- 博客(15)
- 收藏
- 关注
原创 MMDetection安装报错
AssertionError: MMCV==1.4.2 is used but incompatible. Please install mmcv>=1.2.4,
2023-07-21 10:14:58
864
1
原创 YOLOv5复现问题小记
*数据集构造/转换参照Yolov5训练自己的数据集(详细完整版)遇到问题:解决方案:卸载掉comet_ml具体操作如下:pip uninstall comet_ml遇到问题:解决方案:在utils/general.py的import os。
2023-07-06 08:12:21
1770
3
原创 Ubuntu下Python程序font = ImageFont.truetype(‘arial.ttf‘, 24)报错OSError: cannot open resource
OSError: cannot open resource
2022-12-08 21:15:46
1182
原创 之前可跑通的程序在重新配置虚拟环境后报错RuntimeError: CUDA error: device-side assert triggered
之前可跑通的程序在重新配置虚拟环境后报错RuntimeError: CUDA error: device-side assert triggered
2022-09-06 09:04:19
2103
原创 服务器NVIDIA A30报错:NVIDIA A30 with CUDA capability sm_ 80
服务器NVIDIA A30使用PyTorch训练模型时报错:NVIDIA A30 with CUDA capability sm_ 80 is not compatible with the current PyTorch installation.
2022-08-30 21:21:56
3524
原创 报错:ValueError: At least one stride in the given numpy array is negative, and tensors with negative s
Torch 数据导入时报错:ValueError: At least one stride in the given numpy array is negative, and tensors with negative strides are not currently supported. (You can probably work around this by making a copy of your array with array.copy())
2022-07-20 16:55:02
3285
1
原创 PVTv2 ——KeyError:‘RetinaNet:“pvt_v2_b1:\‘pretrained‘“‘
PVTv2 ——KeyError:‘RetinaNet:“pvt_v2_b1:‘pretrained’”’#问题说明:以retinanet_pvt_v2_b1_fpn_1x_coco.py为例,测试时提示错误KeyError:'RetinaNet:“pvt_v2_b1:‘pretrained’”'#解决方案:将pvt_v2.py文件中class pvt_v2_b1(PyramidVisionTransformerV2)中super().__init__中的pretrained=kwargs[‘pret
2022-01-07 21:52:11
2312
1
空空如也
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人