Create mode

This notebook depicts the create mode. It is used to generate the underlying data for a task. Task creation is the very first step that needs to be taken in order to use it for processing. The mml-tasks plugin provides the creators to generate a variety of tasks. But we stick to the mml_fake_task implemented in mml-core for simplicity.

!mml create task_list=[mml_fake_task]

[2025-05-02 22:13:11,242][mml][INFO] - Started MML 1.0.4 on Python 3.10.17 with mode CREATE.
[2025-05-02 22:13:11,242][mml][INFO] - Plugins loaded: ['mml-tasks']
[2025-05-02 22:13:11,599][mml.core.scripts.schedulers.create_scheduler][INFO] - Skipping creation of task mml_fake_task because there already seems to be a RAW version of that.
[2025-05-02 22:13:11,600][mml][INFO] - MML init time was 0.0h 0.0m  0.36s.
[2025-05-02 22:13:11,602][mml.core.scripts.schedulers.create_scheduler][INFO] - Starting task creation!
[2025-05-02 22:13:11,602][mml.core.data_loading.file_manager][INFO] - A total of 0 paths have been created during this run.
[2025-05-02 22:13:11,603][mml.core.scripts.schedulers.base_scheduler][INFO] - Successfully finished all experiments!
[2025-05-02 22:13:11,603][mml][INFO] - MML run time was 0.0h 0.0m  0.00s.

After creation we can see some information on the task with info mode.

!mml info task_list=[mml_fake_task]

[2025-05-02 22:13:21,893][mml][INFO] - Started MML 1.0.4 on Python 3.10.17 with mode INFO.
[2025-05-02 22:13:21,893][mml][INFO] - Plugins loaded: ['mml-tasks']
[2025-05-02 22:13:22,231][mml.core.scripts.schedulers.info_scheduler][INFO] - Was given no study name to search for, so showing all studies with project prefix.
[2025-05-02 22:13:22,234][mml][INFO] - MML init time was 0.0h 0.0m  0.34s.
[2025-05-02 22:13:22,236][mml.core.scripts.schedulers.base_scheduler][INFO] - Preparing experiment ...
[2025-05-02 22:13:22,237][mml.core.scripts.schedulers.base_scheduler][INFO] - Starting experiment!
[2025-05-02 22:13:22,238][mml.core.scripts.schedulers.info_scheduler][INFO] - Starting info on task mml_fake_task
[2025-05-02 22:13:22,238][mml.core.scripts.schedulers.info_scheduler][INFO] - Task name: mml_fake_task
Task type: classification
Num classes: 10
Means: RGBInfo(r=0.49829596281051636, g=0.49836331605911255, b=0.498288631439209)
Stds: RGBInfo(r=0.1517328917980194, g=0.15169444680213928, b=0.1517714262008667)
Sizes: Sizes(min_height=256, max_height=256, min_width=256, max_width=256)
Class occ: {'H': 89, 'E': 93, 'A': 93, 'J': 96, 'C': 103, 'F': 93, 'I': 127, 'G': 90, 'D': 121, 'B': 95}
Preprocessed: default
Task keywords: ['artificial']
paths: {}
models: []
[2025-05-02 22:13:22,250][mml.core.scripts.schedulers.info_scheduler][INFO] - Num samples (full train set): 1000
[2025-05-02 22:13:22,250][mml.core.scripts.schedulers.info_scheduler][INFO] - Default validation class occurrences are: {2: 21, 7: 18, 9: 19, 0: 19, 3: 24, 8: 25, 4: 19, 1: 19, 5: 19, 6: 18}
[2025-05-02 22:13:22,251][mml.core.scripts.schedulers.info_scheduler][INFO] - Finished info on task mml_fake_task
[2025-05-02 22:13:22,252][mml.core.scripts.schedulers.info_scheduler][INFO] - Starting plotting sample grid of all tasks.
Loading samples:   0%|                                    | 0/1 [00:00<?, ?it/s][2025-05-02 22:13:22,254][py.warnings][WARNING] - /home/scholzpa/Documents/development/github/mml/src/mml/core/data_loading/lightning_datamodule.py:319: Deactivated normalization for task mml_fake_task.

Loading samples: 100%|████████████████████████████| 1/1 [00:00<00:00, 35.90it/s]
[2025-05-02 22:13:22,294][mml.core.scripts.schedulers.info_scheduler][INFO] - Finished plotting sample grid. Can be found at /home/scholzpa/Documents/exp/mml_results/default/PLOTS/sample_grid/grid_0112.png.
[2025-05-02 22:13:22,305][mml.core.scripts.schedulers.info_scheduler][INFO] - Finished plotting individual samples for each task.
[2025-05-02 22:13:22,307][mml.core.scripts.schedulers.info_scheduler][INFO] - Model info shows only loaded models (not all existing!). Use reuse.models=... to select a project to load models from.
+------+---------+------+-------------+-----------------+---------+--------+
| task | created | fold | performance | training (secs) | params? | preds? |
+------+---------+------+-------------+-----------------+---------+--------+
+------+---------+------+-------------+-----------------+---------+--------+
[2025-05-02 22:13:22,308][mml.core.scripts.schedulers.info_scheduler][INFO] - No models found for 1 tasks (['mml_fake_task']).
[2025-05-02 22:13:22,309][mml.core.scripts.schedulers.info_scheduler][INFO] - Total number of all samples: 1000.
[2025-05-02 22:13:22,309][mml.core.data_loading.file_manager][INFO] - A total of 2 paths have been created during this run.
[2025-05-02 22:13:22,309][mml.core.scripts.schedulers.base_scheduler][INFO] - Successfully finished all experiments!
[2025-05-02 22:13:22,309][mml][INFO] - MML run time was 0.0h 0.0m  0.08s.