Cloning into 'yolov5'... remote: Enumerating objects: 64, done. remote: Counting objects: 100% (64/64), done. remote: Compressing objects: 100% (42/42), done. remote: Total 1711 (delta 39), reused 42 (delta 22), pack-reused 1647 Receiving objects: 100% (1711/1711), 4.62 MiB | 1.13 MiB/s, done. Resolving deltas: 100% (1152/1152), done. Checking connectivity... done. % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 279 0 279 0 0 1017 0 --:--:-- --:--:-- --:--:-- 1014 100 408 0 408 0 0 362 0 --:--:-- 0:00:01 --:--:-- 61000 0 0 0 0 0 0 0 0 --:--:-- 0:00:01 --:--:-- 0 0 0 0 0 0 0 0 0 --:--:-- 0:00:02 --:--:-- 0 100 21.0M 0 21.0M 0 0 7429k 0 --:--:-- 0:00:02 --:--:-- 7429k Downloading https://drive.google.com/uc?export=download&id=1n_oKgR81BJtqk75b00eAjdv03qVCQn2f as coco128.zip... unzipping... Done (7.3s) ***************************************** Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. ***************************************** Using CUDA Apex device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device1 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Using CUDA Apex device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device1 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Namespace(batch_size=8, bucket='', cache_images=False, cfg='models/yolov5s.yaml', data='data/coco128.yaml', device='0,1', epochs=3, evolve=False, hyp='', img_size=[320, 320], local_rank=1, multi_scale=False, name='', noautoanchor=False, nosave=False, notest=False, rect=False, resume=False, single_cls=False, sync_bn=False, total_batch_size=16, weights='yolov5s.pt', world_size=2) Hyperparameters {'optimizer': 'SGD', 'lr0': 0.01, 'momentum': 0.937, 'weight_decay': 0.0005, 'giou': 0.05, 'cls': 0.58, 'cls_pw': 1.0, 'obj': 1.0, 'obj_pw': 1.0, 'iou_t': 0.2, 'anchor_t': 4.0, 'fl_gamma': 0.0, 'hsv_h': 0.014, 'hsv_s': 0.68, 'hsv_v': 0.36, 'degrees': 0.0, 'translate': 0.0, 'scale': 0.5, 'shear': 0.0} Namespace(batch_size=8, bucket='', cache_images=False, cfg='models/yolov5s.yaml', data='data/coco128.yaml', device='0,1', epochs=3, evolve=False, hyp='', img_size=[320, 320], local_rank=0, multi_scale=False, name='', noautoanchor=False, nosave=False, notest=False, rect=False, resume=False, single_cls=False, sync_bn=False, total_batch_size=16, weights='yolov5s.pt', world_size=2) Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/ from n params module arguments 0 -1 1 3520 models.common.Focus [3, 32, 3] 1 -1 1 18560 models.common.Conv [32, 64, 3, 2] 2 -1 1 19904 models.common.BottleneckCSP [64, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 1 161152 models.common.BottleneckCSP [128, 128, 3] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 1 641792 models.common.BottleneckCSP [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]] 9 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 10 -1 1 131584 models.common.Conv [512, 256, 1, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 models.common.Concat [1] 13 -1 1 378624 models.common.BottleneckCSP [512, 256, 1, False] 14 -1 1 33024 models.common.Conv [256, 128, 1, 1] 15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 16 [-1, 4] 1 0 models.common.Concat [1] 17 -1 1 95104 models.common.BottleneckCSP [256, 128, 1, False] 18 -1 1 32895 torch.nn.modules.conv.Conv2d [128, 255, 1, 1] 19 -2 1 147712 models.common.Conv [128, 128, 3, 2] 20 [-1, 14] 1 0 models.common.Concat [1] 21 -1 1 313088 models.common.BottleneckCSP [256, 256, 1, False] 22 -1 1 65535 torch.nn.modules.conv.Conv2d [256, 255, 1, 1] 23 -2 1 590336 models.common.Conv [256, 256, 3, 2] 24 [-1, 10] 1 0 models.common.Concat [1] 25 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 26 -1 1 130815 torch.nn.modules.conv.Conv2d [512, 255, 1, 1] 27 [-1, 22, 18] 1 0 models.yolo.Detect [80, [[116, 90, 156, 198, 373, 326], [30, 61, 62, 45, 59, 119], [10, 13, 16, 30, 33, 23]]] Model Summary: 191 layers, 7.46816e+06 parameters, 7.46816e+06 gradients Optimizer groups: 62 .bias, 70 conv.weight, 59 other Hyperparameters {'optimizer': 'SGD', 'lr0': 0.01, 'momentum': 0.937, 'weight_decay': 0.0005, 'giou': 0.05, 'cls': 0.58, 'cls_pw': 1.0, 'obj': 1.0, 'obj_pw': 1.0, 'iou_t': 0.2, 'anchor_t': 4.0, 'fl_gamma': 0.0, 'hsv_h': 0.014, 'hsv_s': 0.68, 'hsv_v': 0.36, 'degrees': 0.0, 'translate': 0.0, 'scale': 0.5, 'shear': 0.0} from n params module arguments 0 -1 1 3520 models.common.Focus [3, 32, 3] 1 -1 1 18560 models.common.Conv [32, 64, 3, 2] 2 -1 1 19904 models.common.BottleneckCSP [64, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 1 161152 models.common.BottleneckCSP [128, 128, 3] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 1 641792 models.common.BottleneckCSP [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]] 9 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 10 -1 1 131584 models.common.Conv [512, 256, 1, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 models.common.Concat [1] 13 -1 1 378624 models.common.BottleneckCSP [512, 256, 1, False] 14 -1 1 33024 models.common.Conv [256, 128, 1, 1] 15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 16 [-1, 4] 1 0 models.common.Concat [1] 17 -1 1 95104 models.common.BottleneckCSP [256, 128, 1, False] 18 -1 1 32895 torch.nn.modules.conv.Conv2d [128, 255, 1, 1] 19 -2 1 147712 models.common.Conv [128, 128, 3, 2] 20 [-1, 14] 1 0 models.common.Concat [1] 21 -1 1 313088 models.common.BottleneckCSP [256, 256, 1, False] 22 -1 1 65535 torch.nn.modules.conv.Conv2d [256, 255, 1, 1] 23 -2 1 590336 models.common.Conv [256, 256, 3, 2] 24 [-1, 10] 1 0 models.common.Concat [1] 25 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 26 -1 1 130815 torch.nn.modules.conv.Conv2d [512, 255, 1, 1] 27 [-1, 22, 18] 1 0 models.yolo.Detect [80, [[116, 90, 156, 198, 373, 326], [30, 61, 62, 45, 59, 119], [10, 13, 16, 30, 33, 23]]] Model Summary: 191 layers, 7.46816e+06 parameters, 7.46816e+06 gradients Optimizer groups: 62 .bias, 70 conv.weight, 59 other Downloading https://drive.google.com/uc?export=download&id=1R5T6rIyy3lLwgFXNms8whc-387H0tMQO as yolov5s.pt... % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 279 0 279 0 0 757 0 --:--:-- --:--:-- --:--:-- 758 100 408 0 408 0 0 371 0 --:--:-- 0:00:01 --:--:-- 119k 0 0 0 0 0 0 0 0 --:--:-- 0:00:01 --:--:-- 0 0 0 0 0 0 0 0 0 --:--:-- 0:00:02 --:--:-- 0 100 14.4M 0 14.4M 0 0 4661k 0 --:--:-- 0:00:03 --:--:-- 43.8M Done (5.7s) Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Analyzing anchors... Best Possible Recall (BPR) = 0.9591. Attempting to generate improved anchors, please wait... WARNING: Extremely small objects found. 