'프로그램 사용' 카테고리의 글 목록 (7 Page)

'프로그램 사용'에 해당되는 글 2525건

2026.06.18 moviad stfpm
2026.06.17 ubuntu 26.04 + 3070 + tensorflow + python 3.14 + docker...
2026.06.10 gcc -MD -MF 의존성 확인
2026.06.10 gcc ld rpah=$ORIGIN 동적링크 경로
2026.06.01 STFPM 실행
2026.05.27 blener ll3m
2026.05.27 openscad
2026.05.27 딥러닝 학습 관련(epoch, loss)
2026.05.21 NAS - Neural Architecture Search
2026.05.21 MCUNet

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

저작자표시 (새창열림)

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

moviad ssd-mobilenetv2 on imx8mp (0)	2026.06.24
moviad / stftpm / 1080 ti / webcam feat claude (0)	2026.06.19
ubuntu 26.04 + 3070 + tensorflow + python 3.14 + docker... (0)	2026.06.17
STFPM 실행 (0)	2026.06.01
딥러닝 학습 관련(epoch, loss) (0)	2026.05.27

Posted by 구차니

ubuntu 26.04 + 3070 + tensorflow + python 3.14 + docker...

개발환경 구축하려고 이것저것 보고 있는데

구형 모델을 쓰려고 하다보니(ssd mobilenet v3) 영 환경 잡는것 자체가 쉽지 않아 보인다.

[링크 : https://www.tensorflow.org/install/docker?hl=ko] 텐서플로우 docker

[링크 : https://github.com/NVIDIA/nvidia-docker] 요건 deprecated

[링크 : https://github.com/NVIDIA/nvidia-container-toolkit] 요걸 쓰라고

tensorflow 2.21.0

2026.03.07 release

지원 파이썬 버전 3.13 최대.

[링크 : https://pypi.org/project/tensorflow/]

tensorflow-gpu 2.12.0

2023.01.25 release

[링크 : https://pypi.org/project/tensorflow-gpu/]

그 와중에 우분투 26.04 의 파이썬 버전이 3.14 대이고

$ python3 --version
Python 3.14.4

텐서플로우는 지원버전이 없다고 배째고

$ pip3 install tensorflow
ERROR: Could not find a version that satisfies the requirement tensorflow (from versions: none)
ERROR: No matching distribution found for tensorflow

난 어짜피 gpu 가속받을거니까~ 하고 실행하는데 에러

$ pip3 install tensorflow-gpu
Collecting tensorflow-gpu
  Using cached tensorflow-gpu-2.12.0.tar.gz (2.6 kB)
  Installing build dependencies ... done
  Getting requirements to build wheel ... error
  error: subprocess-exited-with-error

  × Getting requirements to build wheel did not run successfully.
  │ exit code: 1
  ╰─> [77 lines of output]
      Traceback (most recent call last):
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/_vendor/packaging/requirements.py", line 36, in __init__
          parsed = _parse_requirement(requirement_string)
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/_vendor/packaging/_parser.py", line 71, in parse_requirement
          return _parse_requirement(Tokenizer(source, rules=DEFAULT_RULES))
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/_vendor/packaging/_parser.py", line 89, in _parse_requirement
          url, specifier, marker = _parse_requirement_details(tokenizer)
                                   ~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/_vendor/packaging/_parser.py", line 135, in _parse_requirement_details
          marker = _parse_requirement_marker(
              tokenizer,
          ...<5 lines>...
              ),
          )
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/_vendor/packaging/_parser.py", line 156, in _parse_requirement_marker
          tokenizer.raise_syntax_error(
          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
              f"Expected {expected} or end",
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
              span_start=span_start,
              ^^^^^^^^^^^^^^^^^^^^^^
              span_end=None,
              ^^^^^^^^^^^^^^
          )
          ^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/_vendor/packaging/_tokenizer.py", line 166, in raise_syntax_error
          raise ParserSyntaxError(
          ...<3 lines>...
          )
      packaging._tokenizer.ParserSyntaxError: Expected semicolon (after name with no version specifier) or end
          python_version>"3.7"
                        ^

      The above exception was the direct cause of the following exception:

