltralyticsAssistant <web@ultralytics.com> * Ultralytics Actions JSON, CSS and autolabel support (ultralytics#15599) Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Remove unnecessary assignments (ultralytics#15582) * YOLO Vision 2024 updates https://ultralytics.com/events/yolovision (ultralytics#15602) Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> --------- Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Adamcode <86471590+sunmooncode@users.noreply.github.com> Co-authored-by: UltralyticsAssistant <web@ultralytics.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: Sergiu Waxmann <47978446+sergiuwaxmann@users.noreply.github.com> Co-authored-by: Lakshantha Dissanayake <lakshanthad@yahoo.com> Co-authored-by: Muhammad Rizwan Munawar <muhammadrizwanmunawar123@gmail.com> Co-authored-by: Richard Abrich <richard.abrich@gmail.com> Co-authored-by: Abirami Vina <abirami.vina@gmail.com> Co-authored-by: Kayzwer <68285002+Kayzwer@users.noreply.github.com> Co-authored-by: inisis <desmond.yao@buaa.edu.cn> Co-authored-by: inisis <46103969+inisis@users.noreply.github.com> Co-authored-by: RainRat <rainrat78@yahoo.ca> Co-authored-by: czy10383 <81451461+czy10383@users.noreply.github.com> Co-authored-by: kubade-ashish <ashish.ak@perfios.com> Co-authored-by: Burhan <62214284+Burhan-Q@users.noreply.github.com> Co-authored-by: Ultralytics Assistant <135830346+UltralyticsAssistant@users.noreply.github.com> Co-authored-by: Paula Derrenger <107626595+pderrenger@users.noreply.github.com> Co-authored-by: Ivor Zhu <130829914+IvorZhu331@users.noreply.github.com> Co-authored-by: Ahmed Mahfouz <ahmelsamahy@gmail.com> Co-authored-by: Lakshantha <lakshantha@ultralytics.com> Co-authored-by: alexwine36 <alexwine36@gmail.com> Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com> Co-authored-by: Francesco Mattioli <Francesco.mttl@gmail.com> Co-authored-by: Laughing-q <1185102784@qq.com> Co-authored-by: xiaoluohao <blackhu@users.noreply.github.com> Co-authored-by: Alex Pasquali <alexpasquali98@gmail.com> Co-authored-by: Mughees Ahmad <itsmughees@gmail.com> Co-authored-by: Erlangga Yudi Pradana <erlanggayudipradana@gmail.com> Co-authored-by: Muhammad Amir Abdurrozaq <m.amir.hs19@gmail.com> Co-authored-by: DeepDiver <zhaoxu1015@gmail.com> Co-authored-by: jackwolfey <91648300+jackwolfey@users.noreply.github.com> Co-authored-by: bobyard-com <154289614+bobyard-com@users.noreply.github.com> Co-authored-by: zzzer <48149018+zhixuwei@users.noreply.github.com> Co-authored-by: AAOMM <52826299+Chayanonjackal@users.noreply.github.com> Co-authored-by: fatih c. akyon <34196005+fcakyon@users.noreply.github.com> Co-authored-by: RizwanMunawar <chr043416@gmail.com> Co-authored-by: JF Chen <k-2feng@hotmail.com> Co-authored-by: Antônio Martos Harres <tom.mharres@gmail.com> Co-authored-by: Alejandro Casanova <36633619+Alejandro-Casanova@users.noreply.github.com> Co-authored-by: Alexis IMBERT <97242148+Alexis-IMBERT@users.noreply.github.com> Co-authored-by: Sheffey <57262511+SheffeyG@users.noreply.github.com> Co-authored-by: Lucas Buligon Antunes <lukasbuligonantunes@gmail.com> Co-authored-by: Chia-Hsiang Tsai <84863554+Tsai-chia-hsiang@users.noreply.github.com> Co-authored-by: Nguyễn Anh Bình <sometimesocrazy@gmail.com> Co-authored-by: Johnny <johnnynuca14@gmail.com> Co-authored-by: Haonan Liu <hnliu_2@stu.xidian.edu.cn> Co-authored-by: Aryan Jassal <84617406+aryanjassal@users.noreply.github.com> Co-authored-by: zhiqiang yang <89863442+yang-0201@users.noreply.github.com> Co-authored-by: Ariel Kukulanski <akukulanski@gmail.com> Co-authored-by: Mohammed Yasin <32206511+Y-T-G@users.noreply.github.com> Co-authored-by: suke <32221726+wangsrGit119@users.noreply.github.com> Co-authored-by: wangsrGit119 <1215618342@email.com> Co-authored-by: Felipe Parodi <fparodi@pennmedicine.upenn.edu> Co-authored-by: rulosant <rulosanti@gmail.com> Co-authored-by: Miguel Angel Lopez <malopez00@gmail.com> Co-authored-by: malopez <miguelangel.lopez@solute.es> Co-authored-by: Hassan Ghaffari <53246858+hasanghaffari93@users.noreply.github.com> Co-authored-by: Kevin <kevin.qh.lin@gmail.com> Co-authored-by: Jan Knobloch <116908874+jk4e@users.noreply.github.com> Co-authored-by: dearMOMO <34231503+dearMOMO@users.noreply.github.com> Co-authored-by: Haris Rehman <115916265+Harri200191@users.noreply.github.com> Co-authored-by: Haris Rehman <haris.rehman.cowlar@gmail.com> Co-authored-by: Zeel B Patel <patel_zeel@iitgn.ac.in> Co-authored-by: Maxi <60036186+mfloto@users.noreply.github.com> Co-authored-by: Sergiu Waxmann <waxmann.sergiu@me.com> Co-authored-by: Galasnow <854932917@qq.com> Co-authored-by: alanZee <92136487+alanZee@users.noreply.github.com>

