diff --git a/configs/faster_rcnn_r34_fpn_1x.yml b/configs/faster_rcnn_r34_fpn_1x.yml new file mode 100644 index 0000000000000000000000000000000000000000..b97dc33d5d0ededf5036017e3e04a2bacf96f7bc --- /dev/null +++ b/configs/faster_rcnn_r34_fpn_1x.yml @@ -0,0 +1,106 @@ +architecture: FasterRCNN +max_iters: 90000 +use_gpu: true +snapshot_iter: 10000 +log_smooth_window: 20 +save_dir: output +pretrain_weights: ResNet34_pretrained +metric: COCO +weights: output/faster_rcnn_r34_fpn_1x/model_final +num_classes: 81 + +FasterRCNN: + backbone: ResNet + fpn: FPN + rpn_head: FPNRPNHead + roi_extractor: FPNRoIAlign + bbox_head: BBoxHead + bbox_assigner: BBoxAssigner + +ResNet: + norm_type: bn + norm_decay: 0. + depth: 34 + feature_maps: [2, 3, 4, 5] + freeze_at: 2 + +FPN: + min_level: 2 + max_level: 6 + num_chan: 256 + spatial_scale: [0.03125, 0.0625, 0.125, 0.25] + +FPNRPNHead: + anchor_generator: + anchor_sizes: [32, 64, 128, 256, 512] + aspect_ratios: [0.5, 1.0, 2.0] + stride: [16.0, 16.0] + variance: [1.0, 1.0, 1.0, 1.0] + anchor_start_size: 32 + min_level: 2 + max_level: 6 + num_chan: 256 + rpn_target_assign: + rpn_batch_size_per_im: 256 + rpn_fg_fraction: 0.5 + rpn_positive_overlap: 0.7 + rpn_negative_overlap: 0.3 + rpn_straddle_thresh: 0.0 + train_proposal: + min_size: 0.0 + nms_thresh: 0.7 + pre_nms_top_n: 2000 + post_nms_top_n: 2000 + test_proposal: + min_size: 0.0 + nms_thresh: 0.7 + pre_nms_top_n: 1000 + post_nms_top_n: 1000 + +FPNRoIAlign: + canconical_level: 4 + canonical_size: 224 + min_level: 2 + max_level: 5 + box_resolution: 7 + sampling_ratio: 2 + +BBoxAssigner: + batch_size_per_im: 512 + bbox_reg_weights: [0.1, 0.1, 0.2, 0.2] + bg_thresh_lo: 0.0 + bg_thresh_hi: 0.5 + fg_fraction: 0.25 + fg_thresh: 0.5 + +BBoxHead: + head: TwoFCHead + nms: + keep_top_k: 100 + nms_threshold: 0.5 + score_threshold: 0.05 + +TwoFCHead: + mlp_dim: 1024 + +LearningRate: + base_lr: 0.02 + schedulers: + - !PiecewiseDecay + gamma: 0.1 + milestones: [60000, 80000] + - !LinearWarmup + start_factor: 0.1 + steps: 1000 + +OptimizerBuilder: + optimizer: + momentum: 0.9 + type: Momentum + regularizer: + factor: 0.0001 + type: L2 + +_READER_: 'faster_fpn_reader.yml' +TrainReader: + batch_size: 2 diff --git a/configs/faster_rcnn_r34_vd_fpn_1x.yml b/configs/faster_rcnn_r34_vd_fpn_1x.yml new file mode 100644 index 0000000000000000000000000000000000000000..68e061c62f65ea74a35f761bde31aeaa38baf848 --- /dev/null +++ b/configs/faster_rcnn_r34_vd_fpn_1x.yml @@ -0,0 +1,107 @@ +architecture: FasterRCNN +max_iters: 90000 +use_gpu: true +snapshot_iter: 10000 +log_smooth_window: 20 +save_dir: output +pretrain_weights: https://paddle-imagenet-models-name.bj.bcebos.com/ResNet34_vd_pretrained.tar +metric: COCO +weights: output/faster_rcnn_r34_fpn_1x/model_final +num_classes: 81 + +FasterRCNN: + backbone: ResNet + fpn: FPN + rpn_head: FPNRPNHead + roi_extractor: FPNRoIAlign + bbox_head: BBoxHead + bbox_assigner: BBoxAssigner + +ResNet: + norm_type: bn + norm_decay: 0. + depth: 34 + feature_maps: [2, 3, 4, 5] + freeze_at: 2 + variant: d + +FPN: + min_level: 2 + max_level: 6 + num_chan: 256 + spatial_scale: [0.03125, 0.0625, 0.125, 0.25] + +FPNRPNHead: + anchor_generator: + anchor_sizes: [32, 64, 128, 256, 512] + aspect_ratios: [0.5, 1.0, 2.0] + stride: [16.0, 16.0] + variance: [1.0, 1.0, 1.0, 1.0] + anchor_start_size: 32 + min_level: 2 + max_level: 6 + num_chan: 256 + rpn_target_assign: + rpn_batch_size_per_im: 256 + rpn_fg_fraction: 0.5 + rpn_positive_overlap: 0.7 + rpn_negative_overlap: 0.3 + rpn_straddle_thresh: 0.0 + train_proposal: + min_size: 0.0 + nms_thresh: 0.7 + pre_nms_top_n: 2000 + post_nms_top_n: 2000 + test_proposal: + min_size: 0.0 + nms_thresh: 0.7 + pre_nms_top_n: 1000 + post_nms_top_n: 1000 + +FPNRoIAlign: + canconical_level: 4 + canonical_size: 224 + min_level: 2 + max_level: 5 + box_resolution: 7 + sampling_ratio: 2 + +BBoxAssigner: + batch_size_per_im: 512 + bbox_reg_weights: [0.1, 0.1, 0.2, 0.2] + bg_thresh_lo: 0.0 + bg_thresh_hi: 0.5 + fg_fraction: 0.25 + fg_thresh: 0.5 + +BBoxHead: + head: TwoFCHead + nms: + keep_top_k: 100 + nms_threshold: 0.5 + score_threshold: 0.05 + +TwoFCHead: + mlp_dim: 1024 + +LearningRate: + base_lr: 0.02 + schedulers: + - !PiecewiseDecay + gamma: 0.1 + milestones: [60000, 80000] + - !LinearWarmup + start_factor: 0.1 + steps: 1000 + +OptimizerBuilder: + optimizer: + momentum: 0.9 + type: Momentum + regularizer: + factor: 0.0001 + type: L2 + +_READER_: 'faster_fpn_reader.yml' +TrainReader: + batch_size: 2 diff --git a/docs/MODEL_ZOO.md b/docs/MODEL_ZOO.md index dcf66d2cf1d591b203dc6e05193c3cc47f0f7b3c..2198db54a680db2322382ac99a9b2a5e8cac6939 100644 --- a/docs/MODEL_ZOO.md +++ b/docs/MODEL_ZOO.md @@ -41,6 +41,8 @@ The backbone models pretrained on ImageNet are available. All backbone models ar | ResNet50 | Mask | 1 | 1x | 11.615 | 36.5 | 32.2 | [model](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_1x.tar) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/mask_rcnn_r50_1x.yml) | | ResNet50 | Mask | 1 | 2x | 11.494 | 38.2 | 33.4 | [model](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_2x.tar) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/mask_rcnn_r50_2x.yml) | | ResNet50-vd | Faster | 1 | 1x | 12.575 | 36.4 | - | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_vd_1x.tar) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r50_vd_1x.yml) | +| ResNet34-FPN | Faster | 2 | 1x | - | 36.7 | - | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r34_fpn_1x.pdparams) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r34_fpn_1x.yml) | +| ResNet34-vd-FPN | Faster | 2 | 1x | - | 37.4 | - | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r34_vd_fpn_1x.pdparams) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r34_vd_fpn_1x.yml) | | ResNet50-FPN | Faster | 2 | 1x | 22.273 | 37.2 | - | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_fpn_1x.tar) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r50_fpn_1x.yml) | | ResNet50-FPN | Faster | 2 | 2x | 22.297 | 37.7 | - | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_fpn_2x.tar) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r50_fpn_2x.yml) | | ResNet50-FPN | Mask | 1 | 1x | 15.184 | 37.9 | 34.2 | [model](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_fpn_1x.tar) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/mask_rcnn_r50_fpn_1x.yml) | diff --git a/docs/MODEL_ZOO_cn.md b/docs/MODEL_ZOO_cn.md index 778d866e2261fdb572645e8d1a5fce7df05b659b..4454367b64178304f9374217d401d5de18c7e17c 100644 --- a/docs/MODEL_ZOO_cn.md +++ b/docs/MODEL_ZOO_cn.md @@ -38,6 +38,8 @@ Paddle提供基于ImageNet的骨架网络预训练模型。所有预训练模型 | ResNet50 | Mask | 1 | 1x | 11.615 | 36.5 | 32.2 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_1x.tar) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/mask_rcnn_r50_1x.yml) | | ResNet50 | Mask | 1 | 2x | 11.494 | 38.2 | 33.4 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_2x.tar) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/mask_rcnn_r50_2x.yml) | | ResNet50-vd | Faster | 1 | 1x | 12.575 | 36.4 | - | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_vd_1x.tar) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r50_vd_1x.yml) | +| ResNet34-FPN | Faster | 2 | 1x | - | 36.7 | - | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r34_fpn_1x.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r34_fpn_1x.yml) | +| ResNet34-vd-FPN | Faster | 2 | 1x | - | 37.4 | - | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r34_vd_fpn_1x.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r34_vd_fpn_1x.yml) | | ResNet50-FPN | Faster | 2 | 1x | 22.273 | 37.2 | - | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_fpn_1x.tar) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r50_fpn_1x.yml) | | ResNet50-FPN | Faster | 2 | 2x | 22.297 | 37.7 | - | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_fpn_2x.tar) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/faster_rcnn_r50_fpn_2x.yml) | | ResNet50-FPN | Mask | 1 | 1x | 15.184 | 37.9 | 34.2 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_fpn_1x.tar) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/mask_rcnn_r50_fpn_1x.yml) |