更换文档检测模型

2024-08-27 14:42:45 +08:00
parent aea6f19951
commit 1514e09c40
2072 changed files with 254336 additions and 4967 deletions
--- a/paddle_detection/configs/rcnn_enhance/README.md
+++ b/paddle_detection/configs/rcnn_enhance/README.md
@@ -0,0 +1,12 @@
+## 服务器端实用目标检测方案
+
+### 简介
+
+* 近年来，学术界和工业界广泛关注图像中目标检测任务。基于[PaddleClas](https://github.com/PaddlePaddle/PaddleClas)中SSLD蒸馏方案训练得到的ResNet50_vd预训练模型(ImageNet1k验证集上Top1 Acc为82.39%)，结合PaddleDetection中的丰富算子，飞桨提供了一种面向服务器端实用的目标检测方案PSS-DET(Practical Server Side Detection)。基于COCO2017目标检测数据集，V100单卡预测速度为61FPS时，COCO mAP可达41.2%。
+
+
+### 模型库
+
+| 骨架网络             | 网络类型       | 每张GPU图片个数 | 学习率策略 |推理时间(fps) | Box AP | Mask AP |                           下载                          | 配置文件 |
+| :---------------------- | :-------------:  | :-------: | :-----: | :------------: | :----: | :-----: | :-------------: | :-----: |
+| ResNet50-vd-FPN-Dcnv2         | Faster     |     2     |   3x    |     61.425     |  41.5  |    -    | [下载链接](https://paddledet.bj.bcebos.com/models/faster_rcnn_enhance_3x_coco.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/rcnn_enhance/faster_rcnn_enhance_3x_coco.yml) |
--- a/paddle_detection/configs/rcnn_enhance/README_en.md
+++ b/paddle_detection/configs/rcnn_enhance/README_en.md
@@ -0,0 +1,12 @@
+## Practical Server Side Detection
+
+### Introduction
+
+* In recent years, the object detection task in image has been widely concerned by academia and industry. ResNet50vd pretraining model based on SSLD distillation program training in [PaddleClas](https://github.com/PaddlePaddle/PaddleClas) (Top1 on ImageNet1k verification set) Acc is 82.39%), combined with the rich operator of PaddleDetection, PaddlePaddle provides a practical server side detection scheme PSS-DET(Practical Server Side Detection). Based on COCO2017 object detection dataset, V100 single gpu prediction speed is 61FPS, COCO mAP can reach 41.2%.
+
+
+### Model library
+
+| Backbone              | Network type | Number of images per GPU | Learning rate strategy | Inferring time(fps) | Box AP | Mask AP |                                      Download                                       |                                                           Configuration File                                                            |
+| :-------------------- | :----------: | :----------------------: | :--------------------: | :-----------------: | :----: | :-----: | :---------------------------------------------------------------------------------: | :-------------------------------------------------------------------------------------------------------------------------------------: |
+| ResNet50-vd-FPN-Dcnv2 |    Faster    |            2             |           3x           |       61.425        |  41.5  |    -    | [link](https://paddledet.bj.bcebos.com/models/faster_rcnn_enhance_3x_coco.pdparams) | [Configuration File](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/rcnn_enhance/faster_rcnn_enhance_3x_coco.yml) |
--- a/paddle_detection/configs/rcnn_enhance/_base_/faster_rcnn_enhance.yml
+++ b/paddle_detection/configs/rcnn_enhance/_base_/faster_rcnn_enhance.yml
@@ -0,0 +1,81 @@
+architecture: FasterRCNN
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/ResNet50_vd_ssld_v2_pretrained.pdparams
+
+FasterRCNN:
+  backbone: ResNet
+  neck: FPN