      Traceback (most recent call last):
        File "/home/minimonk/venv/lib/python3.14/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 389, in <module>
          main()
          ~~~~^^
        File "/home/minimonk/venv/lib/python3.14/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 373, in main
          json_out["return_val"] = hook(**hook_input["kwargs"])
                                   ~~~~^^^^^^^^^^^^^^^^^^^^^^^^
        File "/home/minimonk/venv/lib/python3.14/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 143, in get_requires_for_build_wheel
          return hook(config_settings)
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/build_meta.py", line 333, in get_requires_for_build_wheel
          return self._get_build_requires(config_settings, requirements=[])
                 ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/build_meta.py", line 301, in _get_build_requires
          self.run_setup()
          ~~~~~~~~~~~~~~^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/build_meta.py", line 520, in run_setup
          super().run_setup(setup_script=setup_script)
          ~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/build_meta.py", line 317, in run_setup
          exec(code, locals())
          ~~~~^^^^^^^^^^^^^^^^
        File "<string>", line 40, in <module>
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/__init__.py", line 114, in setup
          _install_setup_requires(attrs)
          ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/__init__.py", line 85, in _install_setup_requires
          dist.parse_config_files(ignore_option_errors=True)
          ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/dist.py", line 764, in parse_config_files
          self._finalize_requires()
          ~~~~~~~~~~~~~~~~~~~~~~~^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/dist.py", line 382, in _finalize_requires
          self._normalize_requires()
          ~~~~~~~~~~~~~~~~~~~~~~~~^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/dist.py", line 400, in _normalize_requires
          self.install_requires = list_(map(str, _reqs.parse(install_requires)))
                                  ~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "/tmp/pip-build-env-kb1b92jj/overlay/lib/python3.14/site-packages/setuptools/_vendor/packaging/requirements.py", line 38, in __init__
          raise InvalidRequirement(str(e)) from e
      packaging.requirements.InvalidRequirement: Expected semicolon (after name with no version specifier) or end
          python_version>"3.7"
                        ^
      [end of output]

  note: This error originates from a subprocess, and is likely not a problem with pip.
error: subprocess-exited-with-error

× Getting requirements to build wheel did not run successfully.
│ exit code: 1
╰─> See above for output.

note: This error originates from a subprocess, and is likely not a problem with pip.

tensorflow 도 이제 인기가 없어졌다고 해야하나.. 지는 해가 되어버렸나 보다.

[링크 : https://github.com/tensorflow/tensorflow/issues/102890]

짧고 명확하다... deprecation. 후..

TensorFlow Object Detection API
TensorFlow 2.2 TensorFlow 1.15 Python 3.6

Deprecation

[링크 : https://github.com/tensorflow/models/blob/master/research/object_detection/README.md]

keras도 희망이 없고, pytorch도 좋은소리 안나오고 어떻게 해야하나..

by GPT

by claude

저작자표시 (새창열림)

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

moviad / stftpm / 1080 ti / webcam feat claude (0)	2026.06.19
moviad stfpm (0)	2026.06.18
STFPM 실행 (0)	2026.06.01
딥러닝 학습 관련(epoch, loss) (0)	2026.05.27
NAS - Neural Architecture Search (0)	2026.05.21

Posted by 구차니

gcc -MD -MF 의존성 확인

msvc의 md 옵션과는 의미가 다르고(claude에게 낚임!)

그냥 빌드시 디버깅용 플래그라고 해야하나..?

동적 링크시 어떤 파일을 참조하는지 .d 파일에 주로 쓰도록 하는 것 같다.

[링크 : https://dmake.tistory.com/26]

[링크 : https://m.blog.naver.com/wonmylover/220771036728]

-M
Instead of outputting the result of preprocessing, output a rule suitable for make describing the dependencies of the main source file. The preprocessor outputs one make rule containing the object file name for that source file, a colon, and the names of all the included files, including those coming from -include or -imacros command line options.

Unless specified explicitly (with -MT or -MQ), the object file name consists of the name of the source file with any suffix replaced with object file suffix and with any leading directory parts removed. If there are many included files then the rule is split into several lines using \-newline. The rule has no commands.
This option does not suppress the preprocessor's debug output, such as -dM. To avoid mixing such debug output with the dependency rules you should explicitly specify the dependency output file with -MF, or use an environment variable like DEPENDENCIES_OUTPUT . Debug output will still be sent to the regular output stream as normal.