35 of 929 labels are < 3 pixels in width or height. Running kmeans for 9 anchors on 927 points... thr=0.25: 0.9731 best possible recall, 3.74 anchors past thr n=9, img_size=320, metric_all=0.261/0.654-mean/best, past_thr=0.471-mean: 9,12, 32,19, 27,47, 73,43, 53,91, 77,161, 161,107, 174,237, 299,195 Evolving anchors with Genetic Algorithm: fitness = 0.6627: 100%|█| 1000/1000 [00 thr=0.25: 0.9957 best possible recall, 3.79 anchors past thr n=9, img_size=320, metric_all=0.262/0.662-mean/best, past_thr=0.473-mean: 7,8, 17,12, 24,31, 58,39, 50,86, 71,146, 148,116, 144,240, 293,213 New anchors saved to model. Update model *.yaml to use these anchors in the future. Image sizes 320 train, 320 test Using 8 dataloader workers Starting training for 3 epochs... Epoch gpu_mem GIoU obj cls total targets img_size 0/2 0.719G 0.132 0.1036 0.05205 0.2876 55 320 Class Images Targets P R mAP@.5 all 128 929 0.138 0.603 0.337 0.132 Epoch gpu_mem GIoU obj cls total targets img_size 1/2 1.74G 0.1227 0.1017 0.04107 0.2655 94 320 Class Images Targets P R mAP@.5 all 128 929 0.152 0.619 0.426 0.217 Epoch gpu_mem GIoU obj cls total targets img_size 2/2 1.74G 0.1122 0.07557 0.05006 0.2379 85 320 Class Images Targets P R mAP@.5 all 128 929 0.14 0.636 0.448 0.225 Optimizer stripped from runs/exp0/weights/last.pt, 15.1MB Optimizer stripped from runs/exp0/weights/best.pt, 15.1MB No handles with labels found to put in legend. 3 epochs completed in 0.006 hours. Using CUDA Apex device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device1 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Namespace(batch_size=16, bucket='', cache_images=False, cfg='models/yolov5s.yaml', data='data/coco128.yaml', device='0,1', epochs=3, evolve=False, hyp='', img_size=[320, 320], local_rank=-1, multi_scale=False, name='', noautoanchor=False, nosave=False, notest=False, rect=False, resume=False, single_cls=False, sync_bn=False, total_batch_size=16, weights='yolov5s.pt', world_size=1) Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/ Hyperparameters {'optimizer': 'SGD', 'lr0': 0.01, 'momentum': 0.937, 'weight_decay': 0.0005, 'giou': 0.05, 'cls': 0.58, 'cls_pw': 1.0, 'obj': 1.0, 'obj_pw': 1.0, 'iou_t': 0.2, 'anchor_t': 4.0, 'fl_gamma': 0.0, 'hsv_h': 0.014, 'hsv_s': 0.68, 'hsv_v': 0.36, 'degrees': 0.0, 'translate': 0.0, 'scale': 0.5, 'shear': 0.0} from n params module arguments 0 -1 1 3520 models.common.Focus [3, 32, 3] 1 -1 1 18560 models.common.Conv [32, 64, 3, 2] 2 -1 1 19904 models.common.BottleneckCSP [64, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 1 161152 models.common.BottleneckCSP [128, 128, 3] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 1 641792 models.common.BottleneckCSP [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]] 9 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 10 -1 1 131584 models.common.Conv [512, 256, 1, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 models.common.Concat [1] 13 -1 1 378624 models.common.BottleneckCSP [512, 256, 1, False] 14 -1 1 33024 models.common.Conv [256, 128, 1, 1] 15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 16 [-1, 4] 1 0 models.common.Concat [1] 17 -1 1 95104 models.common.BottleneckCSP [256, 128, 1, False] 18 -1 1 32895 torch.nn.modules.conv.Conv2d [128, 255, 1, 1] 19 -2 1 147712 models.common.Conv [128, 128, 3, 2] 20 [-1, 14] 1 0 models.common.Concat [1] 21 -1 1 313088 models.common.BottleneckCSP [256, 256, 1, False] 22 -1 1 65535 torch.nn.modules.conv.Conv2d [256, 255, 1, 1] 23 -2 1 590336 models.common.Conv [256, 256, 3, 2] 24 [-1, 10] 1 0 models.common.Concat [1] 25 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 26 -1 1 130815 torch.nn.modules.conv.Conv2d [512, 255, 1, 1] 27 [-1, 22, 18] 1 0 models.yolo.Detect [80, [[116, 90, 156, 198, 373, 326], [30, 61, 62, 45, 59, 119], [10, 13, 16, 30, 33, 23]]] Model Summary: 191 layers, 7.46816e+06 parameters, 7.46816e+06 gradients Optimizer groups: 62 .bias, 70 conv.weight, 59 other Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Analyzing anchors... Best Possible Recall (BPR) = 0.9591. Attempting to generate improved anchors, please wait... WARNING: Extremely small objects found. 35 of 929 labels are < 3 pixels in width or height. Running kmeans for 9 anchors on 927 points... thr=0.25: 0.9752 best possible recall, 3.75 anchors past thr n=9, img_size=320, metric_all=0.261/0.655-mean/best, past_thr=0.473-mean: 8,11, 26,19, 30,47, 75,43, 52,92, 77,162, 160,107, 174,237, 299,195 Evolving anchors with Genetic Algorithm: fitness = 0.6584: 100%|█| 1000/1000 [00 thr=0.25: 0.9828 best possible recall, 3.84 anchors past thr n=9, img_size=320, metric_all=0.266/0.660-mean/best, past_thr=0.474-mean: 9,10, 25,15, 30,41, 64,43, 50,86, 69,147, 144,108, 183,196, 314,210 New anchors saved to model. Update model *.yaml to use these anchors in the future. Image sizes 320 train, 320 test Using 8 dataloader workers Starting training for 3 epochs... Epoch gpu_mem GIoU obj cls total targets img_size 0/2 0.677G 0.1284 0.1091 0.05425 0.2918 170 320 Class Images Targets P R mAP@.5 all 128 929 0.138 0.636 0.352 0.137 Epoch gpu_mem GIoU obj cls total targets img_size 1/2 1.24G 0.1186 0.08992 0.04769 0.2562 193 320 Class Images Targets P R mAP@.5 all 128 929 0.148 0.626 0.435 0.22 Epoch gpu_mem GIoU obj cls total targets img_size 2/2 1.24G 0.1099 0.09116 0.04624 0.2473 200 320 Class Images Targets P R mAP@.5 all 128 929 0.143 0.645 0.449 0.236 Optimizer stripped from runs/exp1/weights/last.pt, 15.1MB Optimizer stripped from runs/exp1/weights/best.pt, 15.1MB No handles with labels found to put in legend. 3 epochs completed in 0.008 hours. Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.4, device='0,1', img_size=640, iou_thres=0.5, output='inference/output', save_txt=False, source='inference/images', update=False, view_img=False, weights=['yolov5s.pt']) Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device1 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 7.45958e+06 gradients image 1/2 yolov5/inference/images/bus.jpg: 640x512 4 persons, 1 buss, Done. (0.013s) image 2/2 yolov5/inference/images/zidane.jpg: 384x640 2 persons, 1 ties, Done. (0.013s) Results saved to yolov5/inference/output Done. (0.130s) Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.4, device='0,1', img_size=640, iou_thres=0.5, output='inference/output', save_txt=False, source='inference/images', update=False, view_img=False, weights=['runs/exp0/weights/last.pt']) Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device1 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 6.61683e+06 gradients image 1/2 yolov5/inference/images/bus.jpg: 640x512 5 persons, 2 traffic lights, 1 umbrellas, Done. (0.013s) image 2/2 yolov5/inference/images/zidane.jpg: 384x640 2 persons, 1 ties, Done. (0.013s) Results saved to yolov5/inference/output Done. (0.134s) Namespace(augment=False, batch_size=32, conf_thres=0.001, data='data/coco128.yaml', device='0,1', img_size=640, iou_thres=0.65, merge=False, save_json=False, save_txt=False, single_cls=False, task='val', verbose=False, weights=['yolov5s.pt']) Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device1 