search

triple-Mu / ncnn-examples

Learning ncnn with some examples
59 stars 10 forks source link

为什么你的param模型里有out0 out1 out2 我只有out0 #8

Open l7518597 opened 1 year ago

l7518597 commented 1 year ago

为什么你的param模型里有out0 out1 out2 我只有out0

triple-Mu commented 1 year ago

为什么你的param模型里有out0 out1 out2 我只有out0

因为你没看我的知乎，导出的torchscript不对

l7518597 commented 1 year ago

抱歉 我重新弄一次

l7518597 commented 1 year ago

File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics__init.py", line 5, in from ultralytics.hub import start File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\hub__init__.py", line 5, in from ultralytics.hub.auth import Auth File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\hub\auth.py", line 5, in from ultralytics.hub.utils import HUB_API_ROOT, PREFIX, request_with_credentials File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\hub\utils.py", line 14, in from ultralytics.yolo.utils import (ENVIRONMENT, LOGGER, ONLINE, RANK, SETTINGS, TESTS_RUNNING, TQDM_BAR_FORMAT, File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\yolo\init.py", line 3, in from . import v8 File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\yolo\v8\init.py", line 3, in from ultralytics.yolo.v8 import classify, detect, pose, segment File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\yolo\v8\classify\init__.py", line 4, in from ultralytics.yolo.v8.classify.train import ClassificationTrainer, train File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\yolo\v8\classify\train.py", line 6, in from ultralytics.nn.tasks import ClassificationModel, attempt_load_one_weight File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\nn\tasks.py", line 11, in from ultralytics.nn.modules import (C1, C2, C3, C3TR, SPP, SPPF, Bottleneck, BottleneckCSP, C2f, C3Ghost, C3x, Classify, File "C:\Users\Administrator\Desktop\yolov8-triplemu\ultralytics\nn\modules.py", line 558 return x, mc, p ^ IndentationError: unexpected indent

我改的V8源码下的ultralytics\nn\modules.py 内容 按知乎改的

triple-Mu commented 1 year ago

这是缩进错误！是 python 语法层面的错误！

l7518597 commented 1 year ago

可以帮忙实现一下吗 有偿 QQ275173418

JuZi233 commented 5 months ago

我用ncnn官方仓库里的onnx2ncnn转出来的param里面就只有out0，然后ncnn自己的样例代码里面使用的又都是out0,out1,out2,令人费解

cookielive commented 2 months ago

我用ncnn官方仓库里的onnx2ncnn转出来的param里面就只有out0，然后ncnn自己的样例代码里面使用的又都是out0,out1,out2,令人费解

我也有同样的疑问，yolov8转的onnx还是直接ncnn用可视化工具看都是一个输出，但是ncnn的样例需要三个输出。可视化对比了别人的三个输出的模型，发现一个输出的比三个输出的前面是一样的但后续多了很多流程，难道多的这些就是后处理？