Passing -M to the driver implies -E, and suppresses warnings with an implicit -w.

-MM
Like -M but do not mention header files that are found in system header directories, nor header files that are included, directly or indirectly, from such a header.

This implies that the choice of angle brackets or double quotes in an #include directive does not in itself determine whether that header will appear in -MM dependency output. This is a slight change in semantics from GCC versions 3.0 and earlier.
-MF file
When used with -M or -MM, specifies a file to write the dependencies to. If no -MF switch is given the preprocessor sends the rules to the same place it would have sent preprocessed output.
When used with the driver options -MD or -MMD, -MF overrides the default dependency output file.

-MG
In conjunction with an option such as -M requesting dependency generation, -MG assumes missing header files are generated files and adds them to the dependency list without raising an error. The dependency filename is taken directly from the "#include" directive without prepending any path. -MG also suppresses preprocessed output, as a missing header file renders this useless.

This feature is used in automatic updating of makefiles.
-MP
This option instructs CPP to add a phony target for each dependency other than the main file, causing each to depend on nothing. These dummy rules work around errors make gives if you remove header files without updating the Makefile to match.

This is typical output:
test.o: test.c test.h

test.h:
-MT target
Change the target of the rule emitted by dependency generation. By default CPP takes the name of the main input file, deletes any directory components and any file suffix such as .c, and appends the platform's usual object suffix. The result is the target.
An -MT option will set the target to be exactly the string you specify. If you want multiple targets, you can specify them as a single argument to -MT, or use multiple -MT options.

For example, -MT '$(objpfx)foo.o' might give

$(objpfx)foo.o: foo.c
-MQ target
Same as -MT, but it quotes any characters which are special to Make. -MQ '$(objpfx)foo.o' gives
$$(objpfx)foo.o: foo.c
The default target is automatically quoted, as if it were given with -MQ.
-MD
-MD is equivalent to -M -MF file, except that -E is not implied. The driver determines file based on whether an -o option is given. If it is, the driver uses its argument but with a suffix of .d, otherwise it takes the name of the input file, removes any directory components and suffix, and applies a .d suffix.

If -MD is used in conjunction with -E, any -o switch is understood to specify the dependency output file, but if used without -E, each -o is understood to specify a target object file.
Since -E is not implied, -MD can be used to generate a dependency output file as a side-effect of the compilation process.

-MMD
Like -MD except mention only user header files, not system header files.

[링크 : https://linux.die.net/man/1/gcc]

저작자표시 (새창열림)

'프로그램 사용 > gcc' 카테고리의 다른 글

gcc ld rpah=$ORIGIN 동적링크 경로 (0)	2026.06.10
링커 스크립트를 이용한 변수 값 할당 (0)	2026.03.13
gcc -D 를 이용하여 명령줄에서 define 하기 (0)	2026.03.12
gcc __cplusplus 선언 (0)	2026.02.18
gcc __attribute__((weak)) 테스트 (0)	2026.01.29

Posted by 구차니

gcc ld rpah=$ORIGIN 동적링크 경로

리눅스에서 gcc로 빌드하면 시스템 절대 경로라고 해야하나.

아래의 경우, /lib/x86_64-linux-gnu/ 의 경로에 있는 so들을 보도록 되어있는데 (LD_LIBRARY_PATH)

이걸 빌드 시에 위치 기준 상대 경로를 보게 하는 옵션 인듯.

so 파일과 실행파일을 같이 배포할때 쓰이려나?

