Collections: - Name: SOLO Metadata: Training Data: COCO Training Techniques: - SGD with Momentum - Weight Decay Training Resources: 8x V100 GPUs Architecture: - FPN - Convolution - ResNet Paper: https://arxiv.org/abs/1912.04488 README: configs/solo/README.md Models: - Name: decoupled-solo_r50_fpn_1x_coco In Collection: SOLO Config: configs/solo/decoupled-solo_r50_fpn_1x_coco.py Metadata: Training Memory (GB): 7.8 Epochs: 12 inference time (ms/im): - value: 116.4 hardware: V100 backend: PyTorch batch size: 1 mode: FP32 resolution: (1333, 800) Results: - Task: Instance Segmentation Dataset: COCO Metrics: mask AP: 33.9 Weights: https://download.openmmlab.com/mmdetection/v2.0/solo/decoupled_solo_r50_fpn_1x_coco/decoupled_solo_r50_fpn_1x_coco_20210820_233348-6337c589.pth - Name: decoupled-solo_r50_fpn_3x_coco In Collection: SOLO Config: configs/solo/decoupled-solo_r50_fpn_3x_coco.py Metadata: Training Memory (GB): 7.9 Epochs: 36 inference time (ms/im): - value: 117.2 hardware: V100 backend: PyTorch batch size: 1 mode: FP32 resolution: (1333, 800) Results: - Task: Instance Segmentation Dataset: COCO Metrics: mask AP: 36.7 Weights: https://download.openmmlab.com/mmdetection/v2.0/solo/decoupled_solo_r50_fpn_3x_coco/decoupled_solo_r50_fpn_3x_coco_20210821_042504-7b3301ec.pth - Name: decoupled-solo-light_r50_fpn_3x_coco In Collection: SOLO Config: configs/solo/decoupled-solo-light_r50_fpn_3x_coco.py Metadata: Training Memory (GB): 2.2 Epochs: 36 inference time (ms/im): - value: 35.0 hardware: V100 backend: PyTorch batch size: 1 mode: FP32 resolution: (852, 512) Results: - Task: Instance Segmentation Dataset: COCO Metrics: mask AP: 32.9 Weights: https://download.openmmlab.com/mmdetection/v2.0/solo/decoupled_solo_light_r50_fpn_3x_coco/decoupled_solo_light_r50_fpn_3x_coco_20210906_142703-e70e226f.pth - Name: solo_r50_fpn_3x_coco In Collection: SOLO Config: configs/solo/solo_r50_fpn_3x_coco.py Metadata: Training Memory (GB): 7.4 Epochs: 36 inference time (ms/im): - value: 94.2 hardware: V100 backend: PyTorch batch size: 1 mode: FP32 resolution: (1333, 800) Results: - Task: Instance Segmentation Dataset: COCO Metrics: mask AP: 35.9 Weights: https://download.openmmlab.com/mmdetection/v2.0/solo/solo_r50_fpn_3x_coco/solo_r50_fpn_3x_coco_20210901_012353-11d224d7.pth - Name: solo_r50_fpn_1x_coco In Collection: SOLO Config: configs/solo/solo_r50_fpn_1x_coco.py Metadata: Training Memory (GB): 8.0 Epochs: 12 inference time (ms/im): - value: 95.1 hardware: V100 backend: PyTorch batch size: 1 mode: FP32 resolution: (1333, 800) Results: - Task: Instance Segmentation Dataset: COCO Metrics: mask AP: 33.1 Weights: https://download.openmmlab.com/mmdetection/v2.0/solo/solo_r50_fpn_1x_coco/solo_r50_fpn_1x_coco_20210821_035055-2290a6b8.pth