Potential issue in patch map display
taijizhao opened this issue · 3 comments
First congratulations for the great project!
I tried to run the run.sh file with the --visualization-flag on, the resulting patch maps were always aligned to the left border of the image and do not correspond to the activated regions in the heat maps.
I haven't figured out if this is a problem only in the visualization or it also affects the prediction accuracy.
Thank you very much!
Moreover, when I use MODEL_INDEX='4' to run the inference, it returns the following error:
Stage 1: Crop Mammograms Error: the directory to save cropped images already exists. Stage 2: Extract Centers Stage 3: Run Classifier 0%| | 0/4 [00:00<?, ?it/s] Traceback (most recent call last): File "src/scripts/run_model.py", line 302, in <module> main() File "src/scripts/run_model.py", line 292, in main start_experiment( File "src/scripts/run_model.py", line 255, in start_experiment output_df = run_single_model(single_model_path, data_path, parameters, turn_on_visualization) File "src/scripts/run_model.py", line 220, in run_single_model output_df = run_model(model, exam_list, parameters, turn_on_visualization) File "src/scripts/run_model.py", line 183, in run_model output = model(tensor_batch) File "/home/zhao/anaconda3/envs/pytorch/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl return forward_call(*input, **kwargs) File "/media/zhao/HD1/work_zhao/GMIC-master/src/modeling/gmic.py", line 126, in forward self.patch_locations = self._convert_crop_position(small_x_locations, self.cam_size, x_original) File "/media/zhao/HD1/work_zhao/GMIC-master/src/modeling/gmic.py", line 80, in _convert_crop_position assert np.min(top_k_prop_y) >= 0.0, "top_k_prop_y <= 0.0" AssertionError: top_k_prop_y <= 0.0
Hi @taijizhao ,
I am glad this repo is helpful to you. The cause is probably the torch version. This script was developed under torch 1.1.0. Please see this post for more info: #13
Best,
Hi @taijizhao ,
I am glad this repo is helpful to you. The cause is probably the torch version. This script was developed under torch 1.1.0. Please see this post for more info: #13
Best,
Thank you very much!
I should have checked previous posts more carefully.