Closed piyushandilya closed 1 year ago
@joslefaure I'd really appreciate some help from you in this.
I have been looking into it, but I can't reproduce the issue. You can try to uncomment line 27 in train_net.py
but it will help only if the error has to do with Cuda.
@joslefaure Can you share your environment/requirements.txt ? Maybe it's related to using a different version of torch (or any other dependency).
I changed my pytorch version to 1.7.1 and the error got resolved.
i wonder how u solve such question? @piyushandilya looking forward to your reply
I am trying to train the model (JHMDB) with the weights SlowFast-ResNet50-4x16.pth (soft link created) at path HIT/data/models/pretrained_models/
However, when I try to train the model using
python train_net.py --config-file "config_files/hitnet.yaml"
I get an error saying that Conv3D received invalid arguments:This error is possibly originating at the
reduce_person_dim
function. Am I using the project incorrectly? Can you help me fix this?