Dustin Franklin

Hi @Gianni2045 @GianniM2045, I'm having trouble downloading all the individual folders you sent me on Google Drive. Can you compress your entire dataset into a single .zip or .tar.gz archive...

> https://drive.google.com/file/d/1QaMNDQ-UAQogwbqAomApEzW1D6OlN2V_/view?usp=drive_web Hi @GianniM2045, is this the link? I've requested access to download it.

> suddes.zip OK, I requested access to this one also

Hi @eanmikale, what is the command you are using to run your model? It looks like the labels.txt has more entries than expected - your labels.txt has 280 classes, while...

Hi @FantasticCHH, at the end of each epoch you can get the training/validation accuracy from the AverageMeter [`top1.avg` ](https://github.com/dusty-nv/pytorch-classification/blob/393914b484fc946876e0a1fe37c971acdfb1fec6/train.py#L448) value. Then store these in list(s) and make the plot at...

Hi @irshadkn, it uses CUDA kernels for pre/post-processing and TensorRT for the DNN inferencing. The CUDA/TensorRT stuff is all done inside the C++ code. The GStreamer pipelines get the data,...

> 1. is it feasible to use labelimg tool for labelling? if it is not then recommend me the another external tool for labelling the images (not CVAT ). You...

@smartsmit407 the contents of train.txt/trainval.txt/ect under ImageSets/ is just a list of the Image IDs (typically the Image ID's are the filenames of the XML/JPG files, without the file extensions)....

labelimg does generate valid XML annotation files, it just doesn't put them in the correct directory structure for Pascal VOC or make the labels.txt/ImageSets files. But you can do those...

> So, what is the correct directory structure for putting the files.. Download the original Pascal VOC dataset and replicate it's directory structure.