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com> Co-authored-by: DeepDiver <zhaoxu1015@gmail.com> Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com>

zxDeepDiver and others added 2 commits June 24, 2024 17:37

Bugfix: Changed the order of sigmoid binarization and upsampling to a…

f24a45a

…chieve better segmentation edges. (#13912) Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Update Dockerfile-conda

f051067

glenn-jocher added 3 commits June 24, 2024 18:07

Docker (#13941)

e431693

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

conda install pytorch torchvision pytorch-cuda=12.1

887cd29

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Merge branch 'main' into segment-fix

94dad41

glenn-jocher changed the title ~~Segment Sigmoid() fix~~ ultralytics 8.2.42 Segment Sigmoid() fix Jun 24, 2024

glenn-jocher changed the title ~~ultralytics 8.2.42 Segment Sigmoid() fix~~ ultralytics 8.2.42 YOLO Segment sigmoid() fix Jun 24, 2024

Update __init__.py

3c2aa07

glenn-jocher added 8 commits June 24, 2024 20:35

conda install pytorch torchvision pytorch-cuda=12.1

c4e2d81

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Update ci.yaml

bc1ffb0

Update ci.yaml

8557141

conda install pytorch torchvision pytorch-cuda=12.1

9dc5ccd

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Merge remote-tracking branch 'origin/segment-fix' into segment-fix

d882166

conda install pytorch torchvision pytorch-cuda=12.1

01bd8a6

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

Update ops.py

c86a3e8

Update ops.py

a4eba8e

glenn-jocher changed the title ~~ultralytics 8.2.42 YOLO Segment sigmoid() fix~~ YOLO Segment sigmoid() fix Jun 24, 2024

Merge branch 'main' into segment-fix

da3aa13

glenn-jocher merged commit b10e0f3 into main Jun 25, 2024
16 checks passed

glenn-jocher deleted the segment-fix branch June 25, 2024 09:53

Alarmod added a commit to Alarmod/MRI_MedicalAnalysis that referenced this pull request Jul 10, 2024

Supported non sigmoid mode (see ultralytics/ultralytics#13939)

f2b81b7

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

YOLO Segment `sigmoid()` fix #13939

YOLO Segment `sigmoid()` fix #13939

glenn-jocher commented Jun 24, 2024 •

edited by github-actions bot

Loading

codecov bot commented Jun 24, 2024 •

edited

Loading

glenn-jocher commented Jun 24, 2024

glenn-jocher commented Jun 24, 2024

glenn-jocher commented Jun 24, 2024 •

edited

Loading

Laughing-q commented Jun 25, 2024

Laughing-q commented Jun 25, 2024 •

edited

Loading

glenn-jocher commented Jun 25, 2024

glenn-jocher commented Jun 25, 2024

glenn-jocher commented Jun 25, 2024

Alarmod commented Jul 6, 2024

Alarmod commented Jul 6, 2024

glenn-jocher commented Jul 6, 2024

yangkai1129 commented Aug 9, 2024

pderrenger commented Aug 9, 2024

YOLO Segment sigmoid() fix #13939

YOLO Segment sigmoid() fix #13939

Conversation

glenn-jocher commented Jun 24, 2024 • edited by github-actions bot Loading

🛠️ PR Summary

🌟 Summary

📊 Key Changes

🎯 Purpose & Impact

codecov bot commented Jun 24, 2024 • edited Loading

Codecov Report

glenn-jocher commented Jun 24, 2024

Main branch results:

glenn-jocher commented Jun 24, 2024

PR results

glenn-jocher commented Jun 24, 2024 • edited Loading

Laughing-q commented Jun 25, 2024

Laughing-q commented Jun 25, 2024 • edited Loading

glenn-jocher commented Jun 25, 2024

glenn-jocher commented Jun 25, 2024

glenn-jocher commented Jun 25, 2024

Alarmod commented Jul 6, 2024

Alarmod commented Jul 6, 2024

glenn-jocher commented Jul 6, 2024

yangkai1129 commented Aug 9, 2024

pderrenger commented Aug 9, 2024

YOLO Segment `sigmoid()` fix #13939

YOLO Segment `sigmoid()` fix #13939

glenn-jocher commented Jun 24, 2024 •

edited by github-actions bot

Loading

codecov bot commented Jun 24, 2024 •

edited

Loading

glenn-jocher commented Jun 24, 2024 •

edited

Loading

Laughing-q commented Jun 25, 2024 •

edited

Loading