测试并保存分割结果基础使用当需要保存测试输出的分割结果，用 --out 指定分割结果输出路径

python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} --out ${OUTPUT_DIR} 以保存模型 fcn_r50-d8_4xb4-80k_ade20k-512x512 在 ADE20K 验证数据集上的结果为例：

python tools/test.py configs/fcn/fcn_r50-d8_4xb4-80k_ade20k-512x512.py ckpt/fcn_r50-d8_512x512_80k_ade20k_20200614_144016-f8ac5082.pth --out work_dirs/format_results 或者通过配置文件定义 output_dir。例如在 configs/fcn/fcn_r50-d8_4xb4-80k_ade20k-512x512.py 添加 test_evaluator 定义：

test_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'], output_dir='work_dirs/format_results') 然后执行相同功能的命令不需要再使用 --out：

python tools/test.py configs/fcn/fcn_r50-d8_4xb4-80k_ade20k-512x512.py ckpt/fcn_r50-d8_512x512_80k_ade20k_20200614_144016-f8ac5082.pth 当测试的数据集没有提供标注，评测时没有真值可以参与计算，因此需要设置 format_only=True，同时需要修改 test_dataloader，由于没有标注，我们需要在数据增强变换中删掉 dict(type='LoadAnnotations')，以下是一个配置示例：

test_evaluator = dict( type='IoUMetric', iou_metrics=['mIoU'], format_only=True, output_dir='work_dirs/format_results') test_dataloader = dict( batch_size=1, num_workers=4, persistent_workers=True, sampler=dict(type='DefaultSampler', shuffle=False), dataset=dict( type = 'ADE20KDataset' data_root='data/ade/release_test', data_prefix=dict(img_path='testing'),

测试数据变换中没有加载标注

    pipeline=[
        dict(type='LoadImageFromFile'),
        dict(type='Resize', scale=(2048, 512), keep_ratio=True),
        dict(type='PackSegInputs')
    ]))

然后执行测试命令：

python tools/test.py configs/fcn/fcn_r50-d8_4xb4-80k_ade20k-512x512.py ckpt/fcn_r50-d8_512x512_80k_ade20k_20200614_144016-f8ac5082.pth

When I run my file, both changing the terminal command and adding a key_val to the evaluate dict are not work well. Let me show you:

This one came out when I type in this command:python tools/test.py configs/mask_rcnn_semi_auto_germinal_center/config_inference.py work_dirs/mask-rcnn_r50-caffe_fpn_ms-poly-1x_germinal_center/epoch_15.pth

after_run: (BELOW_NORMAL) LoggerHook

Traceback (most recent call last): File "tools/test.py", line 149, in main() File "tools/test.py", line 139, in main assert args.out.endswith(('.pkl', '.pickle')), \ AssertionError: The dump file must be a pkl file.

Then I found a .json file including the evaluation results and the setting recording file.

When I chose to add a key-value to the config file. One type error came out: File "/home/xuym/miniconda3/envs/openmmlab/lib/python3.8/site-packages/mmengine/registry/registry.py", line 570, in build return self.build_func(cfg, *args, kwargs, registry=self) File "/home/xuym/miniconda3/envs/openmmlab/lib/python3.8/site-packages/mmengine/registry/build_functions.py", line 121, in build_from_cfg obj = obj_cls(args) # type: ignore TypeError: init() got an unexpected keyword argument 'output_dir'

open-mmlab / mmsegmentation

The test and inference function in our docs are not work well #3678

测试数据变换中没有加载标注