You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
On line 103 in inference.py the weights arguments are not inserted.
The correct line should be model = infer_engine.initialize_model_from_cfg(args.weights)
Edit:
After giving the weights file the following error occurs
'
Traceback (most recent call last):
File "tools/inference.py", line 160, in
main(args)
File "tools/inference.py", line 103, in main
model = infer_engine.initialize_model_from_cfg(args.weights)
File "/detectron/detectron/core/test_engine.py", line 328, in initialize_model_from_cfg
model = model_builder.create(cfg.MODEL.TYPE, train=False, gpu_id=gpu_id)
File "/detectron/detectron/modeling/model_builder.py", line 120, in create
init_params=train
File "/detectron/detectron/modeling/detector.py", line 49, in init
assert self.num_classes > 0, 'num_classes must be > 0'
AssertionError: num_classes must be > 0
'
I am using the docker image of the detectron. And was able to run the demo of infer_simple.py on a picture. so i think the detectron setup is not a problem.
The text was updated successfully, but these errors were encountered:
On line 103 in inference.py the weights arguments are not inserted.
The correct line should be
model = infer_engine.initialize_model_from_cfg(args.weights)
Edit:
After giving the weights file the following error occurs
'
Traceback (most recent call last):
File "tools/inference.py", line 160, in
main(args)
File "tools/inference.py", line 103, in main
model = infer_engine.initialize_model_from_cfg(args.weights)
File "/detectron/detectron/core/test_engine.py", line 328, in initialize_model_from_cfg
model = model_builder.create(cfg.MODEL.TYPE, train=False, gpu_id=gpu_id)
File "/detectron/detectron/modeling/model_builder.py", line 120, in create
init_params=train
File "/detectron/detectron/modeling/detector.py", line 49, in init
assert self.num_classes > 0, 'num_classes must be > 0'
AssertionError: num_classes must be > 0
'
I am using the docker image of the detectron. And was able to run the demo of infer_simple.py on a picture. so i think the detectron setup is not a problem.
The text was updated successfully, but these errors were encountered: