Ayush Chaurasia comments

Results 416 comments of


                                            Ayush Chaurasia

Python sdk image segmentation does not correspond to categories

This should be fixed. Please let me know if it isnt

Question about Yolov8 architecture and Layers in details ?

Citation is available on the repo

Using v5loader=true results in non-fatal exception: _imagingft C module is not installed

Is this problem reproducible in the colab? I cannot reproduce it on my system. Maybe it might be something with your env? try setting up a new env and install...

Using v5loader=true results in non-fatal exception: _imagingft C module is not installed

What happens when you run it without `v5Loader=True`

Using v5loader=true results in non-fatal exception: _imagingft C module is not installed

@FrancescoSaverioZuppichini @sstainba I'm sorry you guys got off on the wrong foot. I sincerely hope it was just 2 nice people having a not-so-nice day and that we can move...

Using v5loader=true results in non-fatal exception: _imagingft C module is not installed

Okay got it! I'll try to fix this issue or provide a better error msg if I come across this myself.

Tensorflow module path

You can start working in nn.modules.tf ..we can change it later if we need to

speed differences between train and test

No, it's not caching. It might have a been bug that's been resolved. Please try again after updating to the latest version. Thanks!

speed differences between train and test

When you're running model.predict for the 1st time, it warms up the model and so the time should be doubled. After the 1st prediction, the times should be consistent. Try...

speed differences between train and test

Yeah but synchronize is less important. Are you manually warming up the model as I said in my previous comment?