How do I re-train my model?

I am trying to re-train my model as I enlarged my dataset for the object detection project, I would now need to retrain the model, how would I do that, as I have seen there is a keyword “resume” but not quite sure if it is the right one.

Hi,

This depends on the frameworks you use.
In general, you can restart the training job and set the current model as the pre-trained version.

Thanks.

The line of code I was using for training at first was the following:

"python3 train_ssd.py --dataset-type=voc --data=data/{name_of_your_model_directory} --model-dir=models/{name_of_your_model_directory} --epochs={num_epochs} "

However if I run this command again for the new dataset it gives me some errors:

Do you have any clue what it could be?

Hi,

Do you change the class number or just expands the database?

Thanks.

Just enlarged the dataset, the classes are still the same

Hi @fratanto, can you try downloading the mobilenet-v1-ssd-mp-0_675.pth again? It appears to have gotten corrupted.

$ cd jetson-inference/python/training/detection/ssd
$ wget https://nvidia.box.com/shared/static/djf5w54rjvpqocsiztzaandq1m3avr7c.pth -O models/mobilenet-v1-ssd-mp-0_675.pth

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.