Hydra

https://hydra.cc/docs
Admin
Hydra is a framework for configuring complex applications, enabling elegant composition of...
Tokens:132,206
Snippets:1,116
Trust Score:10
Update:2 weeks ago
Show doc for...
Tokens:
Raw
### Install the Hydra app example with pip

Source: https://hydra.cc/docs/advanced/app_packaging

Install the example Hydra application as a Python package. The output confirms successful installation of hydra-app-0.1.

```bash
$ pip install examples/advanced/hydra_app_example  

...  

Successfully installed hydra-app-0.1  

```

--------------------------------

### Launch Ray launcher with new cluster

Source: https://hydra.cc/docs/plugins/ray_launcher

Launches the example application with the Ray launcher, starting a new Ray cluster. Shows the expected output including the Ray dashboard URL.

```shell
$ python my_app.py  --multirun hydra/launcher=ray  

[HYDRA] Ray Launcher is launching 1 jobs, sweep output dir: multirun/2020-11-10/15-16-28  

[HYDRA] Initializing ray with config: {}  

INFO services.py:1164 -- View the Ray dashboard at http://127.0.0.1:8266  

[HYDRA]        #0 :  

(pid=97801) [__main__][INFO] - Executing task 1  

```

--------------------------------

### Start local development server with corepack pnpm start

Source: https://hydra.cc/docs/development/documentation

Starts a local development server and opens a browser window. Most changes are reflected live without restarting the server.

```bash
$ corepack pnpm start
```

--------------------------------

### Example multirun with launcher

Source: https://hydra.cc/docs/plugins/submitit_launcher

Example of launching a multirun with the launcher; shows the sweep output.

```bash
$ python my_app.py task=1,2,3 --multirun

[HYDRA] Sweep output dir : multirun/2020-05-28/15-05-22

[HYDRA]        #0 : task=1

[HYDRA]        #1 : task=2

[HYDRA]        #2 : task=3
```

--------------------------------

### Install hydra-submitit-launcher

Source: https://hydra.cc/docs/plugins/submitit_launcher

Install the Submitit Launcher plugin for Hydra.

```bash
pip install hydra-submitit-launcher --upgrade
```

--------------------------------

### Install hydra-joblib-launcher

Source: https://hydra.cc/docs/plugins/joblib_launcher

Installs the Joblib Launcher plugin. Use the --upgrade flag to ensure the latest version is installed.

```bash
pip install hydra-joblib-launcher --upgrade
```

--------------------------------

### Example output of --help

Source: https://hydra.cc/docs/configure_hydra/app_help

Example output of the `--help` command for an application with a mandatory defaults list choice. It shows the app name, configuration groups, and the generated config with overrides.

```text
$ python my_app.py --help  

== AwesomeApp ==  

  

This is AwesomeApp!  

You can choose a db driver by appending  

== Configuration groups ==  

Compose your configuration from those groups (db=mysql)  

  

db: mysql, postgresql  

  

  

== Config ==  

This is the config generated for this run.  

You can override everything, for example:  

python my_app.py db.user=foo db.pass=bar  

-------  

db:  

  driver: mysql  

  user: omry  

  pass: secret  

  

-------  

  

Powered by Hydra (https://hydra.cc)  

Use --hydra-help to view Hydra specific help  


```

--------------------------------

### Discover Optuna sweeper parameters

Source: https://hydra.cc/docs/plugins/optuna_sweeper

Discover the Optuna sweeper parameters for the multi-objective example. Run from the directory containing the example.

```bash
python example/multi-objective.py hydra/sweeper=optuna --cfg hydra -p hydra.sweeper  
```

--------------------------------

### Run example app to save job configs

Source: https://hydra.cc/docs/experimental/rerun

Run the example app to save job configs. The output shows the callback saving config.pickle and job_return.pickle in the .hydra subdirectory, and the app logging the output directory and cfg.foo.

```bash
$ python my_app.py
[2022-03-16 14:51:30,905][hydra.experimental.pickle_job_info_callback][INFO] - Saving job configs in /Users/jieru/workspace/hydra/examples/experimental/outputs/2022-03-16/14-51-30/.hydra/config.pickle
[2022-03-16 14:51:30,906][__main__][INFO] - Output_dir=/Users/jieru/workspace/hydra/examples/experimental/outputs/2022-03-16/14-51-30
[2022-03-16 14:51:30,906][__main__][INFO] - cfg.foo=bar
[2022-03-16 14:51:30,906][hydra.experimental.pickle_job_info_callback][INFO] - Saving job_return in /Users/jieru/workspace/hydra/examples/experimental/outputs/2022-03-16/14-51-30/.hydra/job_return.pickle
```

