YOLOV8在coco128上的训练
coco128是coco数据集的子集只有128张图片
训练代码main.py
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.yaml") # build a new model from scratch
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Use the model
model.train(data="coco128.yaml", epochs=3) # train the model
metrics = model.val() # evaluate model performance on the validation set
results = model("https://ultralytics.com/images/bus.jpg") # predict on an image
path = model.export(format="onnx") # export the model to ONNX format
(yolov8) nvidia@nvidia-desktop:~/yolov8$ python main.py
from n params module arguments
0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2]
1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2]
2 -1 1 7360 ultralytics.nn.modules.block.C2f [32, 32, 1, True]
3 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2]
4 -1 2 49664 ultralytics.nn.modules.block.C2f [64, 64, 2, True]
5 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2]
6 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True]
7 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2]
8 -1 1 460288 ultralytics.nn.modules.block.C2f [256, 256, 1, True]
9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5]
10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1]
12 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1]
13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1]
15 -1 1 37248 ultralytics.nn.modules.block.C2f [192, 64, 1]
16 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2]
17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1]
18 -1 1 123648 ultralytics.nn.modules.block.C2f [192, 128, 1]
19 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2]
20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1]
21 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1]
22 [15, 18, 21] 1 897664 ultralytics.nn.modules.head.Detect [80, [64, 128, 256]]
YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs
New https://pypi.org/project/ultralytics/8.1.1 available Update with 'pip install -U ultralytics'
Ultralytics YOLOv8.1.0 Python-3.10.13 torch-2.0.1+cu117 CUDA:0 (NVIDIA GeForce RTX 4060 Ti, 16081MiB)
engine/trainer: task=detect, mode=train, model=yolov8n.pt, data=coco128.yaml, epochs=3, time=None, patience=50, batch=16, imgsz=640, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train3, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=False, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=False, opset=None, workspace=4, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, label_smoothing=0.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, mosaic=1.0, mixup=0.0, copy_paste=0.0, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train3
from n params module arguments
0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2]
1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2]
2 -1 1 7360 ultralytics.nn.modules.block.C2f [32, 32, 1, True]
3 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2]
4 -1 2 49664 ultralytics.nn.modules.block.C2f [64, 64, 2, True]
5 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2]
6 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True]
7 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2]
8 -1 1 460288 ultralytics.nn.modules.block.C2f [256, 256, 1, True]
9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5]
10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1]
12 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1]
13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']
14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1]
15 -1 1 37248 ultralytics.nn.modules.block.C2f [192, 64, 1]
16 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2]
17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1]
18 -1 1 123648 ultralytics.nn.modules.block.C2f [192, 128, 1]
19 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2]
20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1]
21 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1]
22 [15, 18, 21] 1 897664 ultralytics.nn.modules.head.Detect [80, [64, 128, 256]]
Model summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs
Transferred 355/355 items from pretrained weights
Freezing layer 'model.22.dfl.conv.weight'
AMP: running Automatic Mixed Precision (AMP) checks with YOLOv8n...
AMP: checks passed ✅
train: Scanning /home/nvidia/datasets/coco128/labels/train2017.cache... 126 images, 2 backgrounds, 0 corrupt: 100%|██████████| 128/12
val: Scanning /home/nvidia/datasets/coco128/labels/train2017.cache... 126 images, 2 backgrounds, 0 corrupt: 100%|██████████| 128/128
Plotting labels to runs/detect/train3/labels.jpg...
optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically...
optimizer: AdamW(lr=0.000119, momentum=0.9) with parameter groups 57 weight(decay=0.0), 64 weight(decay=0.0005), 63 bias(decay=0.0)
3 epochs...
