更换文档检测模型

paddle_detection/configs/tood/README.md

@@ -0,0 +1,35 @@
+# TOOD
+
+## Introduction
+
+[TOOD: Task-aligned One-stage Object Detection](https://arxiv.org/abs/2108.07755)
+
+TOOD is an object detection model. We reproduced the model of the paper.
+
+
+## Model Zoo
+
+| Backbone | Model | Images/GPU  | Inf time (fps) | Box AP | Config | Download |
+|:------:|:--------:|:--------:|:--------------:|:------:|:------:|:--------:|
+| R-50 | TOOD  | 4 | --- | 42.5 | [config](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/tood/tood_r50_fpn_1x_coco.yml) | [model](https://paddledet.bj.bcebos.com/models/tood_r50_fpn_1x_coco.pdparams) |
+
+**Notes:**
+
+- TOOD is trained on COCO train2017 dataset and evaluated on val2017 results of `mAP(IoU=0.5:0.95)`.
+- TOOD uses 8GPU to train 12 epochs.
+
+GPU multi-card training
+```bash
+export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+python -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/tood/tood_r50_fpn_1x_coco.yml --fleet
+```
+
+## Citations
+```
+@inproceedings{feng2021tood,
+    title={TOOD: Task-aligned One-stage Object Detection},
+    author={Feng, Chengjian and Zhong, Yujie and Gao, Yu and Scott, Matthew R and Huang, Weilin},
+    booktitle={ICCV},
+    year={2021}
+}
+```

paddle_detection/configs/tood/_base_/optimizer_1x.yml

@@ -0,0 +1,19 @@
+epoch: 12
+
+LearningRate:
+  base_lr: 0.01
+  schedulers:
+  - !PiecewiseDecay
+    gamma: 0.1
+    milestones: [8, 11]
+  - !LinearWarmup
+    start_factor: 0.001
+    steps: 500
+
+OptimizerBuilder:
+  optimizer:
+    momentum: 0.9
+    type: Momentum
+  regularizer:
+    factor: 0.0001
+    type: L2

paddle_detection/configs/tood/_base_/tood_r50_fpn.yml

@@ -0,0 +1,42 @@
+architecture: TOOD
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/ResNet50_cos_pretrained.pdparams
+
+TOOD:
+  backbone: ResNet
+  neck: FPN
+  head: TOODHead
+
+ResNet:
+  depth: 50
+  variant: b
+  norm_type: bn
+  freeze_at: 0
+  return_idx: [1, 2, 3]
+  num_stages: 4
+
+FPN:
+  out_channel: 256
+  spatial_scales: [0.125, 0.0625, 0.03125]
+  extra_stage: 2
+  has_extra_convs: true
+  use_c5: false
+
+TOODHead:
+  stacked_convs: 6
+  grid_cell_scale: 8
+  static_assigner_epoch: 4
+  loss_weight: { class: 1.0, iou: 2.0 }
+  static_assigner:
+    name: ATSSAssigner
+    topk: 9
+  assigner:
+    name: TaskAlignedAssigner
+    topk: 13
+    alpha: 1.0
+    beta: 6.0
+  nms:
+    name: MultiClassNMS
+    nms_top_k: 1000
+    keep_top_k: 100
+    score_threshold: 0.05
+    nms_threshold: 0.6

paddle_detection/configs/tood/_base_/tood_reader.yml

@@ -0,0 +1,40 @@
+worker_num: 4
+TrainReader:
+  sample_transforms:
+    - Decode: {}
+    - RandomFlip: {prob: 0.5}
+    - Resize: {target_size: [800, 1333], keep_ratio: true}
+    - NormalizeImage: {is_scale: true, mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225]}
+    - Permute: {}
+  batch_transforms:
+    - PadBatch: {pad_to_stride: 32}
+    - PadGT: {}
+  batch_size: 4
+  shuffle: true
+  drop_last: true
+  collate_batch: true
+  use_shared_memory: true
+
+
+EvalReader:
+  sample_transforms:
+  - Decode: {}
+  - Resize: {target_size: [800, 1333], keep_ratio: True}
+  - NormalizeImage: {is_scale: true, mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225]}
+  - Permute: {}
+  batch_transforms:
+  - PadBatch: {pad_to_stride: 32}
+  batch_size: 1
+  shuffle: false
+
+
+TestReader:
+  sample_transforms:
+  - Decode: {}
+  - Resize: {target_size: [800, 1333], keep_ratio: True}
+  - NormalizeImage: {is_scale: true, mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225]}
+  - Permute: {}
+  batch_transforms:
+  - PadBatch: {pad_to_stride: 32}
+  batch_size: 1
+  shuffle: false

paddle_detection/configs/tood/tood_r50_fpn_1x_coco.yml

@@ -0,0 +1,11 @@
+_BASE_: [
+  '../datasets/coco_detection.yml',
+  '../runtime.yml',
+  '_base_/tood_r50_fpn.yml',
+  '_base_/optimizer_1x.yml',
+  '_base_/tood_reader.yml',
+]
+
+weights: output/tood_r50_fpn_1x_coco/model_final
+find_unused_parameters: True
+log_iter: 100