Kyle Wang
Kyle Wang
you can read the code in ./lib/dataset/pascal_voc.py where the images & masks are loaded, splitted by class/instance, stored in .hkl as cache and then an index database (sdsdb) are created.
ResourceExhaustedError means you don't have enough GPU memory, you can try to use smaller data shape or upgrade your GPU
you can download it from here: http://www.cs.utexas.edu/users/ml/clamp/videoDescription/
I've removed all the tf.get_variable_scope().reuse_variables() and solve this problem, but I'm new to tf and not clear about the consequence.
I've tested it on a single GTX1080, got about 83 fps on forward using caffe time
Hi @MenglingD, I'm also working on this model. Do you think this modification can also be applied to swin transformer v2? And have you tried to convert this model to...
If you've logged in before, you can copy the token file `${HOME}/.gdrive/token_v2.json` to the same place in your new machine, then you can skip the login process.
Seems that some irrelative tests failed, most of which are because of environments or resources problems. Should I retry them or just ignore?
Sure, I'd love to, but unfortunately I'm getting a small surgery right now(not very convenient to bring my laptop)😥. I'll get back to this later this week. Is it ok...
Hi @ezyang @voznesenskym ,I tried to import and add the required modules to `CLOSURE_VARS` in `VariableBuilder.__call__()`method, which I think is a feasible approach but maybe not the laziest one. Maybe...