+  rpn_head: RPNHead
+  bbox_head: BBoxHead
+  # post process
+  bbox_post_process: BBoxPostProcess
+
+
+ResNet:
+  # index 0 stands for res2
+  depth: 50
+  norm_type: bn
+  variant: d
+  freeze_at: 0
+  return_idx: [0,1,2,3]
+  num_stages: 4
+  dcn_v2_stages: [1,2,3]
+  lr_mult_list: [0.05, 0.05, 0.1, 0.15]
+
+FPN:
+  in_channels: [256, 512, 1024, 2048]
+  out_channel: 64
+
+RPNHead:
+  anchor_generator:
+    aspect_ratios: [0.5, 1.0, 2.0]
+    anchor_sizes: [[32], [64], [128], [256], [512]]
+    strides: [4, 8, 16, 32, 64]
+  rpn_target_assign:
+    batch_size_per_im: 256
+    fg_fraction: 0.5
+    negative_overlap: 0.3
+    positive_overlap: 0.7
+    use_random: True
+  train_proposal:
+    min_size: 0.0
+    nms_thresh: 0.7
+    pre_nms_top_n: 2000
+    post_nms_top_n: 2000
+    topk_after_collect: True
+  test_proposal:
+    min_size: 0.0
+    nms_thresh: 0.7
+    pre_nms_top_n: 500
+    post_nms_top_n: 300
+
+
+BBoxHead:
+  head: TwoFCHead
+  roi_extractor:
+    resolution: 7
+    sampling_ratio: 0
+    aligned: True
+  bbox_assigner: BBoxLibraAssigner
+  bbox_loss: DIouLoss
+
+TwoFCHead:
+  out_channel: 1024
+
+BBoxLibraAssigner:
+  batch_size_per_im: 512
+  bg_thresh: 0.5
+  fg_thresh: 0.5
+  fg_fraction: 0.25
+  use_random: True
+
+DIouLoss:
+  loss_weight: 10.0
+  use_complete_iou_loss: true
+
+BBoxPostProcess:
+  decode: RCNNBox
+  nms:
+    name: MultiClassNMS
+    keep_top_k: 100
+    score_threshold: 0.05
+    nms_threshold: 0.5
--- a/paddle_detection/configs/rcnn_enhance/_base_/faster_rcnn_enhance_reader.yml
+++ b/paddle_detection/configs/rcnn_enhance/_base_/faster_rcnn_enhance_reader.yml
@@ -0,0 +1,42 @@
+worker_num: 2
+TrainReader:
+  sample_transforms:
+  - Decode: {}
+  - AutoAugment: {autoaug_type: v1}
+  - RandomResize: {target_size: [[384,1000], [416,1000], [448,1000], [480,1000], [512,1000], [544,1000], [576,1000], [608,1000], [640,1000], [672,1000]], interp: 2, keep_ratio: True}
+  - RandomFlip: {prob: 0.5}
+  - 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: 2
+  shuffle: true
+  drop_last: true
+  collate_batch: false
+  use_shared_memory: true
+
+
+EvalReader:
+  sample_transforms:
+  - Decode: {}
+  - Resize: {interp: 2, target_size: [640, 640], 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
+  drop_last: false
+
+
+TestReader:
+  sample_transforms:
+  - Decode: {}
+  - Resize: {interp: 2, target_size: [640, 640], 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
+  drop_last: false
--- a/paddle_detection/configs/rcnn_enhance/_base_/optimizer_3x.yml
+++ b/paddle_detection/configs/rcnn_enhance/_base_/optimizer_3x.yml
@@ -0,0 +1,19 @@
+epoch: 36
+
+LearningRate:
+  base_lr: 0.02
+  schedulers:
+  - !PiecewiseDecay
+    gamma: 0.1
+    milestones: [24, 33]
+  - !LinearWarmup
+    start_factor: 0.
+    steps: 1000
+
+OptimizerBuilder:
+  optimizer:
+    momentum: 0.9
+    type: Momentum
+  regularizer:
+    factor: 0.0001
+    type: L2
--- a/paddle_detection/configs/rcnn_enhance/faster_rcnn_enhance_3x_coco.yml
+++ b/paddle_detection/configs/rcnn_enhance/faster_rcnn_enhance_3x_coco.yml
@@ -0,0 +1,8 @@
+_BASE_: [
+  '../datasets/coco_detection.yml',
+  '../runtime.yml',
+  '_base_/optimizer_3x.yml',
+  '_base_/faster_rcnn_enhance.yml',
+  '_base_/faster_rcnn_enhance_reader.yml',
+]
+weights: output/faster_rcnn_enhance_r50_3x_coco/model_final