Shawn’s Blog
  • 🗂️ Collections
    • 🖥️ Slides Gallery
    • 💻 LeetCode Notes
    • 🧑‍🍳️ Cooking Ideas
    • 🍱 Cookbook
    • 💬 Language Learning
    • 🎼 Songbook
  • ⚙️ Projects
    • ⚛ Continual Learning Arena
  • 📄 Papers
    • AdaHAT
    • FG-AdaHAT
    • AmnesiacHAT
  • 🎓 CV
    • CV (English)
    • CV (Mandarin)
  • About
  1. Components
  2. Other Configs
  • Welcome to CLArena
  • Getting Started
  • Configure Pipelines
  • Continual Learning (CL)
    • CL Main Experiment
    • Save and Evaluate Model
    • Full Experiment
    • Output Results
  • Continual Unlearning (CUL)
    • CUL Main Experiment
    • Full Experiment
    • Output Results
  • Multi-Task Learning (MTL)
    • MTL Experiment
    • Save and Evaluate Model
    • Output Results
  • Single-Task Learning (STL)
    • STL Experiment
    • Save and Evaluate Model
    • Output Results
  • Components
    • CL Dataset
    • MTL Dataset
    • STL Dataset
    • CL Algorithm
    • CUL Algorithm
    • MTL Algorithm
    • STL Algorithm
    • Backbone Network
    • Optimizer
    • Learning Rate Scheduler
    • Trainer
    • Metrics
    • Lightning Loggers
    • Callbacks
    • Other Configs
  • Custom Implementation
    • CL Dataset
    • MTL Dataset
    • STL Dataset
    • CL Algorithm
    • CUL Algorithm
    • MTL Algorithm
    • STL Algorithm
    • Backbone Network
    • Callback
  • API Reference
  • FAQs

On this page

  • Configs for Hydra
  • Config Tree
  • Metadata
  1. Components
  2. Other Configs

Other Configs

Modified

October 9, 2025

Other configs are less related to the experiment. They include:

  • Configuration for Hydra;
  • Miscellaneous configs, such as config tree printing.

They are sub-configs under the index config of:

  • Continual learning main experiment and evaluation
  • Continual learning full experiment, the reference experiments and evaluation
  • Continual unlearning main experiment and evaluation
  • Continual unlearning full experiment, the reference experiments and evaluation
  • Multi-task learning experiment and evaluation
  • Single-task learning experiment and evaluation

To configure a custom one, create a YAML file in the hydra/ or misc/ folder. We do not provide examples here; keep the defaults from the example configs unless you need to customize them.

Configs for Hydra

Here is the default Hydra config in the example configs:

example_configs/hydra/default.yaml
# configs for Hydra itself
# https://hydra.cc/docs/configure_hydra/intro/

# Hydra is automatically configured with defaults. These are overrides.

# enable hydra-colorlog
defaults:
  - override job_logging: colorlog
  - override hydra_logging: colorlog

job_logging:
  root:
    level: INFO # set pylogger level
  handlers:
    file:
      filename: ${hydra.runtime.output_dir}/console.log

run:
  dir: ${output_dir}

These configs are important for the Hydra system to work properly. Customizable fields include:

  • job_logging.root.level: the logging level for the entire experiment: DEBUG, INFO, WARNING, ERROR, CRITICAL. There are many loggings with these different levels in the codes, you can set the level to filter out some of them. The loggings will be both printed in the console and a .log file in the output folder (see Output Results sections of each paradigm).
  • handlers.file.filename: the filename for the output file. It is set to ${hydra.runtime.output_dir}/console.log by default.

Config Tree

We can print the config as a Rich Tree in both the console and the output folder (see Output Results sections of each paradigm). The config tree is customizable; its configs are in the misc/config_tree/ sub-sub-config.

Required config fields:

Field Description
print Whether to print the config tree in the console
save Whether to save the config tree in the output folder
save_path The path to save the config tree in the output folder
style, guide_style The style of the config tree. Please refer to the Rich documentation for more details
fields_order the order of the fields to be shown in the config tree

Metadata

Some metadata are specified in misc/:

  • timestamp: the timestamp of the pipeline. We use it to name the output folder so that each run can be distinguished. It is set as ${now:%Y-%m-%d_%H-%M-%S}. For example, the output folder can be outputs/${expr_name}/${misc.timestamp}, which becomes like outputs/example_cl_main_expr/2023-10-01_12-00-00 if the timestamp is 2023-10-01_12-00-00.
Back to top
Callbacks
Custom Implementation
 
 

©️ 2025 Pengxiang Wang. All rights reserved.