--------------------------------

### Example RQ launcher multirun output

Source: https://hydra.cc/docs/plugins/rq_launcher

Example output when running a multirun with the RQ launcher. It enqueues five jobs and polls every 1.0 seconds.

```bash
$ python my_app.py --multirun task=1,2,3,4,5  

[HYDRA] RQ Launcher is enqueuing 5 job(s) in queue : default  

[HYDRA] Sweep output dir : multirun/2020-06-15/18-00-00  

[HYDRA] Enqueued 13b3da4e-03f7-4d16-9ca8-cfb3c48afeae  

[HYDRA] 	#1 : task=1  

[HYDRA] Enqueued 00c6a32d-e5a4-432c-a0f3-b9d4ef0dd585  

[HYDRA] 	#2 : task=2  

[HYDRA] Enqueued 63b90f27-0711-4c95-8f63-70164fd850df  

[HYDRA] 	#3 : task=3  

[HYDRA] Enqueued b1d49825-8b28-4516-90ca-8106477e1eb1  

[HYDRA] 	#4 : task=4  

[HYDRA] Enqueued ed96bdaa-087d-4c7f-9ecb-56daf948d5e2  

[HYDRA] 	#5 : task=5  

[HYDRA] Finished enqueuing  

[HYDRA] Polling job statuses every 1.0 sec  
```

--------------------------------

### Install hydra-optuna-sweeper

Source: https://hydra.cc/docs/plugins/optuna_sweeper

Install the Optuna Sweeper plugin via pip.

```bash
pip install --upgrade --pre "hydra-optuna-sweeper>=1.4.0.dev1,<1.5.0.dev0"  
```

--------------------------------

### optimization_results.yaml example

Source: https://hydra.cc/docs/plugins/optuna_sweeper

Example content of the optimization_results.yaml file showing best parameters and value.

```yaml
name: optuna  

best_params:  

  x: 0.0  

  'y': 0  

best_value: 0.0  
```

--------------------------------

### Run the standalone Hydra app example

Source: https://hydra.cc/docs/advanced/app_packaging

Run the standalone Hydra application example from the repository. The output shows the default configuration with the imagenet dataset.

```bash
$ python examples/advanced/hydra_app_example/hydra_app/main.py  

dataset:  

  name: imagenet  

  path: /datasets/imagenet  

```

--------------------------------

### Run sphere.py with --multirun

Source: https://hydra.cc/docs/plugins/optuna_sweeper

Run the optimization example with the Optuna sweeper.

```bash
python example/sphere.py --multirun  
```

--------------------------------

### Install hydra-rq-launcher

Source: https://hydra.cc/docs/plugins/rq_launcher

Install the RQ Launcher plugin. Requires a Redis server; RQ does not support Windows.

```bash
pip install hydra-rq-launcher --upgrade  
```

--------------------------------

### Config directory structure

Source: https://hydra.cc/docs/advanced/defaults_list

Shows the directory layout for the example configs.

```text
├── server  

│   ├── db  

│   │   ├── mysql.yaml  

│   │   └── sqlite.yaml  

│   └── apache.yaml  

└── config.yaml  



```

--------------------------------

### Optimal parameters for sphere example

Source: https://hydra.cc/docs/plugins/optuna_sweeper

The optimal parameters for the sphere example, where the minimum is 0.

```yaml
x: 0  

y: 0  
```

--------------------------------

### Run example hyperparameter search

Source: https://hydra.cc/docs/plugins/nevergrad_sweeper

Run the example hyperparameter search using the Nevergrad sweeper. Execute this command from the plugins/hydra_nevergrad_sweeper directory after cloning the code.

```bash
python example/my_app.py -m  
```