$ ldd untitled
linux-vdso.so.1 (0x00007fff3f7a7000)
libQt5Widgets.so.5 => /lib/x86_64-linux-gnu/libQt5Widgets.so.5 (0x00007a58ffe00000)
libQt5Gui.so.5 => /lib/x86_64-linux-gnu/libQt5Gui.so.5 (0x00007a58ff600000)
libQt5Core.so.5 => /lib/x86_64-linux-gnu/libQt5Core.so.5 (0x00007a58ff000000)
libstdc++.so.6 => /lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007a58fec00000)
libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007a5900f63000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007a58fe800000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007a5900519000)
libGL.so.1 => /lib/x86_64-linux-gnu/libGL.so.1 (0x00007a58ffd79000)
libpng16.so.16 => /lib/x86_64-linux-gnu/libpng16.so.16 (0x00007a59004de000)
libz.so.1 => /lib/x86_64-linux-gnu/libz.so.1 (0x00007a5900f45000)
libharfbuzz.so.0 => /lib/x86_64-linux-gnu/libharfbuzz.so.0 (0x00007a58fef31000)
libmd4c.so.0 => /lib/x86_64-linux-gnu/libmd4c.so.0 (0x00007a59004cc000)
libdouble-conversion.so.3 => /lib/x86_64-linux-gnu/libdouble-conversion.so.3 (0x00007a58ffd64000)
libicui18n.so.70 => /lib/x86_64-linux-gnu/libicui18n.so.70 (0x00007a58fe400000)
libicuuc.so.70 => /lib/x86_64-linux-gnu/libicuuc.so.70 (0x00007a58fe205000)
libpcre2-16.so.0 => /lib/x86_64-linux-gnu/libpcre2-16.so.0 (0x00007a58ff576000)
libzstd.so.1 => /lib/x86_64-linux-gnu/libzstd.so.1 (0x00007a58fee62000)
libglib-2.0.so.0 => /lib/x86_64-linux-gnu/libglib-2.0.so.0 (0x00007a58feac5000)
/lib64/ld-linux-x86-64.so.2 (0x00007a5900faa000)
libGLdispatch.so.0 => /lib/x86_64-linux-gnu/libGLdispatch.so.0 (0x00007a58fe747000)
libGLX.so.0 => /lib/x86_64-linux-gnu/libGLX.so.0 (0x00007a58ffd30000)
libfreetype.so.6 => /lib/x86_64-linux-gnu/libfreetype.so.6 (0x00007a58fe13d000)
libgraphite2.so.3 => /lib/x86_64-linux-gnu/libgraphite2.so.3 (0x00007a58ffd09000)
libicudata.so.70 => /lib/x86_64-linux-gnu/libicudata.so.70 (0x00007a58fc400000)
libpcre.so.3 => /lib/x86_64-linux-gnu/libpcre.so.3 (0x00007a58fea4f000)
libX11.so.6 => /lib/x86_64-linux-gnu/libX11.so.6 (0x00007a58fc2c0000)
libbrotlidec.so.1 => /lib/x86_64-linux-gnu/libbrotlidec.so.1 (0x00007a58ffcfb000)
libxcb.so.1 => /lib/x86_64-linux-gnu/libxcb.so.1 (0x00007a58fee38000)
libbrotlicommon.so.1 => /lib/x86_64-linux-gnu/libbrotlicommon.so.1 (0x00007a58fea2c000)
libXau.so.6 => /lib/x86_64-linux-gnu/libXau.so.6 (0x00007a5900f37000)
libXdmcp.so.6 => /lib/x86_64-linux-gnu/libXdmcp.so.6 (0x00007a59004c4000)
libbsd.so.0 => /lib/x86_64-linux-gnu/libbsd.so.0 (0x00007a58ffce3000)
libmd.so.0 => /lib/x86_64-linux-gnu/libmd.so.0 (0x00007a58ff569000)

[링크 : https://velog.io/@wjddms206/RPATH-한번에-이해하기]

[링크 : https://stackoverflow.com/questions/6324131/rpath-origin-not-having-desired-effect]

[링크 : https://stackoverflow.com/questions/38058041/correct-usage-of-rpath-relative-vs-absolute]

저작자표시 (새창열림)

'프로그램 사용 > gcc' 카테고리의 다른 글

gcc -MD -MF 의존성 확인 (0)	2026.06.10
링커 스크립트를 이용한 변수 값 할당 (0)	2026.03.13
gcc -D 를 이용하여 명령줄에서 define 하기 (0)	2026.03.12
gcc __cplusplus 선언 (0)	2026.02.18
gcc __attribute__((weak)) 테스트 (0)	2026.01.29

Posted by 구차니

STFPM 실행

언제나 그렇듯(!) 1080은 돌리기 힘든 시대구만.. 저번에 찾아둔 걸로 버전 맞추고

pip3 uninstall -y torch torchvision torchaudio xformers
pip3 install torch==2.6.0 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
pip3 install xformers==0.0.29.post2

-mvtec-ad 는 dataset 다운로드 받아 압축푼 최상위 디렉토리를 지정해주면 된다.