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 7.45958e+06 gradients Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Class Images Targets P R mAP@.5 all 128 929 0.381 0.745 0.691 0.451 Speed: 1.4/1.6/2.9 ms inference/NMS/total per 640x640 image at batch-size 32 Namespace(augment=False, batch_size=32, conf_thres=0.001, data='data/coco128.yaml', device='0,1', img_size=640, iou_thres=0.65, merge=False, save_json=False, save_txt=False, single_cls=False, task='val', verbose=False, weights=['runs/exp0/weights/last.pt']) Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device1 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 6.61683e+06 gradients Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Class Images Targets P R mAP@.5 all 128 929 0.139 0.726 0.457 0.221 Speed: 1.5/2.0/3.5 ms inference/NMS/total per 640x640 image at batch-size 32 Using CUDA Apex device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Namespace(batch_size=16, bucket='', cache_images=False, cfg='models/yolov5s.yaml', data='data/coco128.yaml', device='0', epochs=3, evolve=False, hyp='', img_size=[320, 320], local_rank=-1, multi_scale=False, name='', noautoanchor=False, nosave=False, notest=False, rect=False, resume=False, single_cls=False, sync_bn=False, total_batch_size=16, weights='yolov5s.pt', world_size=1) Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/ Hyperparameters {'optimizer': 'SGD', 'lr0': 0.01, 'momentum': 0.937, 'weight_decay': 0.0005, 'giou': 0.05, 'cls': 0.58, 'cls_pw': 1.0, 'obj': 1.0, 'obj_pw': 1.0, 'iou_t': 0.2, 'anchor_t': 4.0, 'fl_gamma': 0.0, 'hsv_h': 0.014, 'hsv_s': 0.68, 'hsv_v': 0.36, 'degrees': 0.0, 'translate': 0.0, 'scale': 0.5, 'shear': 0.0} from n params module arguments 0 -1 1 3520 models.common.Focus [3, 32, 3] 1 -1 1 18560 models.common.Conv [32, 64, 3, 2] 2 -1 1 19904 models.common.BottleneckCSP [64, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 1 161152 models.common.BottleneckCSP [128, 128, 3] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 1 641792 models.common.BottleneckCSP [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]] 9 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 10 -1 1 131584 models.common.Conv [512, 256, 1, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 models.common.Concat [1] 13 -1 1 378624 models.common.BottleneckCSP [512, 256, 1, False] 14 -1 1 33024 models.common.Conv [256, 128, 1, 1] 15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 16 [-1, 4] 1 0 models.common.Concat [1] 17 -1 1 95104 models.common.BottleneckCSP [256, 128, 1, False] 18 -1 1 32895 torch.nn.modules.conv.Conv2d [128, 255, 1, 1] 19 -2 1 147712 models.common.Conv [128, 128, 3, 2] 20 [-1, 14] 1 0 models.common.Concat [1] 21 -1 1 313088 models.common.BottleneckCSP [256, 256, 1, False] 22 -1 1 65535 torch.nn.modules.conv.Conv2d [256, 255, 1, 1] 23 -2 1 590336 models.common.Conv [256, 256, 3, 2] 24 [-1, 10] 1 0 models.common.Concat [1] 25 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 26 -1 1 130815 torch.nn.modules.conv.Conv2d [512, 255, 1, 1] 27 [-1, 22, 18] 1 0 models.yolo.Detect [80, [[116, 90, 156, 198, 373, 326], [30, 61, 62, 45, 59, 119], [10, 13, 16, 30, 33, 23]]] Model Summary: 191 layers, 7.46816e+06 parameters, 7.46816e+06 gradients Optimizer groups: 62 .bias, 70 conv.weight, 59 other Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Analyzing anchors... Best Possible Recall (BPR) = 0.9591. Attempting to generate improved anchors, please wait... WARNING: Extremely small objects found. 35 of 929 labels are < 3 pixels in width or height. Running kmeans for 9 anchors on 927 points... thr=0.25: 0.9752 best possible recall, 3.75 anchors past thr n=9, img_size=320, metric_all=0.261/0.655-mean/best, past_thr=0.473-mean: 8,11, 26,19, 30,47, 75,43, 52,92, 77,162, 160,107, 174,237, 299,195 Evolving anchors with Genetic Algorithm: fitness = 0.6584: 100%|█| 1000/1000 [00 thr=0.25: 0.9828 best possible recall, 3.84 anchors past thr n=9, img_size=320, metric_all=0.266/0.660-mean/best, past_thr=0.474-mean: 9,10, 25,15, 30,41, 64,43, 50,86, 69,147, 144,108, 183,196, 314,210 New anchors saved to model. Update model *.yaml to use these anchors in the future. Image sizes 320 train, 320 test Using 8 dataloader workers Starting training for 3 epochs... Epoch gpu_mem GIoU obj cls total targets img_size 0/2 1.74G 0.128 0.1096 0.05107 0.2887 170 320 Class Images Targets P R mAP@.5 all 128 929 0.144 0.636 0.364 0.137 Epoch gpu_mem GIoU obj cls total targets img_size 1/2 1.72G 0.118 0.08912 0.04238 0.2495 193 320 Class Images Targets P R mAP@.5 all 128 929 0.154 0.627 0.444 0.224 Epoch gpu_mem GIoU obj cls total targets img_size 2/2 1.72G 0.1089 0.09007 0.04059 0.2395 200 320 Class Images Targets P R mAP@.5 all 128 929 0.149 0.646 0.467 0.24 Optimizer stripped from runs/exp2/weights/last.pt, 15.1MB Optimizer stripped from runs/exp2/weights/best.pt, 15.1MB No handles with labels found to put in legend. 3 epochs completed in 0.007 hours. Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.4, device='0', img_size=640, iou_thres=0.5, output='inference/output', save_txt=False, source='inference/images', update=False, view_img=False, weights=['yolov5s.pt']) Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 7.45958e+06 gradients image 1/2 yolov5/inference/images/bus.jpg: 640x512 4 persons, 1 buss, Done. (0.013s) image 2/2 yolov5/inference/images/zidane.jpg: 384x640 2 persons, 1 ties, Done. (0.013s) Results saved to yolov5/inference/output Done. (0.133s) Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.4, device='0', img_size=640, iou_thres=0.5, output='inference/output', save_txt=False, source='inference/images', update=False, view_img=False, weights=['runs/exp0/weights/last.pt']) Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 6.61683e+06 gradients image 1/2 yolov5/inference/images/bus.jpg: 640x512 5 persons, 2 traffic lights, 1 umbrellas, Done. (0.013s) image 2/2 yolov5/inference/images/zidane.jpg: 384x640 2 persons, 1 ties, Done. (0.012s) Results saved to yolov5/inference/output Done. (0.132s) Namespace(augment=False, batch_size=32, conf_thres=0.001, data='data/coco128.yaml', device='0', img_size=640, iou_thres=0.65, merge=False, save_json=False, save_txt=False, single_cls=False, task='val', verbose=False, weights=['yolov5s.pt']) Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 7.45958e+06 gradients Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Class Images Targets P R mAP@.5 all 128 929 0.381 0.745 0.691 0.451 Speed: 1.4/1.7/3.0 ms inference/NMS/total per 640x640 image at batch-size 32 Namespace(augment=False, batch_size=32, conf_thres=0.001, data='data/coco128.yaml', device='0', img_size=640, iou_thres=0.65, merge=False, save_json=False, save_txt=False, single_cls=False, task='val', verbose=False, weights=['runs/exp0/weights/last.pt']) Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 6.61683e+06 gradients Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Class Images Targets P R mAP@.5 all 128 929 0.139 0.726 0.457 0.221 Speed: 1.3/2.0/3.3 ms inference/NMS/total per 640x640 image at batch-size 32 Using CPU Namespace(batch_size=16, bucket='', cache_images=False, cfg='models/yolov5s.yaml', data='data/coco128.yaml', device='cpu', epochs=3, evolve=False, hyp='', img_size=[320, 320], local_rank=-1, multi_scale=False, name='', noautoanchor=False, nosave=False, notest=False, rect=False, resume=False, single_cls=False, sync_bn=False, total_batch_size=16, weights='yolov5s.pt', world_size=1) Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/ Hyperparameters {'optimizer': 'SGD', 'lr0': 0.01, 'momentum': 0.937, 'weight_decay': 0.0005, 'giou': 0.05, 'cls': 0.58, 'cls_pw': 1.0, 'obj': 1.0, 'obj_pw': 1.0, 'iou_t': 0.2, 'anchor_t': 4.0, 'fl_gamma': 0.0, 'hsv_h': 0.014, 'hsv_s': 0.68, 'hsv_v': 0.36, 'degrees': 0.0, 'translate': 0.0, 'scale': 0.5, 'shear': 0.0} from n params module arguments 0 -1 1 3520 models.common.Focus [3, 32, 3] 1 -1 1 18560 models.common.Conv [32, 64, 3, 2] 2 -1 1 19904 models.common.BottleneckCSP [64, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 1 161152 models.common.BottleneckCSP [128, 128, 3] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 1 641792 models.common.BottleneckCSP [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]] 9 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 10 -1 1 131584 models.common.Conv [512, 256, 1, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 models.common.Concat [1] 13 -1 1 378624 models.common.BottleneckCSP [512, 256, 1, False] 14 -1 1 33024 models.common.Conv [256, 128, 1, 1] 15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 16 [-1, 4] 1 0 models.common.Concat [1] 17 -1 1 95104 models.common.BottleneckCSP [256, 128, 1, False] 18 -1 1 32895 torch.nn.modules.conv.Conv2d [128, 255, 1, 1] 19 -2 1 147712 models.common.Conv [128, 128, 3, 2] 20 [-1, 14] 1 0 models.common.Concat [1] 21 -1 1 313088 models.common.BottleneckCSP [256, 256, 1, False] 22 -1 1 65535 torch.nn.modules.conv.Conv2d [256, 255, 1, 1] 23 -2 1 590336 models.common.Conv [256, 256, 3, 2] 24 [-1, 10] 1 0 models.common.Concat [1] 25 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 26 -1 1 130815 torch.nn.modules.conv.Conv2d [512, 255, 1, 1] 27 [-1, 22, 18] 1 0 models.yolo.Detect [80, [[116, 90, 156, 198, 373, 326], [30, 61, 62, 45, 59, 119], [10, 13, 16, 30, 33, 23]]] Model Summary: 191 layers, 7.46816e+06 parameters, 7.46816e+06 gradients Optimizer groups: 62 .bias, 70 conv.weight, 59 other Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Analyzing anchors... Best Possible Recall (BPR) = 0.9591. Attempting to generate improved anchors, please wait... WARNING: Extremely small objects found. 35 of 929 labels are < 3 pixels in width or height. Running kmeans for 9 anchors on 927 points... thr=0.25: 0.9752 best possible recall, 3.75 anchors past thr n=9, img_size=320, metric_all=0.261/0.655-mean/best, past_thr=0.473-mean: 8,11, 26,19, 30,47, 75,43, 52,92, 77,162, 160,107, 174,237, 299,195 Evolving anchors with Genetic Algorithm: fitness = 0.6584: 100%|█| 1000/1000 [00 thr=0.25: 0.9828 best possible recall, 3.84 anchors past thr n=9, img_size=320, metric_all=0.266/0.660-mean/best, past_thr=0.474-mean: 9,10, 25,15, 30,41, 64,43, 50,86, 69,147, 144,108, 183,196, 314,210 New anchors saved to model. Update model *.yaml to use these anchors in the future. Image sizes 320 train, 320 test Using 8 dataloader workers Starting training for 3 epochs... Epoch gpu_mem GIoU obj cls total targets img_size 0/2 0G 0.128 0.1096 0.05105 0.2886 170 320 Class Images Targets P R mAP@.5 all 128 929 0.143 0.635 0.364 0.142 Epoch gpu_mem GIoU obj cls total targets img_size 1/2 0G 0.118 0.08917 0.04235 0.2495 193 320 Class Images Targets P R mAP@.5 all 128 929 0.154 0.628 0.446 0.226 Epoch gpu_mem GIoU obj cls total targets img_size 2/2 0G 0.1089 0.09018 0.04058 0.2396 200 320 Class Images Targets P R mAP@.5 all 128 929 0.149 0.645 0.467 0.239 Optimizer stripped from runs/exp3/weights/last.pt, 15.1MB Optimizer stripped from runs/exp3/weights/best.pt, 15.1MB No handles with labels found to put in legend. 3 epochs completed in 0.032 hours. Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.4, device='cpu', img_size=640, iou_thres=0.5, output='inference/output', save_txt=False, source='inference/images', update=False, view_img=False, weights=['yolov5s.pt']) Using CPU Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 7.45958e+06 gradients image 1/2 yolov5/inference/images/bus.jpg: 640x512 4 persons, 1 buss, Done. (0.156s) image 2/2 yolov5/inference/images/zidane.jpg: 384x640 2 persons, 1 ties, Done. (0.097s) Results saved to yolov5/inference/output Done. (3.529s) Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.4, device='cpu', img_size=640, iou_thres=0.5, output='inference/output', save_txt=False, source='inference/images', update=False, view_img=False, weights=['runs/exp0/weights/last.pt']) Using CPU Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 6.61683e+06 gradients image 1/2 yolov5/inference/images/bus.jpg: 640x512 5 persons, 2 traffic lights, 1 umbrellas, Done. (0.121s) image 2/2 yolov5/inference/images/zidane.jpg: 384x640 2 persons, 1 ties, Done. (0.097s) Results saved to yolov5/inference/output Done. (3.478s) Namespace(augment=False, batch_size=32, conf_thres=0.001, data='data/coco128.yaml', device='cpu', img_size=640, iou_thres=0.65, merge=False, save_json=False, save_txt=False, single_cls=False, task='val', verbose=False, weights=['yolov5s.pt']) Using CPU Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 7.45958e+06 gradients Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Class Images Targets P R mAP@.5 all 128 929 0.382 0.747 0.692 0.45 Speed: 68.1/6.1/74.2 ms inference/NMS/total per 640x640 image at batch-size 32 Namespace(augment=False, batch_size=32, conf_thres=0.001, data='data/coco128.yaml', device='cpu', img_size=640, iou_thres=0.65, merge=False, save_json=False, save_txt=False, single_cls=False, task='val', verbose=False, weights=['runs/exp0/weights/last.pt']) Using CPU Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 6.61683e+06 gradients Scanning labels ../coco128/labels/train2017.cache (126 found, 0 missing, 2 empty Class Images Targets P R mAP@.5 all 128 929 0.139 0.728 0.457 0.222 Speed: 70.6/18.0/88.6 ms inference/NMS/total per 640x640 image at batch-size 32 Using CUDA device0 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device1 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device2 