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
1/3 2.66G 1.226 1.615 1.274 178 640: 100%|██████████| 8/8 [00:03<00:00, 2.63it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 4/4 [00:00<00:00, 4.44it/
all 128 929 0.645 0.533 0.614 0.455
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
2/3 2.67G 1.225 1.514 1.268 231 640: 100%|██████████| 8/8 [00:01<00:00, 5.24it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 4/4 [00:00<00:00, 5.06it/
all 128 929 0.675 0.538 0.626 0.467
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
3/3 2.71G 1.209 1.448 1.222 178 640: 100%|██████████| 8/8 [00:01<00:00, 5.52it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 4/4 [00:00<00:00, 4.99it/
all 128 929 0.676 0.548 0.632 0.47
3 epochs completed in 0.003 hours.
Optimizer stripped from runs/detect/train3/weights/last.pt, 6.5MB
Optimizer stripped from runs/detect/train3/weights/best.pt, 6.5MB
Validating runs/detect/train3/weights/best.pt...
Ultralytics YOLOv8.1.0 Python-3.10.13 torch-2.0.1+cu117 CUDA:0 (NVIDIA GeForce RTX 4060 Ti, 16081MiB)
Model summary (fused): 168 layers, 3151904 parameters, 0 gradients, 8.7 GFLOPs
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 4/4 [00:03<00:00, 1.13it/
all 128 929 0.675 0.549 0.632 0.469
person 128 254 0.802 0.671 0.769 0.542
bicycle 128 6 0.581 0.333 0.332 0.286
car 128 46 0.844 0.217 0.285 0.178
motorcycle 128 5 0.687 0.886 0.938 0.721
airplane 128 6 0.826 0.799 0.903 0.673
bus 128 7 0.745 0.714 0.736 0.648
train 128 3 0.556 0.667 0.83 0.731
truck 128 12 1 0.353 0.498 0.304
boat 128 6 0.293 0.167 0.351 0.225
traffic light 128 14 0.696 0.168 0.202 0.139
stop sign 128 2 0.966 1 0.995 0.711
bench 128 9 0.838 0.575 0.637 0.4
bird 128 16 0.923 0.748 0.88 0.53
cat 128 4 0.866 1 0.995 0.835
dog 128 9 0.704 0.778 0.821 0.633
horse 128 2 0.536 1 0.995 0.511
elephant 128 17 0.851 0.765 0.879 0.669
bear 128 1 0.631 1 0.995 0.995
zebra 128 4 0.857 1 0.995 0.965
giraffe 128 9 0.899 0.993 0.973 0.714
backpack 128 6 0.605 0.333 0.392 0.234
umbrella 128 18 0.71 0.5 0.663 0.453
handbag 128 19 0.518 0.0582 0.18 0.0947
tie 128 7 0.69 0.642 0.641 0.457
suitcase 128 4 0.641 1 0.828 0.596
frisbee 128 5 0.567 0.8 0.759 0.663
skis 128 1 0.473 1 0.995 0.497
snowboard 128 7 0.661 0.714 0.755 0.486
sports ball 128 6 0.703 0.406 0.503 0.29
kite 128 10 0.811 0.5 0.595 0.208
baseball bat 128 4 0.574 0.362 0.414 0.174
baseball glove 128 7 0.672 0.429 0.429 0.295
skateboard 128 5 0.776 0.6 0.6 0.44
tennis racket 128 7 0.742 0.415 0.529 0.367
bottle 128 18 0.508 0.389 0.394 0.24
wine glass 128 16 0.584 0.562 0.581 0.361
cup 128 36 0.632 0.287 0.4 0.279
fork 128 6 0.597 0.167 0.264 0.193
knife 128 16 0.643 0.5 0.612 0.351
spoon 128 22 0.554 0.182 0.332 0.18
bowl 128 28 0.676 0.571 0.615 0.495