--------------------------------

### Output of Banana function run

Source: https://hydra.cc/docs/plugins/ax_sweeper

Example output from running the Banana function example. It shows the parameters being optimized, the Bayesian Optimization generation strategy (Sobol+GPEI), the 5 launched jobs, and the final best value and parameters.

```text
[HYDRA] AxSweeper is optimizing the following parameters:  
banana.x: range=[-5, 5]  
banana.y: range=[-5.0, 10.1]  
ax.modelbridge.dispatch_utils: Using Bayesian Optimization generation strategy: GenerationStrategy(name='Sobol+GPEI', steps=[Sobol for 5 trials, GPEI for subsequent trials]). Iterations after 5 will take longer to generate due to model-fitting.  
[HYDRA] AxSweeper is launching 5 jobs  
[HYDRA] Launching 5 jobs locally  
[HYDRA]        #0 : banana.x=2 banana.y=-0.988  
[__main__][INFO] - Banana_Function(x=2, y=-0.988)=2488.883  
[HYDRA]        #1 : banana.x=-1 banana.y=7.701  
[__main__][INFO] - Banana_Function(x=-1, y=7.701)=4493.987  
[HYDRA]        #2 : banana.x=-1 banana.y=-3.901  
[__main__][INFO] - Banana_Function(x=-1, y=-3.901)=2406.259  
[HYDRA]        #3 : banana.x=-1 banana.y=0.209  
[__main__][INFO] - Banana_Function(x=-1, y=0.209)=66.639  
[HYDRA]        #4 : banana.x=4 banana.y=-4.557  
[__main__][INFO] - Banana_Function(x=4, y=-4.557)=42270.006  
[HYDRA] New best value: 66.639, best parameters: {'banana.x': -1, 'banana.y': 0.209}  
```

--------------------------------

### Directory structure for composition example

Source: https://hydra.cc/docs/intro

Directory structure for the composition example. It shows the conf directory with a config group named db containing mysql.yaml and postgresql.yaml.

```text
├── conf  
│   ├── config.yaml  
│   ├── db  
│   │   ├── mysql.yaml  
│   │   └── postgresql.yaml  
│   └── __init__.py  
└── my_app.py  
```

--------------------------------

### Base configs and directory structure for package examples

Source: https://hydra.cc/docs/advanced/overriding_packages

Base configuration files and directory structure used in the package examples. The config.yaml defaults list includes server/apache, which in turn includes db: mysql, and the directory tree shows the layout of the config files.

```yaml
defaults:  

  - server/apache  

  

debug: false  

  

  

   

```

```yaml
defaults:  

  - db: mysql  

  

name: apache  

  

  

   

```

```yaml
name: mysql  

```

```yaml
name: sqlite  

```

```text
├── server  
│   ├── db  
│   │   ├── mysql.yaml  
│   │   └── sqlite.yaml  
│   └── apache.yaml  
└── config.yaml  

```

--------------------------------

### Optimal parameters for example

Source: https://hydra.cc/docs/plugins/nevergrad_sweeper

The optimal parameter values that minimize the dummy function to 0. These values are used as a reference for the example.

```yaml
db: mnist  

lr: 0.12  

dropout: 0.33  

batch_size: 4  
```

--------------------------------

### Config file search space example

Source: https://hydra.cc/docs/plugins/optuna_sweeper

Shows how to express a command-line override in a config file. For example, `x=range(1,4)` becomes `x: range(1,4)` under the `hydra.sweeper.params` node.

```yaml
x: range(1,4)
```

--------------------------------

### Example classes for partial instantiation

Source: https://hydra.cc/docs/advanced/instantiate_objects/overview

Defines `Optimizer` and `Model` classes used in the partial instantiation example. `Model` accepts a partial optimizer and calls it with `lr`.

```python
class Optimizer:  

    algo: str  

    lr: float  

  

    def __init__(self, algo: str, lr: float) -> None:  

        self.algo = algo  

        self.lr = lr  

  

    def __repr__(self) -> str:  

        return f"Optimizer(algo={self.algo},lr={self.lr})"  

  

  

class Model:  

    def __init__(self, optim_partial: Any, lr: float):  

        super().__init__()  

        self.optim = optim_partial(lr=lr)  

        self.lr = lr  

  

    def __repr__(self) -> str:  

        return f"Model(Optimizer={self.optim},lr={self.lr})"  
```

--------------------------------

### Install hydra-nevergrad-sweeper

Source: https://hydra.cc/docs/plugins/nevergrad_sweeper

Install the Nevergrad sweeper plugin for Hydra. Use the --upgrade flag to ensure the latest version is installed.

```bash
pip install hydra-nevergrad-sweeper --upgrade  
```