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device3 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device4 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device5 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device6 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) device7 _CudaDeviceProperties(name='Tesla V100-SXM2-32GB', total_memory=32480MB) from n params module arguments 0 -1 1 3520 models.common.Focus [3, 32, 3] 1 -1 1 18560 models.common.Conv [32, 64, 3, 2] 2 -1 1 19904 models.common.BottleneckCSP [64, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 1 161152 models.common.BottleneckCSP [128, 128, 3] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 1 641792 models.common.BottleneckCSP [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]] 9 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 10 -1 1 131584 models.common.Conv [512, 256, 1, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 models.common.Concat [1] 13 -1 1 378624 models.common.BottleneckCSP [512, 256, 1, False] 14 -1 1 33024 models.common.Conv [256, 128, 1, 1] 15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 16 [-1, 4] 1 0 models.common.Concat [1] 17 -1 1 95104 models.common.BottleneckCSP [256, 128, 1, False] 18 -1 1 32895 torch.nn.modules.conv.Conv2d [128, 255, 1, 1] 19 -2 1 147712 models.common.Conv [128, 128, 3, 2] 20 [-1, 14] 1 0 models.common.Concat [1] 21 -1 1 313088 models.common.BottleneckCSP [256, 256, 1, False] 22 -1 1 65535 torch.nn.modules.conv.Conv2d [256, 255, 1, 1] 23 -2 1 590336 models.common.Conv [256, 256, 3, 2] 24 [-1, 10] 1 0 models.common.Concat [1] 25 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False] 26 -1 1 130815 torch.nn.modules.conv.Conv2d [512, 255, 1, 1] 27 [-1, 22, 18] 1 0 Detect [80, [[116, 90, 156, 198, 373, 326], [30, 61, 62, 45, 59, 119], [10, 13, 16, 30, 33, 23]]] Model Summary: 191 layers, 7.46816e+06 parameters, 7.46816e+06 gradients Namespace(batch_size=1, img_size=[640, 640], weights='yolov5s.pt') Starting TorchScript export with torch 1.5.1... TorchScript export success, saved as yolov5s.torchscript.pt Starting ONNX export with onnx 1.7.0... Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 7.45958e+06 gradients graph torch-jit-export ( %images[FLOAT, 1x3x640x640] ) initializers ( %model.0.conv.conv.bias[FLOAT, 32] %model.0.conv.conv.weight[FLOAT, 32x12x3x3] %model.1.conv.bias[FLOAT, 64] %model.1.conv.weight[FLOAT, 64x32x3x3] %model.10.conv.bias[FLOAT, 256] %model.10.conv.weight[FLOAT, 256x512x1x1] %model.13.bn.bias[FLOAT, 256] %model.13.bn.num_batches_tracked[INT64, scalar] %model.13.bn.running_mean[FLOAT, 256] %model.13.bn.running_var[FLOAT, 256] %model.13.bn.weight[FLOAT, 256] %model.13.cv1.conv.bias[FLOAT, 128] %model.13.cv1.conv.weight[FLOAT, 128x512x1x1] %model.13.cv2.weight[FLOAT, 128x512x1x1] %model.13.cv3.weight[FLOAT, 128x128x1x1] %model.13.cv4.conv.bias[FLOAT, 256] %model.13.cv4.conv.weight[FLOAT, 256x256x1x1] %model.13.m.0.cv1.conv.bias[FLOAT, 128] %model.13.m.0.cv1.conv.weight[FLOAT, 128x128x1x1] %model.13.m.0.cv2.conv.bias[FLOAT, 128] %model.13.m.0.cv2.conv.weight[FLOAT, 128x128x3x3] %model.14.conv.bias[FLOAT, 128] %model.14.conv.weight[FLOAT, 128x256x1x1] %model.17.bn.bias[FLOAT, 128] %model.17.bn.num_batches_tracked[INT64, scalar] %model.17.bn.running_mean[FLOAT, 128] %model.17.bn.running_var[FLOAT, 128] %model.17.bn.weight[FLOAT, 128] %model.17.cv1.conv.bias[FLOAT, 64] %model.17.cv1.conv.weight[FLOAT, 64x256x1x1] %model.17.cv2.weight[FLOAT, 64x256x1x1] %model.17.cv3.weight[FLOAT, 64x64x1x1] %model.17.cv4.conv.bias[FLOAT, 128] %model.17.cv4.conv.weight[FLOAT, 128x128x1x1] %model.17.m.0.cv1.conv.bias[FLOAT, 64] %model.17.m.0.cv1.conv.weight[FLOAT, 64x64x1x1] %model.17.m.0.cv2.conv.bias[FLOAT, 64] %model.17.m.0.cv2.conv.weight[FLOAT, 64x64x3x3] %model.18.bias[FLOAT, 255] %model.18.weight[FLOAT, 255x128x1x1] %model.19.conv.bias[FLOAT, 128] %model.19.conv.weight[FLOAT, 128x128x3x3] %model.2.bn.bias[FLOAT, 64] %model.2.bn.num_batches_tracked[INT64, scalar] %model.2.bn.running_mean[FLOAT, 64] %model.2.bn.running_var[FLOAT, 64] %model.2.bn.weight[FLOAT, 64] %model.2.cv1.conv.bias[FLOAT, 32] %model.2.cv1.conv.weight[FLOAT, 32x64x1x1] %model.2.cv2.weight[FLOAT, 32x64x1x1] %model.2.cv3.weight[FLOAT, 32x32x1x1] %model.2.cv4.conv.bias[FLOAT, 64] %model.2.cv4.conv.weight[FLOAT, 64x64x1x1] %model.2.m.0.cv1.conv.bias[FLOAT, 32] %model.2.m.0.cv1.conv.weight[FLOAT, 32x32x1x1] %model.2.m.0.cv2.conv.bias[FLOAT, 32] %model.2.m.0.cv2.conv.weight[FLOAT, 32x32x3x3] %model.21.bn.bias[FLOAT, 256] %model.21.bn.num_batches_tracked[INT64, scalar] %model.21.bn.running_mean[FLOAT, 256] %model.21.bn.running_var[FLOAT, 256] %model.21.bn.weight[FLOAT, 256] %model.21.cv1.conv.bias[FLOAT, 128] %model.21.cv1.conv.weight[FLOAT, 128x256x1x1] %model.21.cv2.weight[FLOAT, 128x256x1x1] %model.21.cv3.weight[FLOAT, 128x128x1x1] %model.21.cv4.conv.bias[FLOAT, 256] %model.21.cv4.conv.weight[FLOAT, 256x256x1x1] %model.21.m.0.cv1.conv.bias[FLOAT, 128] %model.21.m.0.cv1.conv.weight[FLOAT, 128x128x1x1] %model.21.m.0.cv2.conv.bias[FLOAT, 128] %model.21.m.0.cv2.conv.weight[FLOAT, 128x128x3x3] %model.22.bias[FLOAT, 255] %model.22.weight[FLOAT, 255x256x1x1] %model.23.conv.bias[FLOAT, 256] %model.23.conv.weight[FLOAT, 256x256x3x3] %model.25.bn.bias[FLOAT, 512] %model.25.bn.num_batches_tracked[INT64, scalar] %model.25.bn.running_mean[FLOAT, 512] %model.25.bn.running_var[FLOAT, 512] %model.25.bn.weight[FLOAT, 512] %model.25.cv1.conv.bias[FLOAT, 256] %model.25.cv1.conv.weight[FLOAT, 256x512x1x1] %model.25.cv2.weight[FLOAT, 256x512x1x1] %model.25.cv3.weight[FLOAT, 256x256x1x1] %model.25.cv4.conv.bias[FLOAT, 512] %model.25.cv4.conv.weight[FLOAT, 512x512x1x1] %model.25.m.0.cv1.conv.bias[FLOAT, 256] %model.25.m.0.cv1.conv.weight[FLOAT, 256x256x1x1] %model.25.m.0.cv2.conv.bias[FLOAT, 256] %model.25.m.0.cv2.conv.weight[FLOAT, 256x256x3x3] %model.26.bias[FLOAT, 255] %model.26.weight[FLOAT, 255x512x1x1] %model.27.anchor_grid[FLOAT, 3x1x3x1x1x2] %model.27.anchors[FLOAT, 3x3x2] %model.3.conv.bias[FLOAT, 128] %model.3.conv.weight[FLOAT, 128x64x3x3] %model.4.bn.bias[FLOAT, 128] %model.4.bn.num_batches_tracked[INT64, scalar] %model.4.bn.running_mean[FLOAT, 128] %model.4.bn.running_var[FLOAT, 128] %model.4.bn.weight[FLOAT, 128] %model.4.cv1.conv.bias[FLOAT, 64] %model.4.cv1.conv.weight[FLOAT, 64x128x1x1] %model.4.cv2.weight[FLOAT, 64x128x1x1] %model.4.cv3.weight[FLOAT, 64x64x1x1] %model.4.cv4.conv.bias[FLOAT, 128] %model.4.cv4.conv.weight[FLOAT, 128x128x1x1] %model.4.m.0.cv1.conv.bias[FLOAT, 64] %model.4.m.0.cv1.conv.weight[FLOAT, 64x64x1x1] %model.4.m.0.cv2.conv.bias[FLOAT, 64] %model.4.m.0.cv2.conv.weight[FLOAT, 64x64x3x3] %model.4.m.1.cv1.conv.bias[FLOAT, 64] %model.4.m.1.cv1.conv.weight[FLOAT, 64x64x1x1] %model.4.m.1.cv2.conv.bias[FLOAT, 64] %model.4.m.1.cv2.conv.weight[FLOAT, 