~/src/STFPM/mvtec-ad$ tree -d -L 2
.
├── bottle
│   ├── ground_truth
│   ├── test
│   └── train
├── cable
│   ├── ground_truth
│   ├── test
│   └── train
├── capsule
│   ├── ground_truth
│   ├── test
│   └── train
├── carpet
│   ├── ground_truth
│   ├── test
│   └── train
├── grid
│   ├── ground_truth
│   ├── test
│   └── train
├── hazelnut
│   ├── ground_truth
│   ├── test
│   └── train
├── leather
│   ├── ground_truth
│   ├── test
│   └── train
├── metal_nut
│   ├── ground_truth
│   ├── test
│   └── train
├── pill
│   ├── ground_truth
│   ├── test
│   └── train
├── screw
│   ├── ground_truth
│   ├── test
│   └── train
├── tile
│   ├── ground_truth
│   ├── test
│   └── train
├── toothbrush
│   ├── ground_truth
│   ├── test
│   └── train
├── transistor
│   ├── ground_truth
│   ├── test
│   └── train
├── wood
│   ├── ground_truth
│   ├── test
│   └── train
└── zipper
    ├── ground_truth
    ├── test
    └── train

$ python3 main.py train --mvtec-ad mvtec-ad --category carpet --epochs 200
...
[198/200] loss: 0.145916
[198/200] loss: 0.147569
[198/200] loss: 0.147157
[198/200] loss: 0.146806
[198/200] loss: 0.147341
[198/200] loss: 0.145462
[198/200] loss: 0.144424
Valid Loss: 0.0001142
[199/200] loss: 0.146722
[199/200] loss: 0.146246
[199/200] loss: 0.145853
[199/200] loss: 0.145966
[199/200] loss: 0.145793
[199/200] loss: 0.145863
[199/200] loss: 0.154857
Valid Loss: 0.0001460

real 9m35.951s
user 31m24.301s
sys 1m16.579s

loss 만 보고 학습을 하게 했네.

그래도 200 epoch 인데 내부는 어떤지 모르겠지만 10분 컷이면 양호하네

cpu 로 1개 코어만 사용해서 하는데 제법 걸린다. 이미지로 판별하고 출력할 수 없나?

$ time python3 main.py test --mvtec-ad mvtec-ad --category carpet --checkpoint snapshots/carpet/best.pth.tar
load snapshots/carpet/best.pth.tar
Catergory: carpet Pixel-AUC: 0.990288 Image-AUC: 0.955859 PRO: 0.959637

real 5m39.944s
user 5m46.592s
sys 0m1.064s

tar 라면서 왜 zip 이래?!?

~/src/STFPM/snapshots/carpet$ file *
best.pth.tar: Zip archive data, at least v0.0 to extract, compression method=store

~/src/STFPM/snapshots/carpet$ ll -h
total 43M
drwxrwxr-x 2 falinux falinux 4.0K  6월  1 11:57 ./
drwxrwxr-x 3 falinux falinux 4.0K  6월  1 11:14 ../
-rw-rw-r-- 1 falinux falinux  43M  6월  1 11:52 best.pth.tar

[링크 : https://github.com/gdwang08/STFPM]

저작자표시 (새창열림)

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

moviad stfpm (0)	2026.06.18
ubuntu 26.04 + 3070 + tensorflow + python 3.14 + docker... (0)	2026.06.17
딥러닝 학습 관련(epoch, loss) (0)	2026.05.27
NAS - Neural Architecture Search (0)	2026.05.21
MCUNet (0)	2026.05.21

Posted by 구차니

blener ll3m

먼가 발견하긴 했는데, 이미 서비스 종료 -_ㅠ

Notice: The model used in the paper, Claude Sonnet 3.7, has been retired. As a result, we have discontinued the LL3M server.