--------------------------------

### Install hydra-ax-sweeper

Source: https://hydra.cc/docs/plugins/ax_sweeper

Install the Ax Sweeper plugin for Hydra. Use the --upgrade flag to ensure the latest version is installed.

```bash
pip install hydra-ax-sweeper --upgrade  
```

--------------------------------

### Sweep initialization and first evaluations output

Source: https://hydra.cc/docs/plugins/nevergrad_sweeper

Example output from the sweep initialization and the first 5 evaluations (out of 100). Shows the optimizer configuration, parametrization, job launches, and the best parameters found.

```text
[2020-10-08 20:13:53,592][HYDRA] NevergradSweeper(optimizer=NGOpt, budget=100, num_workers=10) minimization  

[2020-10-08 20:13:53,593][HYDRA] with parametrization Dict(batch_size=Choice(choices=Tuple(4,8,16),weights=Array{(1,3)}),db=Choice(choices=Tuple(mnist,cifar),weights=Array{(1,2)}),dropout=Scalar{Cl(0,1,b)}[sigma=Log{exp=2.0}],lr=Log{exp=3.162277660168379,Cl(0.001,1,b)}):{'db': 'mnist', 'lr': 0.03162277660168379, 'dropout': 0.5, 'batch_size': 8}  

[2020-10-08 20:13:53,593][HYDRA] Sweep output dir: multirun/2020-10-08/20-13-53  

[2020-10-08 20:13:55,023][HYDRA] Launching 10 jobs locally  

[2020-10-08 20:13:55,023][HYDRA]        #0 : db=mnist lr=0.03162277660168379 dropout=0.5 batch_size=16  

[2020-10-08 20:13:55,217][__main__][INFO] - dummy_training(dropout=0.500, lr=0.032, db=mnist, batch_size=16) = 13.258  

[2020-10-08 20:13:55,218][HYDRA]        #1 : db=cifar lr=0.018178519762066934 dropout=0.5061074452336254 batch_size=4  

[2020-10-08 20:13:55,408][__main__][INFO] - dummy_training(dropout=0.506, lr=0.018, db=cifar, batch_size=4) = 0.278  

[2020-10-08 20:13:55,409][HYDRA]        #2 : db=cifar lr=0.10056825918734161 dropout=0.6399687427725211 batch_size=4  

[2020-10-08 20:13:55,595][__main__][INFO] - dummy_training(dropout=0.640, lr=0.101, db=cifar, batch_size=4) = 0.329  

[2020-10-08 20:13:55,596][HYDRA]        #3 : db=mnist lr=0.06617542958182834 dropout=0.5059497416026679 batch_size=8  

[2020-10-08 20:13:55,812][__main__][INFO] - dummy_training(dropout=0.506, lr=0.066, db=mnist, batch_size=8) = 5.230  

[2020-10-08 20:13:55,813][HYDRA]        #4 : db=mnist lr=0.16717013388679514 dropout=0.6519070394318255 batch_size=4  

...  

[2020-10-08 20:14:27,988][HYDRA] Best parameters: db=cifar lr=0.11961221693764439 dropout=0.37285878409770895 batch_size=4  
```

--------------------------------

### Install hydra_colorlog

Source: https://hydra.cc/docs/plugins/colorlog

Install the Hydra colorlog plugin via pip. Use the --upgrade flag to ensure the latest version is installed.

```bash
pip install hydra_colorlog --upgrade  
```

--------------------------------

### Install hydra-ray-launcher with pip

Source: https://hydra.cc/docs/plugins/ray_launcher

Install the Hydra Ray Launcher plugin using pip. The --upgrade flag ensures the latest version is installed.

```bash
$ pip install hydra-ray-launcher --upgrade  
```

--------------------------------

### Inspect search path with --info flag

Source: https://hydra.cc/docs/advanced/search_path

Use the --info flag with the 'searchpath' argument to display the current search path and loaded configurations. Run from the command line with your application script.

```bash
$ python my_app.py --info searchpath  
```

--------------------------------

### Tag examples

Source: https://hydra.cc/docs/advanced/override_grammar/extended

Examples of tagging sweeps with one or more tags.

```python
tag(log,interval(0,1))          # 1.0 <= x < 1.0, tags=[log]  
tag(foo,bar,interval(0,1))      # 1.0 <= x < 1.0, tags=[foo,bar]  

