Setting Configuration
In order to run the simulation, the Orchestrator instance must receive a settings object that contains all the necessary parameters. It is possible to store those parameters in a JSON format and load them as the Python dictionary by using asociita.utils.helper.load_from_json function. Below is an exemplary settings object embedded as a json file. All the elements are necessary unless stated otherwise.
{
"orchestrator":{
"iterations": int,
"number_of_nodes": int,
"local_warm_start": bool,
"sample_size": int,
"evaluation": "none" | "full"
"save_metrics": bool,
"save_models": bool,
"save_path": str
"nodes": [0,
1,
2]
},
"nodes":{
"local_epochs": int,
"model_settings": {
"optimizer": "RMS",
"batch_size": int,
"learning_rate": float}
}
}
The settings contains two dictionaries: orchestrator and nodes.
orchestrator contains all the settings necessary details of the training:
iterationsis the number of rounds to be performed. Example:iterations:12number_of_nodesis the number of nodes that will be included in the training. Example:number_of_nodes: 10local_warm_startallows to distribute various pre-trained weights to different local clients. Not implemeneted yet. Example:local_warm_start: false.sample_sizeis the size of the sample that will be taken each round. Example:sample_size : 4.evaluationallows to control the evaluation procedure across the clients. Currently, onlynoneorfullare supported. Setting the evaluation to full will perform a full evaluation of every client included in the training. Example:evaluation: fullsave_metricsallows to control whether the metrics should be saved in a csv file. Example:save_metrics: true.save_modelsallows to control whether the models should be saved. Not implemeneted yet. Example:save_metrics: false.save_pathis the system path that will be used when saving the model. It is possible to define a saving_path in a method call.nodesis the list containing the ids of all the nodes participating in the training. Length ofnodesmust be equalnumber_of_nodes.
nodes contains all the necessary configuration for nodes.
"local_epochs":the number of local epochs to be performed on the local nodes."model_settings"is a dictionary containing all the parameters for training the model.optimizeris an optimizer that will be used during the training. Example:optimizer: "RMS"batch_sizeis the batch size that will be used during the training. Example:batch_size: 32learning_rateis the learning rate that will be used during the training. Example:learning_rate: 0.001
1 maintainer
Metadata
1 maintainer
Release files for asociita 0.3.4
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| asociita-0.3.4.tar.gz | 91.2 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| asociita-0.3.4-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 151.4 kB
Release files / asociita-0.3.4.tar.gz
| Download URL | asociita-0.3.4.tar.gz |
|---|---|
| Size | 91.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
40df60c3204e8a734ac2f026f16d2e0ddaa0152c77678ebd12bc92ca26fc44e1
|
|
BLAKE2b-256 checksum How to use checksums |
3205f1774b2f98c1f67c820b04048aed7fe27c4fde0517fbc34f942cd8f4ffa4
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.5.0 CPython/3.10.11 Linux/5.19.0-50-generic
|
Release files / asociita-0.3.4-py3-none-any.whl
| Download URL | asociita-0.3.4-py3-none-any.whl |
|---|---|
| Size | 60.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
ab6014a7a9844960a18952bcb8c6f021019f7935ce0c67197618d78c9bf25478
|
|
BLAKE2b-256 checksum How to use checksums |
98d063c8ce10e613cb7e5355aafe6e9d5f494c424fc2bbc36e69420deeffbfb9
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.5.0 CPython/3.10.11 Linux/5.19.0-50-generic
|
Security for asociita 0.3.4
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
