CODE SAMPLES

Examples

Real-world code examples and usage patterns for building on GOKA.

Basic Inference

Run a simple inference request on the GOKA network

import { Goka } from '@goka/sdk';

const goka = new Goka({ network: 'mainnet-beta' });

const result = await goka.inference.run({
  model: 'gpt-4-turbo',
  input: 'Explain quantum computing in simple terms.',
  maxTokens: 500,
});

console.log(result.output);
// "Quantum computing uses quantum bits (qubits)..."
Streaming Response

Stream inference results for real-time applications

import { Goka } from '@goka/sdk';

const goka = new Goka({ network: 'mainnet-beta' });

// Stream tokens as they're generated
for await (const chunk of goka.inference.stream({
  model: 'gpt-4-turbo',
  input: 'Write a poem about the future of AI.',
})) {
  process.stdout.write(chunk);
}

// Output streams in real-time as it's generated
Distributed Training

Train a model across multiple GOKA compute nodes

import { Goka } from '@goka/sdk';

const goka = new Goka({ network: 'mainnet-beta' });

const job = await goka.train.start({
  dataset: 'ipfs://Qm...',
  modelType: 'transformer',
  epochs: 100,
  batchSize: 32,
  distributed: true,
  nodes: 'auto', // Auto-select optimal nodes
});

// Monitor training progress
job.on('progress', (progress) => {
  console.log(`Epoch ${progress.epoch}: loss=${progress.loss}`);
});

await job.complete();
console.log('Model saved to:', job.modelPath);
Model Deployment

Deploy your trained model to the GOKA marketplace

import { Goka } from '@goka/sdk';

const goka = new Goka({ network: 'mainnet-beta' });

const deployment = await goka.models.deploy({
  name: 'my-custom-model',
  path: 'ipfs://Qm...',
  description: 'Fine-tuned for code generation',
  pricing: {
    perInference: '0.0001 SOL',
    perToken: '0.00001 SOL',
  },
  public: true,
});

console.log('Deployed at:', deployment.endpoint);
// "https://api.goka.ai/v1/models/my-custom-model"
CLI Usage

Use the GOKA CLI for quick operations

# Install CLI globally
npm install -g @goka/cli

# Initialize project
goka init my-project

# Connect wallet
goka auth login

# Run inference
goka run --model gpt-4-turbo --input "Hello GOKA"

# Start training
goka train --dataset ./data --epochs 50

# Deploy model
goka deploy ./model --name my-model --public

# Check status
goka status
Docker Deployment

Run GOKA node in a container

# docker-compose.yml
version: '3.8'
services:
  goka-node:
    image: goka/node:latest
    environment:
      - NETWORK=mainnet-beta
      - WALLET_PATH=/wallet/wallet.json
      - GPU_ENABLED=true
    volumes:
      - ./wallet:/wallet
      - ./data:/data
    ports:
      - "8080:8080"
    deploy:
      resources:
        reservations:
          devices:
            - capabilities: [gpu]

# Run with: docker-compose up -d

Ready to build on decentralized AI?