Xieyangyang

@Note-Liu have you solverd it？？

@AthenaAlala you need to change the learningrate #146

/pytorch/aten/src/ATen/native/cuda/IndexKernel.cu:60: operator(): block: [41/pytorch/aten/src/ATen/native/cuda/IndexKernel.cu:60: operator(): block: [41,0,0,0], thread: [0,0,0], thread: [0,0,0] Assertion `index >= -sizes[i] && index < sizes[i] && "index out of bounds",0] Assertion `index >= -sizes[i] && index...

@wuzhihao7788 你好，单卡没有问题； coco数据转换过来，标签是从0开始吧？

@wuzhihao7788 你能发一份COCO的标签文件给我吗？train.txt val.txt和label.names 我先验证一下我的数据制作有没有问题。我的邮箱[email protected]

@wuzhihao7788 我发现在yolodet/models/heads/yolo.py文件中 “”“ pi = pred b, a, gj, gi = indices # image, anchor, gridy, gridx print("pi: ", pi.shape) print("b: ", b) print("a: ", a) print("gj: ", gj) print("gi:...

@wuzhihao7788 问题应该是多卡训练时候图片被分到了多张卡，但是计算loss时，对应的标签用的是整个batchsize的。

@wuzhihao7788 你好，可以了

@wuzhihao7788 测试的时候遇到错误 Traceback (most recent call last): File "tools/train.py", line 144, in main() File "tools/train.py", line 140, in main train_detector(model,datasets,cfg,validate=args.validate,timestamp=timestamp,meta=meta) File "/data1/xieyangyang/yolodet-pytorch/yolodet/apis/train.py", line 161, in train_detector runner.run(data_loaders, cfg.workflow, cfg.total_epochs) File...

@wuzhihao7788 你好，训练和测试都正常了； 我在coco上训练yolov5 4个epoch后的测试结果 Class Images Targets P R [email protected] [email protected]:.95: all 4.94e+03 3.68e+04 0.109 0.013 0.0268 0.0155 这个正常嘛？ 能发一份你的训练log给我嘛 [email protected]