```

--------------------------------

### Interval sweep examples

Source: https://hydra.cc/docs/advanced/override_grammar/extended

Examples of interval sweeps, showing auto-casting to floats.

```python
interval(1.0,5.0)  # 1.0 <= x < 5.0  
interval(1,5)      # 1.0 <= x < 5.0, auto-cast to floats  

```

--------------------------------

### Running with default dataset (imagenet)

Source: https://hydra.cc/docs/patterns/specializing_config

Running the example with default settings (imagenet) does not apply the specialized config, so num_layers remains 7.

```bash
$ python example.py  

dataset:  

  name: imagenet  

  path: /datasets/imagenet  

model:  

  num_layers: 7  

  type: alexnet  

```

--------------------------------

### Demonstrate `_convert_` strategies with `instantiate`

Source: https://hydra.cc/docs/advanced/instantiate_objects/overview

Shows how the `_convert_` parameter changes the types of arguments passed to the instantiated target. `"none"` keeps OmegaConf containers, `"partial"` converts non-Structured Configs to dicts, `"object"` converts Structured Configs to dataclass instances, and `"all"` converts everything to primitive containers. The `_execution_whitelist_` is required to allow instantiating `__main__.*` classes.

```python
from dataclasses import dataclass  

from omegaconf import DictConfig, OmegaConf  

from hydra.utils import instantiate  

  

@dataclass  

class Foo:  

    a: int = 123  

  

class MyTarget:  

    def __init__(self, foo, bar):  

        self.foo = foo  

        self.bar = bar  

  

cfg = OmegaConf.create(  

    {  

        "_target_": "__main__.MyTarget",  

        "foo": Foo(),  

        "bar": {"b": 456},  

    }  

)  

  

obj_none = instantiate(cfg, _convert_="none", _execution_whitelist_="__main__.*")  

assert isinstance(obj_none, MyTarget)  

assert isinstance(obj_none.foo, DictConfig)  

assert isinstance(obj_none.bar, DictConfig)  

  

obj_partial = instantiate(cfg, _convert_="partial", _execution_whitelist_="__main__.*")  

assert isinstance(obj_partial, MyTarget)  

assert isinstance(obj_partial.foo, DictConfig)  

assert isinstance(obj_partial.bar, dict)  

  

obj_object = instantiate(cfg, _convert_="object", _execution_whitelist_="__main__.*")  

assert isinstance(obj_object, MyTarget)  

assert isinstance(obj_object.foo, Foo)  

assert isinstance(obj_object.bar, dict)  

  

obj_all = instantiate(cfg, _convert_="all", _execution_whitelist_="__main__.*")  

assert isinstance(obj_all, MyTarget)  

assert isinstance(obj_all.foo, dict)  

assert isinstance(obj_all.bar, dict)  
```

--------------------------------

### Install website dependencies with corepack pnpm install

Source: https://hydra.cc/docs/development/documentation

Installs website dependencies from the website directory. pnpm is configured to avoid resolving npm package versions published in the last 10 days during dependency updates.

```bash
$ corepack pnpm install
```

--------------------------------

### Install Hydra 1.4 development release for testing

Source: https://hydra.cc/docs/upgrades/1.3_to_1.4/prepare_for_1_4

Use this pip command to install a prerelease of Hydra 1.4 for early migration testing. The upper bound prevents installation of a later Hydra release. Install matching Hydra 1.4 versions of any launchers or sweepers used by the application. Do not use a Hydra 1.4 development release in production.

```bash
python -m pip install --upgrade --pre "hydra-core>=1.4.0.dev0,<1.5"
```

--------------------------------

### Example config.yaml and foo/bar.yaml

Source: https://hydra.cc/docs/upgrades/1.0_to_1.1/default_composition_order

Two YAML config files used to illustrate the composition order change. The primary config `config.yaml` has a defaults list and a `foo` key, while `foo/bar.yaml` is a config from the defaults list.

```yaml
defaults:  
  - foo: bar  

foo:  
  x: 10  
```

```yaml
# @package _group_  
x: 20  
```