[링크 : https://github.com/threedle/ll3m]

[링크 : https://threedle.github.io/ll3m/]

저작자표시 (새창열림)

'프로그램 사용 > Blender' 카테고리의 다른 글

blender bone (armature) (0)	2026.06.24
3d model 다운로드 / blender import (0)	2026.06.24
gemini + blender (0)	2025.09.01
blender로 stl 구멍 메우기 (0)	2025.08.20
blender 휠 에뮬레이트 하기 (0)	2025.05.18

Posted by 구차니

openscad

freecad나 blender cad 확장으로는 이래저래 힘들고

python 통해서 제어하면 이상하게 나와서 제어가 유리한 무언가를 찾는 중

[링크 : https://github.com/fogleman/sdf]

[링크 : https://news.hada.io/topic?id=25230]

저작자표시 (새창열림)

'프로그램 사용 > openSCAD' 카테고리의 다른 글

openSCAD + claude.ai = 파라메트릭 모델링 (0)	2026.07.13
openSCAD MCP (0)	2026.06.20
pythonSCAD (0)	2026.06.20
openSCAD cheatsheet (0)	2026.06.20
openSCAD 설치 (0)	2026.06.20

Posted by 구차니

딥러닝 학습 관련(epoch, loss)

-----

metric 함수. 얘도 함수였구나.

아무튼 metric은 격하게 오락가락 하고

[링크 : https://modulabs.co.kr/blog/loss-versus-accuracy]

loss 가 점진적 감소 -> 잘 학습

loss 급격히 감소 후 유지 -> 정상

loss 일정 시점 이후 증가 -> over fitting

[링크 : https://m.blog.naver.com/datapreprep/223748771141]

근데.. underfitting은 먼지 감이 안오네.

When to change the number of epochs (training cycles)

Selecting the appropriate number of epochs is a balance between underfitting and overfitting.

Underfitting: One of the most straightforward indicators of underfitting is if the model performs poorly on the training data. This can be observed in Edge Impulse Studio through metrics such as accuracy, or loss, depending on the type of problem (classification or regression). If these metrics indicate poor performance, it suggests that the model has not learned the patterns of the data well. In that case, increasing the number of epochs can improve your model performance. Please note that other solutions exist such as increasing your neural network architecture complexity, changing the preprocessing technique or reducing regularization.

Overfitting: Detecting overfitting involves recognizing when the model has learned too much from the training data, including its noise and outliers, to the detriment of its performance on new, unseen data. Overfitting is characterized by the model performing exceptionally well on the training data but poorly on the validation or test data. Evaluating overfitting can be achieved by comparing the performance of the model between the training set and the validation set during training. When the performance on the validation set starts to degrade, it might indicate that the model is beginning to overfit the training data. In that case, decreasing the number of epochs can improve your model performance. As with underfitting, other solutions exist to reduce overfitting such as increasing the number of training data, adding regularization techniques to add penalties on large weights, adding dropout layers, simplifying the model architecture and even using early stopping.

[링크 : https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/neural-networks/epochs]

overfitting 중인 loss 값 이라는데,

학습셋이 너무 적을 경우 overfitting이 빠르게 발생하는 것으로 보인다.

그나저나 저 130 miliion parameter는 ssd300 에서 어떻게 산출된걸까?

What you are experiencing is called overfitting and it happens because of your very small dataset.
All the model cares about is performance on the training dataset, so given the opportunity, it will simply attempt to memorize it. This is what happens in you case, you feed a model which contains over 130 Million parameters less than 319 images.
So regarding your questions:

The loss function shows a clear case of overfitting.
On general, it is okay to use a trained model, especially when you only have a small dataset, but in your case, the dataset is too small for any deep-learning model. When I say small dataset, I mean 10k images, not several hundreds.
You should not train for longer time, once the validation loss stops improving, it is a clear sign to stop. There is even a training technique named "early stopping" which is designed to stop training once the validation loss stops to drop.