64x64x3x3] %model.4.m.2.cv1.conv.bias[FLOAT, 64] %model.4.m.2.cv1.conv.weight[FLOAT, 64x64x1x1] %model.4.m.2.cv2.conv.bias[FLOAT, 64] %model.4.m.2.cv2.conv.weight[FLOAT, 64x64x3x3] %model.5.conv.bias[FLOAT, 256] %model.5.conv.weight[FLOAT, 256x128x3x3] %model.6.bn.bias[FLOAT, 256] %model.6.bn.num_batches_tracked[INT64, scalar] %model.6.bn.running_mean[FLOAT, 256] %model.6.bn.running_var[FLOAT, 256] %model.6.bn.weight[FLOAT, 256] %model.6.cv1.conv.bias[FLOAT, 128] %model.6.cv1.conv.weight[FLOAT, 128x256x1x1] %model.6.cv2.weight[FLOAT, 128x256x1x1] %model.6.cv3.weight[FLOAT, 128x128x1x1] %model.6.cv4.conv.bias[FLOAT, 256] %model.6.cv4.conv.weight[FLOAT, 256x256x1x1] %model.6.m.0.cv1.conv.bias[FLOAT, 128] %model.6.m.0.cv1.conv.weight[FLOAT, 128x128x1x1] %model.6.m.0.cv2.conv.bias[FLOAT, 128] %model.6.m.0.cv2.conv.weight[FLOAT, 128x128x3x3] %model.6.m.1.cv1.conv.bias[FLOAT, 128] %model.6.m.1.cv1.conv.weight[FLOAT, 128x128x1x1] %model.6.m.1.cv2.conv.bias[FLOAT, 128] %model.6.m.1.cv2.conv.weight[FLOAT, 128x128x3x3] %model.6.m.2.cv1.conv.bias[FLOAT, 128] %model.6.m.2.cv1.conv.weight[FLOAT, 128x128x1x1] %model.6.m.2.cv2.conv.bias[FLOAT, 128] %model.6.m.2.cv2.conv.weight[FLOAT, 128x128x3x3] %model.7.conv.bias[FLOAT, 512] %model.7.conv.weight[FLOAT, 512x256x3x3] %model.8.cv1.conv.bias[FLOAT, 256] %model.8.cv1.conv.weight[FLOAT, 256x512x1x1] %model.8.cv2.conv.bias[FLOAT, 512] %model.8.cv2.conv.weight[FLOAT, 512x1024x1x1] %model.9.bn.bias[FLOAT, 512] %model.9.bn.num_batches_tracked[INT64, scalar] %model.9.bn.running_mean[FLOAT, 512] %model.9.bn.running_var[FLOAT, 512] %model.9.bn.weight[FLOAT, 512] %model.9.cv1.conv.bias[FLOAT, 256] %model.9.cv1.conv.weight[FLOAT, 256x512x1x1] %model.9.cv2.weight[FLOAT, 256x512x1x1] %model.9.cv3.weight[FLOAT, 256x256x1x1] %model.9.cv4.conv.bias[FLOAT, 512] %model.9.cv4.conv.weight[FLOAT, 512x512x1x1] %model.9.m.0.cv1.conv.bias[FLOAT, 256] %model.9.m.0.cv1.conv.weight[FLOAT, 256x256x1x1] %model.9.m.0.cv2.conv.bias[FLOAT, 256] %model.9.m.0.cv2.conv.weight[FLOAT, 256x256x3x3] ) { %167 = Constant[value = ]() %168 = Constant[value = ]() %169 = Constant[value = ]() %170 = Constant[value = ]() %171 = Slice(%images, %168, %169, %167, %170) %172 = Constant[value = ]() %173 = Constant[value = ]() %174 = Constant[value = ]() %175 = Constant[value = ]() %176 = Slice(%171, %173, %174, %172, %175) %177 = Constant[value = ]() %178 = Constant[value = ]() %179 = Constant[value = ]() %180 = Constant[value = ]() %181 = Slice(%images, %178, %179, %177, %180) %182 = Constant[value = ]() %183 = Constant[value = ]() %184 = Constant[value = ]() %185 = Constant[value = ]() %186 = Slice(%181, %183, %184, %182, %185) %187 = Constant[value = ]() %188 = Constant[value = ]() %189 = Constant[value = ]() %190 = Constant[value = ]() %191 = Slice(%images, %188, %189, %187, %190) %192 = Constant[value = ]() %193 = Constant[value = ]() %194 = Constant[value = ]() %195 = Constant[value = ]() %196 = Slice(%191, %193, %194, %192, %195) %197 = Constant[value = ]() %198 = Constant[value = ]() %199 = Constant[value = ]() %200 = Constant[value = ]() %201 = Slice(%images, %198, %199, %197, %200) %202 = Constant[value = ]() %203 = Constant[value = ]() %204 = Constant[value = ]() %205 = Constant[value = ]() %206 = Slice(%201, %203, %204, %202, %205) %207 = Concat[axis = 1](%176, %186, %196, %206) %208 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%207, %model.0.conv.conv.weight, %model.0.conv.conv.bias) %209 = LeakyRelu[alpha = 0.100000001490116](%208) %210 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%209, %model.1.conv.weight, %model.1.conv.bias) %211 = LeakyRelu[alpha = 0.100000001490116](%210) %212 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%211, %model.2.cv1.conv.weight, %model.2.cv1.conv.bias) %213 = LeakyRelu[alpha = 0.100000001490116](%212) %214 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%213, %model.2.m.0.cv1.conv.weight, %model.2.m.0.cv1.conv.bias) %215 = LeakyRelu[alpha = 0.100000001490116](%214) %216 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%215, %model.2.m.0.cv2.conv.weight, %model.2.m.0.cv2.conv.bias) %217 = LeakyRelu[alpha = 0.100000001490116](%216) %218 = Add(%213, %217) %219 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%218, %model.2.cv3.weight) %220 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%211, %model.2.cv2.weight) %221 = Concat[axis = 1](%219, %220) %222 = BatchNormalization[epsilon = 0.00100000004749745, momentum = 0.990000009536743](%221, %model.2.bn.weight, %model.2.bn.bias, %model.2.bn.running_mean, %model.2.bn.running_var) %223 = LeakyRelu[alpha = 0.100000001490116](%222) %224 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%223, %model.2.cv4.conv.weight, %model.2.cv4.conv.bias) %225 = LeakyRelu[alpha = 0.100000001490116](%224) %226 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%225, %model.3.conv.weight, %model.3.conv.bias) %227 = LeakyRelu[alpha = 0.100000001490116](%226) %228 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%227, %model.4.cv1.conv.weight, %model.4.cv1.conv.bias) %229 = LeakyRelu[alpha = 0.100000001490116](%228) %230 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%229, %model.4.m.0.cv1.conv.weight, %model.4.m.0.cv1.conv.bias) %231 = LeakyRelu[alpha = 0.100000001490116](%230) %232 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%231, %model.4.m.0.cv2.conv.weight, %model.4.m.0.cv2.conv.bias) %233 = LeakyRelu[alpha = 0.100000001490116](%232) %234 = Add(%229, %233) %235 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%234, %model.4.m.1.cv1.conv.weight, %model.4.m.1.cv1.conv.bias) %236 = LeakyRelu[alpha = 0.100000001490116](%235) %237 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%236, %model.4.m.1.cv2.conv.weight, %model.4.m.1.cv2.conv.bias) %238 = LeakyRelu[alpha = 0.100000001490116](%237) %239 = Add(%234, %238) %240 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%239, %model.4.m.2.cv1.conv.weight, %model.4.m.2.cv1.conv.bias) %241 = LeakyRelu[alpha = 0.100000001490116](%240) %242 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%241, %model.4.m.2.cv2.conv.weight, %model.4.m.2.cv2.conv.bias) %243 = LeakyRelu[alpha = 0.100000001490116](%242) %244 = Add(%239, %243) %245 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%244, %model.4.cv3.weight) %246 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%227, %model.4.cv2.weight) %247 = Concat[axis = 1](%245, %246) %248 = BatchNormalization[epsilon = 0.00100000004749745, momentum = 0.990000009536743](%247, %model.4.bn.weight, %model.4.bn.bias, %model.4.bn.running_mean, %model.4.bn.running_var) %249 = LeakyRelu[alpha = 0.100000001490116](%248) %250 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%249, %model.4.cv4.conv.weight, %model.4.cv4.conv.bias) %251 = LeakyRelu[alpha = 0.100000001490116](%250) %252 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%251, %model.5.conv.weight, %model.5.conv.bias) %253 = LeakyRelu[alpha = 0.100000001490116](%252) %254 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%253, %model.6.cv1.conv.weight, %model.6.cv1.conv.bias) %255 = LeakyRelu[alpha = 0.100000001490116](%254) %256 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%255, %model.6.m.0.cv1.conv.weight, %model.6.m.0.cv1.conv.bias) %257 = LeakyRelu[alpha = 0.100000001490116](%256) %258 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%257, %model.6.m.0.cv2.conv.weight, %model.6.m.0.cv2.conv.bias) %259 = LeakyRelu[alpha = 0.100000001490116](%258) %260 = Add(%255, %259) %261 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%260, %model.6.m.1.cv1.conv.weight, %model.6.m.1.cv1.conv.bias) %262 = LeakyRelu[alpha = 0.100000001490116](%261) %263 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%262, %model.6.m.1.cv2.conv.weight, %model.6.m.1.cv2.conv.bias) %264 = LeakyRelu[alpha = 0.100000001490116](%263) %265 = Add(%260, %264) %266 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%265, %model.6.m.2.cv1.conv.weight, %model.6.m.2.cv1.conv.bias) %267 = LeakyRelu[alpha = 0.100000001490116](%266) %268 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%267, %model.6.m.2.cv2.conv.weight, %model.6.m.2.cv2.conv.bias) %269 = LeakyRelu[alpha = 0.100000001490116](%268) %270 = Add(%265, %269) %271 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%270, %model.6.cv3.weight) %272 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%253, %model.6.cv2.weight) %273 = Concat[axis = 1](%271, %272) %274 = BatchNormalization[epsilon = 0.00100000004749745, momentum = 0.990000009536743](%273, %model.6.bn.weight, %model.6.bn.bias, %model.6.bn.running_mean, %model.6.bn.running_var) %275 = LeakyRelu[alpha = 0.100000001490116](%274) %276 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%275, %model.6.cv4.conv.weight, %model.6.cv4.conv.bias) %277 = LeakyRelu[alpha = 0.100000001490116](%276) %278 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%277, %model.7.conv.weight, %model.7.conv.bias) %279 = LeakyRelu[alpha = 0.100000001490116](%278) %280 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%279, %model.8.cv1.conv.weight, %model.8.cv1.conv.bias) %281 = LeakyRelu[alpha = 0.100000001490116](%280) %282 = MaxPool[ceil_mode = 0, kernel_shape = [5, 5], pads = [2, 2, 2, 2], strides = [1, 1]](%281) %283 = MaxPool[ceil_mode = 0, kernel_shape = [9, 9], pads = [4, 4, 4, 4], strides = [1, 1]](%281) %284 = MaxPool[ceil_mode = 0, kernel_shape = [13, 13], pads = [6, 6, 6, 6], strides = [1, 1]](%281) %285 = Concat[axis = 1](%281, %282, %283, %284) %286 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%285, %model.8.cv2.conv.weight, %model.8.cv2.conv.bias) %287 = LeakyRelu[alpha = 0.100000001490116](%286) %288 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%287, %model.9.cv1.conv.weight, %model.9.cv1.conv.bias) %289 = LeakyRelu[alpha = 0.100000001490116](%288) %290 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%289, %model.9.m.0.cv1.conv.weight, %model.9.m.0.cv1.conv.bias) %291 = LeakyRelu[alpha = 0.100000001490116](%290) %292 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%291, %model.9.m.0.cv2.conv.weight, %model.9.m.0.cv2.conv.bias) %293 = LeakyRelu[alpha = 0.100000001490116](%292) %294 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%293, %model.9.cv3.weight) %295 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%287, %model.9.cv2.weight) %296 = Concat[axis = 1](%294, %295) %297 = BatchNormalization[epsilon = 0.00100000004749745, momentum = 0.990000009536743](%296, %model.9.bn.weight, %model.9.bn.bias, %model.9.bn.running_mean, %model.9.bn.running_var) %298 = LeakyRelu[alpha = 0.100000001490116](%297) %299 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%298, %model.9.cv4.conv.weight, %model.9.cv4.conv.bias) %300 = LeakyRelu[alpha = 0.100000001490116](%299) %301 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%300, %model.10.conv.weight, %model.10.conv.bias) %302 = LeakyRelu[alpha = 0.100000001490116](%301) %303 = Shape(%302) %304 = Constant[value = ]() %305 = Gather[axis = 0](%303, %304) %306 = Cast[to = 1](%305) %307 = Constant[value = ]() %308 = Mul(%306, %307) %309 = Cast[to = 1](%308) %310 = Floor(%309) %311 = Shape(%302) %312 = Constant[value = ]() %313 = Gather[axis = 0](%311, %312) %314 = Cast[to = 1](%313) %315 = Constant[value = ]() %316 = Mul(%314, %315) %317 = Cast[to = 1](%316) %318 = Floor(%317) %319 = Unsqueeze[axes = [0]](%310) %320 = Unsqueeze[axes = [0]](%318) %321 = Concat[axis = 0](%319, %320) %322 = Constant[value = ]() %323 = Shape(%302) %324 = Constant[value = ]() %325 = Constant[value = ]() %326 = Constant[value = ]() %327 = Slice(%323, %325, %326, %324) %328 = Cast[to = 7](%321) %329 = Concat[axis = 0](%327, %328) %330 = Constant[value = ]() %331 = Resize[coordinate_transformation_mode = 'asymmetric', cubic_coeff_a = -0.75, mode = 'nearest', nearest_mode = 'floor'](%302, %322, %330, %329) %332 = Concat[axis = 1](%331, %277) %333 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%332, %model.13.cv1.conv.weight, %model.13.cv1.conv.bias) %334 = LeakyRelu[alpha = 0.100000001490116](%333) %335 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%334, %model.13.m.0.cv1.conv.weight, %model.13.m.0.cv1.conv.bias) %336 = LeakyRelu[alpha = 0.100000001490116](%335) %337 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%336, %model.13.m.0.cv2.conv.weight, %model.13.m.0.cv2.conv.bias) %338 = LeakyRelu[alpha = 0.100000001490116](%337) %339 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%338, %model.13.cv3.weight) %340 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%332, %model.13.cv2.weight) %341 = Concat[axis = 1](%339, %340) %342 = BatchNormalization[epsilon = 0.00100000004749745, momentum = 0.990000009536743](%341, %model.13.bn.weight, %model.13.bn.bias, %model.13.bn.running_mean, %model.13.bn.running_var) %343 = LeakyRelu[alpha = 0.100000001490116](%342) %344 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%343, %model.13.cv4.conv.weight, %model.13.cv4.conv.bias) %345 = LeakyRelu[alpha = 0.100000001490116](%344) %346 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%345, %model.14.conv.weight, %model.14.conv.bias) %347 = LeakyRelu[alpha = 0.100000001490116](%346) %348 = Shape(%347) %349 = Constant[value = ]() %350 = Gather[axis = 0](%348, %349) %351 = Cast[to = 1](%350) %352 = Constant[value = ]() %353 = Mul(%351, %352) %354 = Cast[to = 1](%353) %355 = Floor(%354) %356 = Shape(%347) %357 = Constant[value = ]() %358 = Gather[axis = 0](%356, %357) %359 = Cast[to = 1](%358) %360 = Constant[value = ]() %361 = Mul(%359, %360) %362 = Cast[to = 1](%361) %363 = Floor(%362) %364 = Unsqueeze[axes = [0]](%355) %365 = Unsqueeze[axes = [0]](%363) %366 = Concat[axis = 0](%364, %365) %367 = Constant[value = ]() %368 = Shape(%347) %369 = Constant[value = ]() %370 = Constant[value = ]() %371 = Constant[value = ]() %372 = Slice(%368, %370, %371, %369) %373 = Cast[to = 7](%366) %374 = Concat[axis = 0](%372, %373) %375 = Constant[value = ]() %376 = Resize[coordinate_transformation_mode = 'asymmetric', cubic_coeff_a = -0.75, mode = 'nearest', nearest_mode = 'floor'](%347, %367, %375, %374) %377 = Concat[axis = 1](%376, %251) %378 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%377, %model.17.cv1.conv.weight, %model.17.cv1.conv.bias) %379 = LeakyRelu[alpha = 0.100000001490116](%378) %380 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%379, %model.17.m.0.cv1.conv.weight, %model.17.m.0.cv1.conv.bias) %381 = LeakyRelu[alpha = 0.100000001490116](%380) %382 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%381, %model.17.m.0.cv2.conv.weight, %model.17.m.0.cv2.conv.bias) %383 = LeakyRelu[alpha = 0.100000001490116](%382) %384 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%383, %model.17.cv3.weight) %385 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%377, %model.17.cv2.weight) %386 = Concat[axis = 1](%384, %385) %387 = BatchNormalization[epsilon = 0.00100000004749745, momentum = 0.990000009536743](%386, %model.17.bn.weight, %model.17.bn.bias, %model.17.bn.running_mean, %model.17.bn.running_var) %388 = LeakyRelu[alpha = 0.100000001490116](%387) %389 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%388, %model.17.cv4.conv.weight, %model.17.cv4.conv.bias) %390 = LeakyRelu[alpha = 0.100000001490116](%389) %391 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%390, %model.18.weight, %model.18.bias) %392 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%390, %model.19.conv.weight, %model.19.conv.bias) %393 = LeakyRelu[alpha = 0.100000001490116](%392) %394 = Concat[axis = 1](%393, %347) %395 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%394, %model.21.cv1.conv.weight, %model.21.cv1.conv.bias) %396 = LeakyRelu[alpha = 0.100000001490116](%395) %397 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%396, %model.21.m.0.cv1.conv.weight, %model.21.m.0.cv1.conv.bias) %398 = LeakyRelu[alpha = 0.100000001490116](%397) %399 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%398, %model.21.m.0.cv2.conv.weight, %model.21.m.0.cv2.conv.bias) %400 = LeakyRelu[alpha = 0.100000001490116](%399) %401 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%400, %model.21.cv3.weight) %402 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%394, %model.21.cv2.weight) %403 = Concat[axis = 1](%401, %402) %404 = BatchNormalization[epsilon = 0.00100000004749745, momentum = 0.990000009536743](%403, %model.21.bn.weight, %model.21.bn.bias, %model.21.bn.running_mean, %model.21.bn.running_var) %405 = LeakyRelu[alpha = 0.100000001490116](%404) %406 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%405, %model.21.cv4.conv.weight, %model.21.cv4.conv.bias) %407 = LeakyRelu[alpha = 0.100000001490116](%406) %408 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%407, %model.22.weight, %model.22.bias) %409 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%407, %model.23.conv.weight, %model.23.conv.bias) %410 = LeakyRelu[alpha = 0.100000001490116](%409) %411 = Concat[axis = 1](%410, %302) %412 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%411, %model.25.cv1.conv.weight, %model.25.cv1.conv.bias) %413 = LeakyRelu[alpha = 0.100000001490116](%412) %414 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%413, %model.25.m.0.cv1.conv.weight, %model.25.m.0.cv1.conv.bias) %415 = LeakyRelu[alpha = 0.100000001490116](%414) %416 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%415, %model.25.m.0.cv2.conv.weight, %model.25.m.0.cv2.conv.bias) %417 = LeakyRelu[alpha = 0.100000001490116](%416) %418 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%417, %model.25.cv3.weight) %419 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%411, %model.25.cv2.weight) %420 = Concat[axis = 1](%418, %419) %421 = BatchNormalization[epsilon = 0.00100000004749745, momentum = 0.990000009536743](%420, %model.25.bn.weight, %model.25.bn.bias, %model.25.bn.running_mean, %model.25.bn.running_var) %422 = LeakyRelu[alpha = 0.100000001490116](%421) %423 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%422, %model.25.cv4.conv.weight, %model.25.cv4.conv.bias) %424 = LeakyRelu[alpha = 0.100000001490116](%423) %425 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%424, %model.26.weight, %model.26.bias) %426 = Shape(%425) %427 = Constant[value = ]() %428 = Gather[axis = 0](%426, %427) %429 = Shape(%425) %430 = Constant[value = ]() %431 = Gather[axis = 0](%429, %430) %432 = Shape(%425) %433 = Constant[value = ]() %434 = Gather[axis = 0](%432, %433) %435 = Constant[value = ]() %436 = Constant[value = ]() %437 = Unsqueeze[axes = [0]](%428) %438 = Unsqueeze[axes = [0]](%435) %439 = Unsqueeze[axes = [0]](%436) %440 = Unsqueeze[axes = [0]](%431) %441 = Unsqueeze[axes = [0]](%434) %442 = Concat[axis = 0](%437, %438, %439, %440, %441) %443 = Reshape(%425, %442) %output = Transpose[perm = [0, 1, 3, 4, 2]](%443) %445 = Shape(%408) %446 = Constant[value = ]() %447 = Gather[axis = 0](%445, %446) %448 = Shape(%408) %449 = Constant[value = ]() %450 = Gather[axis = 0](%448, %449) %451 = Shape(%408) %452 = Constant[value = ]() %453 = Gather[axis = 0](%451, %452) %454 = Constant[value = ]() %455 = Constant[value = ]() %456 = Unsqueeze[axes = [0]](%447) %457 = Unsqueeze[axes = [0]](%454) %458 = Unsqueeze[axes = [0]](%455) %459 = Unsqueeze[axes = [0]](%450) %460 = Unsqueeze[axes = [0]](%453) %461 = Concat[axis = 0](%456, %457, %458, %459, %460) %462 = Reshape(%408, %461) %463 = Transpose[perm = [0, 1, 3, 4, 2]](%462) %464 = Shape(%391) %465 = Constant[value = ]() %466 = Gather[axis = 0](%464, %465) %467 = Shape(%391) %468 = Constant[value = ]() %469 = Gather[axis = 0](%467, %468) %470 = Shape(%391) %471 = Constant[value = ]() %472 = Gather[axis = 0](%470, %471) %473 = Constant[value = ]() %474 = Constant[value = ]() %475 = Unsqueeze[axes = [0]](%466) %476 = Unsqueeze[axes = [0]](%473) %477 = Unsqueeze[axes = [0]](%474) %478 = Unsqueeze[axes = [0]](%469) %479 = Unsqueeze[axes = [0]](%472) %480 = Concat[axis = 0](%475, %476, %477, %478, %479) %481 = Reshape(%391, %480) %482 = Transpose[perm = [0, 1, 3, 4, 2]](%481) return %output, %463, %482 } ONNX export success, saved as yolov5s.onnx WARNING:root:TensorFlow version 2.2.0 detected. Last version known to be fully compatible is 1.14.0 . Starting CoreML export with coremltools 3.4... CoreML export failure: module 'coremltools' has no attribute 'convert' Export complete. Visualize with https://github.com/lutzroeder/netron. ​