--------------------------------

### my_app.py - Structured Configs example

Source: https://hydra.cc/docs/advanced/instantiate_objects/structured_config

Defines DBConnection classes and structured configs (DBConfig, MySQLConfig, PostGreSQLConfig, Config). Stores configs in ConfigStore and uses @hydra.main with instantiate to create a connection. The _execution_whitelist_ restricts instantiation to my_app.*.

```python
class DBConnection:  

    def __init__(self, driver: str, host: str, port: int) -> None:  

        self.driver = driver  

        self.host = host  

        self.port = port  

  

    def connect(self) -> None:  

        print(f"{self.driver} connecting to {self.host}")  

  

class MySQLConnection(DBConnection):  

    def __init__(self, driver: str, host: str, port: int) -> None:  

        super().__init__(driver=driver, host=host, port=port)  

  

class PostgreSQLConnection(DBConnection):  

    def __init__(self, driver: str, host: str, port: int, timeout: int) -> None:  

        super().__init__(driver=driver, host=host, port=port)  

        self.timeout = timeout  

  

@dataclass  

class DBConfig:  

    driver: str = MISSING  

    host: str = "localhost"  

    port: int = 80  

  

@dataclass  

class MySQLConfig(DBConfig):  

    _target_: str = "my_app.MySQLConnection"  

    driver: str = "MySQL"  

    port: int = 1234  

  

@dataclass  

class PostGreSQLConfig(DBConfig):  

    _target_: str = "my_app.PostgreSQLConnection"  

    driver: str = "PostgreSQL"  

    port: int = 5678  

    timeout: int = 10  

  

@dataclass  

class Config:  

    defaults: list[Any] = field(default_factory=lambda: [{"db": "mysql"}])  

    db: DBConfig = MISSING  

  

  

cs = ConfigStore.instance()  

cs.store(name="config", node=Config)  

cs.store(group="db", name="mysql", node=MySQLConfig)  

cs.store(group="db", name="postgresql", node=PostGreSQLConfig)  

  

@hydra.main(config_name="config")  

def my_app(cfg: Config) -> None:  

    connection = instantiate(cfg.db, _execution_whitelist_="my_app.*")  

    connection.connect()  

  

if __name__ == "__main__":  

    my_app()  
```

--------------------------------

### Install Hydra Core 1.4

Source: https://hydra.cc/docs/plugins/optuna_sweeper

Install the required Hydra Core version for the Optuna Sweeper plugin.

```bash
pip install --upgrade --pre "hydra-core>=1.4.0.dev1,<1.5.0.dev0"  
```

--------------------------------

### Output Config Example

Source: https://hydra.cc/docs/advanced/defaults_list

Output of running 'python my_app.py --cfg job' showing the composed Output Config.

```yaml
server:  

  db:  

    name: mysql  

  name: apache  

debug: false  

```

--------------------------------

### Range sweep examples

Source: https://hydra.cc/docs/advanced/override_grammar/extended

Examples of range sweeps with integer and floating-point steps, including negative step.

```python
num=range(5)                          # 0,1,2,3,4  
num=range(0,5)                        # 0,1,2,3,4  
num=range(0,5,2)                      # 0,2,4  
num=range(0,10,3.3)                   # 0.0,3.3,6.6,9.9  
num=range(-5,step=-1)                 # 0,-1,-2,-3,-4  

```

--------------------------------

### Run multirun with Joblib launcher

Source: https://hydra.cc/docs/plugins/joblib_launcher

Example command and output showing five parallel executions with the Joblib launcher. The output shows the Joblib.Parallel configuration and the process IDs for each task.

```bash
$ python my_app.py --multirun task=1,2,3,4,5

[HYDRA] Joblib.Parallel(n_jobs=10,backend=loky,prefer=processes,require=None,verbose=0,timeout=None,pre_dispatch=2*n_jobs,batch_size=auto,temp_folder=None,max_nbytes=None,mmap_mode=r) is launching 5 jobs

[HYDRA] Launching jobs, sweep output dir : multirun/2020-02-18/10-00-00

[__main__][INFO] - Process ID 14336 executing task 2 ...

[__main__][INFO] - Process ID 14333 executing task 1 ...

[__main__][INFO] - Process ID 14334 executing task 3 ...

[__main__][INFO] - Process ID 14335 executing task 4 ...

[__main__][INFO] - Process ID 14337 executing task 5 ...
```