You have to understand that currently, your dataset of 300 images, is irrelevant to the world of deep-learning. So if you still want to use it for object detection, you need to revert to more classic computer-vision techniques like using HOG or SIFT features, or even manually engineering the features for your special case.

[링크 : https://datascience.stackexchange.com/questions/46456/issues-with-training-ssd-on-own-dataset]

The SSD300 has 35.6 million parameters, while the SSDlite320 has about a tenth of them, i.e., 3.4 million.

[링크 : https://www.mdpi.com/2673-6470/3/3/12]

저작자표시 (새창열림)

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

ubuntu 26.04 + 3070 + tensorflow + python 3.14 + docker... (0)	2026.06.17
STFPM 실행 (0)	2026.06.01
NAS - Neural Architecture Search (0)	2026.05.21
MCUNet (0)	2026.05.21
STFPM - Student-Teacher Feature Pyramid Matching (0)	2026.05.20

Posted by 구차니

NAS - Neural Architecture Search

MCUNet 보다 보니 NAS 라는게 보여서 찾아봄

Search NN model on an existing library e.g., ProxylessNAS, MnasNet

[링크 : https://hanlab18.mit.edu/projects/tinyml/mcunet/assets/MCUNet-slides.pdf]

Neural Architecture Search

[링크 : https://doing-ai.tistory.com/entry/Neural-Architecture-Search-NAS-Hardware-Aware-NAS-ProxylessNAS]

EfficientNAS 이후에 나온 더욱 효율적인 NAS를 위한 논문, ProxylessNAS다. TuNAS의 근간이 된 논문이기도 하며, 구체적으로 알아보자.

[링크 : https://seokdonge.tistory.com/27] ProxylessNAS

[링크 : https://ech97.tistory.com/entry/MnasNet] MnasNet

TuNAS

[링크 : https://seokdonge.tistory.com/26]

[링크 : https://openaccess.thecvf.com/content_CVPR_2020/papers/Bender_Can_Weight_Sharing_Outperform_Random_Architecture_Search_An_Investigation_With_CVPR_2020_paper.pdf]

[링크 : https://github.com/google-research/google-research/tree/master/tunas]

저작자표시 (새창열림)

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

STFPM 실행 (0)	2026.06.01
딥러닝 학습 관련(epoch, loss) (0)	2026.05.27
MCUNet (0)	2026.05.21
STFPM - Student-Teacher Feature Pyramid Matching (0)	2026.05.20
EfficientAD (0)	2026.05.20

Posted by 구차니

MCUNet

STFPM -> PaSTe -> MCUNet 으로 찾아옴

[링크 : https://github.com/AMCO-UniPD/PaSTe]

MCU 에서 돌릴만큼 경량 네트워크

[링크 : https://cowkim-svd.tistory.com/7]

[링크 : https://github.com/mit-han-lab/mcunet]

[링크 : https://www.youtube.com/watch?v=YvioBgtec4U]

[링크 : https://www.youtube.com/watch?v=0pUFZYdoMY8]

저작자표시 (새창열림)

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

딥러닝 학습 관련(epoch, loss) (0)	2026.05.27
NAS - Neural Architecture Search (0)	2026.05.21
STFPM - Student-Teacher Feature Pyramid Matching (0)	2026.05.20
EfficientAD (0)	2026.05.20
patchcore (0)	2026.05.20

Posted by 구차니

구차니의 잡동사니 모음

'프로그램 사용'에 해당되는 글 2525건

moviad stfpm

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

ubuntu 26.04 + 3070 + tensorflow + python 3.14 + docker...

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

gcc -MD -MF 의존성 확인

'프로그램 사용 > gcc' 카테고리의 다른 글

gcc ld rpah=$ORIGIN 동적링크 경로

'프로그램 사용 > gcc' 카테고리의 다른 글

STFPM 실행

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

blener ll3m

'프로그램 사용 > Blender' 카테고리의 다른 글

openscad

'프로그램 사용 > openSCAD' 카테고리의 다른 글

딥러닝 학습 관련(epoch, loss)

When to change the number of epochs (training cycles)

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

NAS - Neural Architecture Search

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

MCUNet

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

카테고리

공지사항

태그목록

최근에 올라온 글

최근에 달린 댓글

티스토리툴바