Evaluation about compressai

[jbegaint]

Hi @OzanKara41, I've just updated the example training script to allow training of the predefined models.
You should be able to train the bmshj2018-factorized architecture now.

[OzanKara41]

I'll use this code again, right?
python3 examples/train.py -d /path/to/my/image/dataset/ --epochs 300 -lr 1e-4 --batch-size 16 --cuda --save
After the train, a checkpoint will be created than i'll use this evalutation code below
python3 -m compressai.utils.eval_model checkpoint /path/to/images/folder/ -a bmshj2018-factorized -p checkpoint.pth.tar
Can u check it ? Thank you for quick reply.

[jbegaint]

yes that looks right!

You can set CUDA_VISIBLE_DEVICES to control the number of gpu devices available. Also the --num-workers option might be helpful to maximize the gpu usage.