banana 128 1 0 0 0.166 0.048
sandwich 128 2 0.394 0.5 0.497 0.497
orange 128 4 1 0.313 0.995 0.666
broccoli 128 11 0.471 0.182 0.247 0.221
carrot 128 24 0.739 0.458 0.658 0.411
hot dog 128 2 0.65 0.95 0.828 0.796
pizza 128 5 0.689 1 0.995 0.86
donut 128 14 0.639 1 0.946 0.859
cake 128 4 0.658 1 0.995 0.88
chair 128 35 0.514 0.514 0.468 0.262
couch 128 6 0.746 0.493 0.673 0.497
potted plant 128 14 0.739 0.643 0.722 0.484
bed 128 3 0.768 0.667 0.806 0.636
dining table 128 13 0.484 0.615 0.504 0.415
toilet 128 2 1 0.869 0.995 0.946
tv 128 2 0.384 0.5 0.695 0.656
laptop 128 3 1 0 0.696 0.544
mouse 128 2 1 0 0.0527 0.00527
remote 128 8 0.849 0.5 0.583 0.507
cell phone 128 8 0 0 0.0688 0.0465
microwave 128 3 0.63 0.667 0.863 0.719
oven 128 5 0.472 0.4 0.338 0.269
sink 128 6 0.367 0.167 0.232 0.159
refrigerator 128 5 0.688 0.4 0.647 0.525
book 128 29 0.621 0.114 0.328 0.187
clock 128 9 0.779 0.781 0.893 0.713
vase 128 2 0.418 1 0.828 0.745
scissors 128 1 1 0 0.249 0.0746
teddy bear 128 21 0.884 0.381 0.64 0.429
toothbrush 128 5 0.9 0.6 0.786 0.503
Speed: 3.0ms preprocess, 1.8ms inference, 0.0ms loss, 1.2ms postprocess per image
Results saved to runs/detect/train3
Ultralytics YOLOv8.1.0 Python-3.10.13 torch-2.0.1+cu117 CUDA:0 (NVIDIA GeForce RTX 4060 Ti, 16081MiB)
Model summary (fused): 168 layers, 3151904 parameters, 0 gradients, 8.7 GFLOPs
val: Scanning /home/nvidia/datasets/coco128/labels/train2017.cache... 126 images, 2 backgrounds, 0 corrupt: 100%|██████████| 128/128
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 8/8 [00:03<00:00, 2.06it/
all 128 929 0.658 0.548 0.627 0.466
person 128 254 0.808 0.68 0.773 0.542
bicycle 128 6 0.569 0.333 0.326 0.283
car 128 46 0.805 0.217 0.285 0.178
motorcycle 128 5 0.689 0.896 0.898 0.697
airplane 128 6 0.827 0.803 0.903 0.681
bus 128 7 0.681 0.714 0.736 0.648
train 128 3 0.551 0.667 0.83 0.731
truck 128 12 1 0.374 0.494 0.295
boat 128 6 0.261 0.167 0.324 0.14
traffic light 128 14 0.696 0.168 0.202 0.139
stop sign 128 2 0.93 1 0.995 0.711
bench 128 9 0.842 0.596 0.636 0.4
bird 128 16 0.853 0.75 0.866 0.5
cat 128 4 0.863 1 0.995 0.835
dog 128 9 0.682 0.778 0.821 0.626
horse 128 2 0.53 1 0.995 0.515
elephant 128 17 0.849 0.765 0.879 0.669
bear 128 1 0.626 1 0.995 0.995
zebra 128 4 0.854 1 0.995 0.965
giraffe 128 9 0.744 0.971 0.943 0.732
backpack 128 6 0.613 0.333 0.391 0.238
umbrella 128 18 0.675 0.5 0.657 0.453
handbag 128 19 0.554 0.0698 0.173 0.0932
tie 128 7 0.698 0.665 0.641 0.457
suitcase 128 4 0.633 1 0.828 0.596
frisbee 128 5 0.563 0.8 0.759 0.663
skis 128 1 0.461 1 0.995 0.497
snowboard 128 7 0.657 0.714 0.757 0.484
sports ball 128 6 0.705 0.411 0.502 0.274
kite 128 10 0.801 0.5 0.598 0.206
baseball bat 128 4 0.577 0.25 0.348 0.174
baseball glove 128 7 0.638 0.429 0.429 0.295
