# 2019 第五届“四维图新”杯创新大赛 [自动驾驶视觉综合感知算法赛](https://www.datafountain.cn/competitions/366) detection baseline
**testA object detection map**: 0.22
**testA semantic segmentation miou**: 0.46
**author**： zhengye
for the **semantic segmentation** part, please refer to my teammate's git repo: https://github.com/SHERLOCKLS/datafountain_siweituxin_autodriver_seg
## 环境配置
请按照[mmdetection](https://github.com/open-mmlab/mmdetection)说明进行安装配置
### 训练
- **数据准备**
将训练数据所有训练图片放置于 data/siweituxin/train_image
两批次的数据label分别位于'data/dataset1/train.txt' 和 'data/DF_1018/train.txt'
合并label： python tools/convert_datasets/merage_txt_label.py
txt转coco json： python tools/convert_datasets/trans_txt2json.py
- **模型训练**
请依据mmdetection说明依据自身显卡情况线性调整lr，这里以4卡为例
CUDA_VISIBLE_DEVICES=0,1,2,3 ./tools/dist_train.sh config/siweituxin/faster_rcnn_r50_fpn.py 4
### 预测
- **数据准备**
将测试数据所有训练图片放置于 data/siweituxin/test_images
- **inference**
CUDA_VISIBLE_DEVICES=0 python tools/infer_siweituxin.py config/siweituxin/faster_rcnn_r50_fpn.py /work_dirs/faster_rcnn_r50_fpn/latest.pth --out det.txt
## Contact
This repo is currently maintained by Ye Zheng ([@zhengye1995](https://github.com/zhengye1995)).
2019 第五届“四维图新”杯创新大赛 自动驾驶视觉综合感知 检测baseline.zip （211个子文件） 
deform_conv_cuda.cpp 29KB deform_pool_cuda.cpp 4KB roi_align_cuda.cpp 3KB roi_pool_cuda.cpp 3KB masked_conv2d_cuda.cpp 3KB nms_cpu.cpp 2KB sigmoid_focal_loss.cpp 2KB nms_cuda.cpp 575B deform_conv_cuda_kernel.cu 41KB deform_pool_cuda_kernel.cu 16KB roi_align_kernel.cu 11KB roi_pool_kernel.cu 6KB sigmoid_focal_loss_cuda.cu 6KB nms_kernel.cu 5KB masked_conv2d_kernel.cu 5KB pytest.ini 293B LICENSE 11KB README.md 1KB transforms.py 29KB reppoints_head.py 27KB guided_anchor_head.py 25KB htc.py 24KB cascade_rcnn.py 23KB hrnet.py 19KB resnet.py 18KB test_robustness.py 17KB fcos_head.py 16KB fovea_head.py 16KB mean_ap.py 15KB grid_head.py 15KB generalized_attention.py 15KB flops_counter.py 14KB anchor_head.py 13KB deform_conv.py 12KB two_stage.py 12KB guided_anchor_target.py 12KB deform_pool.py 10KB grid_rcnn.py 9KB bbox_head.py 9KB train.py 9KB mask_scoring_rcnn.py 8KB robustness_eval.py 8KB resnext.py 8KB coco_utils.py 8KB test.py 8KB transforms.py 8KB ssd_head.py 8KB double_head_rcnn.py 7KB maskiou_head.py 7KB test_config.py 7KB anchor_target.py 7KB test_mixins.py 7KB fcn_mask_head.py 7KB convfc_bbox_head.py 7KB coco_error_analysis.py 7KB max_iou_assigner.py 6KB setup.py 6KB point_target.py 6KB ghm_loss.py 6KB eval_hooks.py 6KB analyze_logs.py 6KB decorators.py 6KB formating.py 6KB inference.py 6KB recall.py 6KB iou_balanced_neg_sampler.py 6KB sampler.py 6KB conv_module.py 6KB class_names.py 5KB faster_rcnn_r50_fpn.py 5KB ssd_vgg.py 5KB fpn.py 5KB double_bbox_head.py 5KB loading.py 5KB point_assigner.py 5KB base.py 5KB custom.py 5KB ga_rpn_head.py 5KB approx_max_iou_assigner.py 5KB pascal_voc.py 5KB trans_txt2json.py 4KB hooks.py 4KB iou_loss.py 4KB coco.py 4KB rpn_head.py 4KB detectron2pytorch.py 4KB single_level.py 4KB context_block.py 4KB ga_retina_head.py 4KB non_local.py 4KB retina_head.py 4KB siweituxin.py 4KB bfp.py 4KB merge_augs.py 3KB fused_semantic_head.py 3KB anchor_generator.py 3KB rpn.py 3KB train.py 3KB cross_entropy_loss.py 3KB masked_conv.py 3KB共 211 条
- 1
- 2
- 3
资源评论