--------------------------------

### Config directory structure

Source: https://hydra.cc/docs/patterns/select_multiple_configs_from_config_group

Shows the file layout for the example: config.yaml and server directory with apache.yaml and site subdirectory containing amazon.yaml, fb.yaml, and google.yaml.

```text
├── config.yaml  

└── server  

    ├── apache.yaml  

    └── site  

        ├── amazon.yaml  

        ├── fb.yaml  

        └── google.yaml  

```

--------------------------------

### Run multi-objective optimization

Source: https://hydra.cc/docs/plugins/optuna_sweeper

Run the multi-objective optimization example. Execute from the plugins/hydra_optuna_sweeper directory.

```bash
python example/multi-objective.py --multirun  
```

--------------------------------

### Install Hydra stable release

Source: https://hydra.cc/docs/intro

Install the latest stable release of Hydra from PyPI. Use this for production or when you need the stable version.

```bash
pip install hydra-core --upgrade  
```

--------------------------------

### Single-quoted value examples

Source: https://hydra.cc/docs/advanced/override_grammar/basic

Examples of single-quoted values in Hydra overrides, including escaped single quotes, backslashes, and embedded double quotes.

```plaintext
'hello there'  

'escaped \'single quote\''  

'the path is C:\\\'some folder\'\\'  

'1,2,3'  

'{a:10} ${xyz}'  

'"double quoted string"'  

```

--------------------------------

### Double-quoted value examples

Source: https://hydra.cc/docs/advanced/override_grammar/basic

Examples of double-quoted values in Hydra overrides, including escaped double quotes, backslashes, and embedded single quotes.

```plaintext
"hello there"  

"escaped \"double quote\""  

"the path is C:\\\"some folder\"\\"  

"1,2,3"  

"{a:10} ${xyz}"  

"'single quoted string'"  

```

--------------------------------

### Initial config.yaml

Source: https://hydra.cc/docs/patterns/specializing_config

Initial configuration with defaults for dataset and model.

```yaml
defaults:  

  - dataset: imagenet  

  - model: alexnet  

```

--------------------------------

### Basic defaults list example

Source: https://hydra.cc/docs/advanced/defaults_list

Shows config.yaml, server/apache.yaml, server/db/mysql.yaml, server/db/sqlite.yaml, and the output of running the app. The defaults list includes a config group 'server/apache' which itself has a default 'db: mysql'.

```yaml
defaults:  

  - server/apache  

debug: false  


  



```

```yaml
defaults:  

  - db: mysql  

name: apache  


  



```

```yaml
name: mysql

```

```yaml
name: sqlite

```

```text
$ python my_app.py

server:  

  db:  

    name: mysql  

  name: apache  

debug: false  



```

--------------------------------

### Install Hydra development release

Source: https://hydra.cc/docs/intro

Install the development release of Hydra 1.4 from PyPI. Use this to try upcoming features before the stable release is available.

```bash
pip install --pre --upgrade hydra-core  
```

--------------------------------

### Run the installed Hydra app

Source: https://hydra.cc/docs/advanced/app_packaging

Run the installed Hydra application. It uses the packaged configuration files and outputs the same dataset configuration as the source run.

```bash
$ hydra_app  

dataset:  

  name: imagenet  

  path: /datasets/imagenet  

```

--------------------------------

### Instantiation with partial config

Source: https://hydra.cc/docs/advanced/instantiate_objects/overview

Instantiates the model using `instantiate` with an execution whitelist. The output shows the resulting `Model` with the partial optimizer.

```python
with execution_whitelist("my_app.*"):  

    model = instantiate(cfg.model)  

print(model)  

# "Model(Optimizer=Optimizer(algo=SGD,lr=0.01),lr=0.01)  
```

--------------------------------

### shuffle usage examples

Source: https://hydra.cc/docs/advanced/override_grammar/extended

Examples of using shuffle with positional arguments, sweeps, ranges, and lists. Comments indicate the result is a shuffled version of the input.

```python
shuffle(a,b,c)                                       # shuffled a,b,c  

shuffle(choice(a,b,c)), shuffle(sweep=choice(a,b,c)) # shuffled choice(a,b,c)  

shuffle(range(1,10))                                 # shuffled range(1,10)  

shuffle([a,b,c]), shuffle(list=[a,b,c])              # shuffled list [a,b,c]  


```
Tokens:
Raw