skateboard 128 5 0.871 0.6 0.6 0.44
tennis racket 128 7 0.744 0.419 0.529 0.365
bottle 128 18 0.469 0.389 0.358 0.217
wine glass 128 16 0.574 0.562 0.554 0.347
cup 128 36 0.566 0.278 0.401 0.286
fork 128 6 0.59 0.167 0.228 0.195
knife 128 16 0.563 0.5 0.587 0.358
spoon 128 22 0.633 0.182 0.331 0.186
bowl 128 28 0.737 0.643 0.658 0.498
banana 128 1 0 0 0.0995 0.042
sandwich 128 2 0.161 0.241 0.497 0.497
orange 128 4 1 0.321 0.995 0.666
broccoli 128 11 0.501 0.182 0.257 0.207
carrot 128 24 0.728 0.558 0.653 0.418
hot dog 128 2 0.649 0.946 0.828 0.796
pizza 128 5 0.717 1 0.995 0.86
donut 128 14 0.637 1 0.94 0.854
cake 128 4 0.61 1 0.945 0.845
chair 128 35 0.484 0.543 0.472 0.258
couch 128 6 0.613 0.5 0.745 0.578
potted plant 128 14 0.716 0.643 0.723 0.481
bed 128 3 0.757 0.667 0.913 0.661
dining table 128 13 0.457 0.615 0.496 0.391
toilet 128 2 1 0.875 0.995 0.946
tv 128 2 0.378 0.5 0.695 0.656
laptop 128 3 1 0 0.605 0.484
mouse 128 2 1 0 0.0698 0.00698
remote 128 8 0.845 0.5 0.605 0.514
cell phone 128 8 0 0 0.0696 0.0469
microwave 128 3 0.617 0.667 0.863 0.733
oven 128 5 0.431 0.4 0.339 0.27
sink 128 6 0.378 0.167 0.18 0.131
refrigerator 128 5 0.684 0.4 0.65 0.517
book 128 29 0.637 0.122 0.343 0.195
clock 128 9 0.78 0.788 0.894 0.734
vase 128 2 0.407 1 0.828 0.745
scissors 128 1 1 0 0.249 0.0746
teddy bear 128 21 0.883 0.381 0.634 0.42
toothbrush 128 5 0.636 0.6 0.736 0.468
Speed: 3.4ms preprocess, 13.3ms inference, 0.0ms loss, 3.5ms postprocess per image
Results saved to runs/detect/train32
Found https://ultralytics.com/images/bus.jpg locally at bus.jpg
image 1/1 /home/nvidia/yolov8/bus.jpg: 640x480 4 persons, 1 bus, 1 stop sign, 17.2ms
Speed: 4.8ms preprocess, 17.2ms inference, 3.0ms postprocess per image at shape (1, 3, 640, 480)
Ultralytics YOLOv8.1.0 Python-3.10.13 torch-2.0.1+cu117 CPU (Intel Xeon E5-2686 v4 2.30GHz)
PyTorch: starting from 'runs/detect/train3/weights/best.pt' with input shape (1, 3, 640, 640) BCHW and output shape(s) (1, 84, 8400) (6.2 MB)
ONNX: starting export with onnx 1.15.0 opset 17...
============= Diagnostic Run torch.onnx.export version 2.0.1+cu117 =============
verbose: False, log level: Level.ERROR
======================= 0 NONE 0 NOTE 0 WARNING 0 ERROR ========================
ONNX: export success ✅ 0.9s, saved as 'runs/detect/train3/weights/best.onnx' (12.2 MB)
Export complete (2.4s)
Results saved to /home/nvidia/yolov8/runs/detect/train3/weights
Predict: yolo predict task=detect model=runs/detect/train3/weights/best.onnx imgsz=640
Validate: yolo val task=detect model=runs/detect/train3/weights/best.onnx imgsz=640 data=/home/nvidia/anaconda3/envs/yolov8/lib/python3.10/site-packages/ultralytics/cfg/datasets/coco128.yaml
Visualize: https://netron.app 可视化网站