ipai (InterPlanetary Artificial Intelligence)

InterPlanetary Artificial Intelligence (IPAI): store your personal AI model leveraging on InterPlanetary Data Wallet ( IPDW). It enables to build a real censorship resistant research swarm without centralised repositories. Browser based inferences, Personal fine-tuning, Zero-backend privacy, and much more.

Features

• Store your personal AI model on the InterPlanetary Data Wallet (IPDW)
• Publish a without centralized repositories
• Take advantage of browser-based inferences
• Personalize your AI model through fine-tuning
• Enjoy zero-backend privacy for your AI model
• And much more!

Getting Started

To get started with IPAI, you will need to follow these steps:

Clone the repository:

$ git clone https://github.com/ansi-code/ipai.git

Install the required dependencies:

$ npm install

Follow the instructions provided in the IPDW repository to set up your wallet.

Start using IPAI.

const model = await GPTNeoXForCausalLm.Load("ipfs://QmecpDvGdWfcKw7BM4nxyEb7TB856sTY1MqY1dCR45rWjv", console.log);
const tokenizer = await GPT2Tokenizer.Load('ipfs://QmRnFHciVJxtpTtGktB3vLRMMxutEaAybXvwobXKLxRpd9', 'ipfs://QmQWBu2Cd4KnBGeeT9dx7JSG6v9VJg1QeiDg3EbBtSLKkD', console.log);

const prompt = "In a shocking finding, scientists discovered a herd of unicorns living in a remote, previously unexplored valley, in the Andes Mountains.\nEven more surprising to the researchers was the fact that the unicorns spoke perfect English.";
const inputIds = tokenizer.encode(prompt);

const genTokens = await model.generate(inputIds, true, 0.9, 1, 1, 150, async t => {
    process.stdout.write(tokenizer.decode([t]))
});

const genText = tokenizer.decode(genTokens);
console.log("Final text:", genText);

Contributing

We welcome contributions to IPAI! If you would like to contribute, please follow these steps:

1. Fork the repository
2. Create a new branch for your changes
3. Commit your changes and open a pull request
4. Support

If you need help using IPAI or have any questions, please open an issue in this repository and we will be happy to assist you.

TODO

• Migrate matrices computation to NumTs (https://github.com/ansi-code/numts)
• Custom sparse matrices backend

License

This project is licensed under the Apache 2.0 License. See the LICENSE file for details.