moviad stfpm

하.. 갈길이 멀다 ㅠㅠ

patchcore는 cudnn만 꺼주면 어떻게 되는데

stfpm은 소스가 문제라 .. -_-

~/src/moviad/main_scripts$ python3 main_stfpm.py --train --eval     --model_name mobilenet_v2     --categories bottle     --ad_layers 3 4 5     --boot_layer 2     --results_dirpath debug_outputs/metrics     --checkpoint_dir debug_outputs/checkpoints     --seeds 0 --epochs 3 --input_size 224 224     --device cuda:0 ERROR:root:micromind not found in current environment Traceback (most recent call last):   File "/home/falinux/src/moviad/main_scripts/main_stfpm.py", line 12, in <module>     from moviad.trainers.trainer_stfpm import train_param_grid_search ImportError: cannot import name 'train_param_grid_search' from 'moviad.trainers.trainer_stfpm' (/home/falinux/src/moviad/moviad/trainers/trainer_stfpm.py)

 

$ git log commit 5d547292f3e1e4402b91d2950b884be74a37900f (HEAD -> main, origin/main, origin/HEAD) Author: FrancescoBorsatti <francesco.borsatti.1@phd.unipd.it> Date:   Mon Apr 13 18:16:40 2026 +0200     update readme

 

괜히 verified가 붙은게 아닌듯. 저 버전으로는 내려가야 한다. merge 하다가 꼬인듯.

[링크 : https://github.com/AMCO-UniPD/moviad/blob/5bf4b63a89a860ab3d76679fffe35ee50225901d/moviad/trainers/trainer_stfpm.py] 없음

[링크 : https://github.com/AMCO-UniPD/moviad/blob/ec89d3e145615c9c1fbae69480f7da2eb4f9c606/moviad/trainers/trainer_stfpm.py] 있음(verified)

[링크 : https://github.com/AMCO-UniPD/moviad/blob/a209ec364aba6f5f33be010af948a578af4971ce/moviad/trainers/trainer_stfpm.py] 있음

 

+

아래 커밋으로 이동해버리고(!)

git checkout ec89d3e145615c9c1fbae69480f7da2eb4f9c606

 

실행하면 안된다!

정해진 경로에 넣으란거지? -_-+

~/src/moviad/main_scripts$ python3 main_stfpm.py --train --eval     --model_name mobilenet_v2     --categories bottle     --ad_layers 3 4 5     --boot_layer 2     --results_dirpath debug_outputs/metrics     --checkpoint_dir debug_outputs/checkpoints     --seeds 0 --epochs 3 --input_size 224 224     --device cuda:0 ERROR:root:micromind not found in current environment Training with params: {'dataset_path': '../../datasets/mvtec/', 'categories': ['bottle'], 'ad_layers': [[3, 4, 5]], 'epochs': [3], 'seeds': [0], 'batch_size': 64, 'backbone_model_name': 'mobilenet_v2', 'device': device(type='cuda', index=0), 'img_input_size': [224, 224], 'img_output_size': (224, 224), 'early_stopping': None, 'student_bootstrap_layer': [2], 'checkpoint_dir': 'debug_outputs/checkpoints', 'normalize_dataset': True, 'log_dirpath': None, 'contamination_ratio': None, 'test_dataset': False} TRAIN | cat: bottle, ad_layers: [3, 4, 5], epochs: 3, seed: 0, early_stopping: None, bootstrap: 2 Traceback (most recent call last):   File "/home/minimonk/src/moviad/main_scripts/main_stfpm.py", line 405, in <module>     raise e   File "/home/minimonk/src/moviad/main_scripts/main_stfpm.py", line 387, in <module>     main(args)   File "/home/minimonk/src/moviad/main_scripts/main_stfpm.py", line 86, in main     trained_models_filepaths = train_param_grid_search(params)   File "/home/minimonk/src/moviad/moviad/trainers/trainer_stfpm.py", line 321, in train_param_grid_search     log, snapshot_path = train_param_grid_step(   File "/home/minimonk/src/moviad/moviad/trainers/trainer_stfpm.py", line 209, in train_param_grid_step     train_dataset.load_dataset()   File "/home/minimonk/src/moviad/moviad/datasets/mvtec/mvtec_dataset.py", line 175, in load_dataset     raise RuntimeError(msg) RuntimeError: Found 0 images in ../../datasets/mvtec/bottle