mindyolo
mindyolo copied to clipboard
MindSpore YOLO series toolbox and benchmark
 根目录下test.py, network construct返回值unpack失败,需更改为单输出。
模型输出和训练时,配置输入只能是正方形,若要输出长方形则需要修改源码,希望再配置策可以直接支持。 **If this is your first time, please read our contributor guidelines:** 如果这是您第一次提交请先阅读我们提供的贡献指南 https://gitee.com/mindspore/mindspore/blob/master/CONTRIBUTING.md **Is your feature request related to a problem? Please describe.** 您的需求和某个报错有关吗?请描述您的问题 A clear and concise description...
* using example finetue visdrone to train model in visdrone dataset. * can not load yolov8 pretrain model to train model ? # env single nvidia geforce rtx 4090
* 下面是详细链接。 https://gitee.com/mindspore/mindspore/issues/I9FGP1?from=project-issue
mpirun --allow-run-as-root -n 2 python train.py --config ./configs/yolov8/yolov8n1.yaml --is_parallel True 
环境:[MindCV0.2.2]MindSpore2.0.0-Cann6.3.rc1-Python3.7-Euler2.8 启动命令:python demo/predict.py --config ./configs/yolov8/yolov8m.yaml --weight=/home/ma-user/work/mindyolo/yolov8m.ckpt --image_path /home/ma-user/work/mindyolo/images/val2017/000000563653.jpg 报错: 
mpirun --allow-run-as-root -n 8 python train.py --config ./configs/yolov7/yolov7.yaml --device_target Ascend --is_parallel True 分布式训练没有mpirun 指令,单卡可以跑通yolov7
一、问题现象: 在300V Pro上基于minyolo跑yolov8推理样例,报错:[ERROR] KERNEL(90948,fffdbae749f0,python):2024-03-19-17:40:09.790.450 [mindspore/ccsrc/kernel/oplib/op_info_utils.cc:172] LoadOpInfoJson] Get op info json suffix path failed, soc_version: Ascend310P3 [ERROR] KERNEL(90948,fffdbae749f0,python):2024-03-19-17:40:09.790.546 [mindspore/ccsrc/kernel/oplib/op_info_utils.cc:111] GenerateOpInfos] Load op info json failed, version: Ascend310P3  二、软件版本: -- CANN...
将VOC标签转换为YOLO标签格式,文件夹转换为COCO格式后报错:Traceback (most recent call last): File "/home/jiasj/chenruoshui/mindyolo-master/mindyolo-master/train.py", line 321, in train(args) File "/home/jiasj/chenruoshui/mindyolo-master/mindyolo-master/train.py", line 132, in train _dataset = COCODataset( File "/home/jiasj/chenruoshui/mindyolo-master/mindyolo-master/mindyolo/data/dataset.py", line 115, in __init__ raise Exception(f"Error loading data...
一、问题表现 1、单机单卡训练时间正常 2、单机8卡训练卡住在图编译阶段-耗时严重,最后导致建链超时(默认静态模式) 拉起命令: mpirun --allow-run-as-root -n 8 python train.py --config ./configs/yolov8/yolov8n.yaml --device_target Ascend --data_dir /home/code/coco --is_parallel True  3、单机8卡训练(修改计算图为动态模式) - 可以迭代,但迭代数据较慢。  4、同环境上训练densenet121模型单机8卡训练正常。 