Metadata-Version: 2.1
Name: openvalidators
Version: 0.0.1
Summary: Openvalidators is a collection of open source validators for the Bittensor Network.
Home-page: https://github.com/opentensor/foundation_validator
Author: bittensor.com
Author-email: 
Maintainer: 
Maintainer-email: 
License: MIT
Keywords: bittensor,validator,ai,machine-learning,deep-learning,blockchain,pytorch,torch,neural-networks,cryptocurrency
Platform: UNKNOWN
Classifier: Intended Audience :: Developers
Classifier: Topic :: Software Development :: Build Tools
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3 :: Only
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Topic :: Scientific/Engineering
Classifier: Topic :: Scientific/Engineering :: Mathematics
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Topic :: Software Development
Classifier: Topic :: Software Development :: Libraries
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: bittensor (==5.1.0)
Requires-Dist: transformers (==4.20.1)
Requires-Dist: wandb (==0.15.3)
Requires-Dist: datasets (==2.12.0)
Requires-Dist: plotly (==5.14.1)
Requires-Dist: networkx (==3.1)
Requires-Dist: scipy (==1.10.1)

<div align="center">

# **Open Validators** <!-- omit in toc -->
[![Discord Chat](https://img.shields.io/discord/308323056592486420.svg)](https://discord.gg/bittensor)
[![PyPI version](https://badge.fury.io/py/test-openvalidators.svg)](https://badge.fury.io/py/openvalidators-test)
[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT) 

---

</div>


This repository contains Bittensor Validators designed by the OpenTensor Foundation team for the community.
It offers several functionalities, such as:

- Building and running Bittensor validators
- Real-time analysis of validator performance integrated with wandb
- Offline analysis of data generated from the network
- Creation of datasets using network data for training miners 

The main goal of this repository is to facilitate the interaction with the Bittensor network by providing a set of
open-source validators to the community. The current validator implementation queries the network for responses and 
evaluations using carefully crafted prompts, that are later evaluated by a large foundation GPT-J reward model.

Additionally, the repository provides an analysis and data toolkit that allows users to analyze the data generated from
the validator's interaction with the network. By default, the validator collects various data points, such as question 
responses, evaluations, rewards and scorings by UID, and model performance data. This data is then sent to wandb, 
making it publicly accessible to the community.

The toolkit also includes scripts to analyze and extract data from specific validator runs or multiple runs, simplifying
the creation of valuable datasets for the community's miners.

To learn more about the Bittensor validation process, check out this [documentation](https://tensor-wiki.vercel.app/validating/validating).

# Usage
There are currently four main avenues for engaging with this repository:

1. [Validators](#Validators):
   - Designed for TAO holders who aim to build or run validators developed by the foundation.

2. [Real-time performance analysis with wandb integration](#Real-time-performance-analysis-with-wandb-integration):
   - Allows users to analyze the performance of various validators runs in real-time using wandb.

3. [Data analysis](#Data-analysis)
   - Caters to individuals, researchers, and data scientists interested in analyzing the data generated from the
validators' interaction with the Bittensor network.

4. [Dataset creation](#Dataset-creation)
   - Serves individuals, researchers, and developers who seek to create datasets for the community's miners.

# Install
There are two ways to use OpenTensor validators:

1. With pip:
```bash
$ pip3 install open-validators
```

2. From source:
```bash
$ git clone https://github.com/opentensor/openvalidators.git
$ pip install -e openvalidators/
```

You can test the installation by running the following command:
```bash
$ python3 openvalidators/neuron.py --help
```

# Validators
Participation in Network Validation is available to TAO holders. The validation mechanism utilizes a dual
proof-of-stake and proof-of-work system known as Yuma Consensus, which you can learn more about
[here](https://tensor-wiki.vercel.app/validating/validating). To start validating, you will need to have a 
Bittensor wallet with a sufficient amount of TAO tokens staked.

Once you have your wallet ready for validation, you can start the foundation validator by running the following command:
```bash
$ python3 openvalidators/neuron.py --wallet.name <your-wallet-name> --wallet.hotkey <your-wallet-hot-key>
```

# Real-time performance analysis with wandb integration
By default, the validator sends data to wandb, allowing users to analyze the performance of the validator in real-time.
All the data sent to wandb is publicly available to the community at the following [link](https://wandb.ai/opentensor-dev/openvalidators).

You don't need to have a wandb account to access the data or to generate a new run,
but bear in mind that
[data generated by anonymous users will be deleted after 7 days](https://docs.wandb.ai/guides/app/features/anon#:~:text=If%20there's%20no%20account%2C%20we,be%20available%20for%207%20days)
as default wandb policy.

# Data analysis
This repository provides a set of tools to analyze the data generated by the validators. 
The scripts and notebooks are located in the [analysis](./analysis) folder.

# Dataset creation
For the individuals who are eager to create datasets tailored specifically for the community's miners.
With convenient scripts available in the [scripts](./scripts) folder, you can effortlessly download data from specific or multiple runs 
of wandb, empowering you to curate comprehensive and valuable datasets that align with your mining objectives.

# License

The MIT License (MIT) Copyright © 2023 Yuma Rao

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated 
documentation files (the “Software”), to deal in the Software without restriction, including without limitation the
rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit 
persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE 
WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR 
COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

