inference is not work

here is my inference script `import mmcv import os import numpy as np from tqdm import tqdm import json import cv2

def show_result(result): bbox_result = result bboxes = np.vstack(bbox_result) labels = [ np.full(bbox.shape[0], i, dtype=np.int32) for i, bbox in enumerate(bbox_result) ] labels = np.concatenate(labels) return bboxes,labels

config_file = ‘/data/lzy_intern/mmdetection/configs/my_config/image_ann_cascade_rcnn_x101_64x4d_fpn_1x.py’ checkpoint_file = ‘/data/lzy_intern/mmdetection/models/cascade_rcnn_x101_64x4d_fpn_1x.pth’ model = init_detector(config_file, checkpoint_file, device = ‘cuda:2’)

image = mmcv.imread(‘/data/lzy_intern/dataset/coco/train2017/000000174601.jpg’) result = inference_detector(model,image) bbox,label = show_result(result) print(len(bbox)) print(len(label))`

the checkpoint model is download from https://s3.ap-northeast-2.amazonaws.com/open-mmlab/mmdetection/models/cascade_rcnn_x101_64x4d_fpn_1x_20181218-e2dc376a.pth and the config file is the deafult version of mmdetection, but why is output of len(bbox) and len(label) is 0, Is the inference is not work?

[hellock]

You may try device = 'cuda:0'.

[AlaylmYC]

The solution is very simple, but I’ve been bothered by this for a few days -_-||. I want to use 5th gpu, so I set device = 'cuda:5' and run my test script with python test.py this is not work and get nothing. The right way is: use device = 'cuda:0' and run like this: CUDA_VISIBLE_DEVICES=5 python test.py This works for me, but I’m confused. Can you explain it in detail @hellock

I meet the same problem. It bothers me a lot. At last I find that it works only use device = ‘cuda:0’. The result will become abnormal when use other devices such as device = ‘cuda:1’ or as device = ‘cuda:2’ Could you explain it in detail？ Thanks! @hellock