I used the evaluation command to run model inference after training, but there was an error happened.
the command is:
python src/apply_net.py --dataset-dir COCO_DATASET_ROOT --test-dataset coco_2017_custom_val --config-file COCO-Detection/retinanet/retinanet_R_50_FPN_3x_reg_var_es_dropout.yaml --inference-config Inference/bayes_od_mc_dropout.yaml --random-seed 0 --image-corruption-level 0
and the error is:
Traceback (most recent call last):
File "src/apply_net.py", line 112, in
launch(
File "/root/miniconda3/envs/probdet/lib/python3.8/site-packages/detectron2/engine/launch.py", line 84, in launch
main_func(*args)
File "src/apply_net.py", line 81, in main
outputs = predictor(input_im)
File "/root/autodl-tmp/mzy/Git/probdet/src/probabilistic_inference/inference_core.py", line 78, in call
results = self.post_processing_bayes_od(input_im)
File "/root/autodl-tmp/mzy/Git/probdet/src/probabilistic_inference/probabilistic_retinanet_predictor.py", line 400, in post_processing_bayes_od
outputs = self.retinanet_probabilistic_inference(input_im)
File "/root/autodl-tmp/mzy/Git/probdet/src/probabilistic_inference/probabilistic_retinanet_predictor.py", line 53, in retinanet_probabilistic_inference
n_fms = len(self.model.in_features)
File "/root/miniconda3/envs/probdet/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1695, in getattr
raise AttributeError(f"'{type(self).name}' object has no attribute '{name}'")
AttributeError: 'ProbabilisticRetinaNet' object has no attribute 'in_features'
asharakeh
