Error trying to test a trained model (X3D_S)

Hi,

Has anyone tried to test a trained model using steps described [here](https://github.com/facebookresearch/SlowFast/blob/4b61d24136538b7584da7240c8242966980b88d2/GETTING_STARTED.md)?

I'm trying to run the command:
`python tools/run_net.py --cfg configs/Kinetics/X3D_S.yaml DATA.PATH_TO_DATA_DIR TEST.CHECKPOINT_FILE_PATH TRAIN.ENABLE False NUM_GPUS 1 DATA.DECODING_BACKEND pyav MODEL.NUM_CLASSES 226 `

but the script ends with after this exception has been raised:
`RuntimeError: Failed to fetch video after 10 retries.`

I have trained a X3D_L model from scratch and the during the train some video could not been decoded but it has never stopped the training process (as said [here](https://github.com/facebookresearch/SlowFast/issues/123)). But now, I want to test the last checkpoint saved and I have the exception described above.

In advance, thanks for your time.