Model Hub

import { Tenzro } from 'tenzro';

const tenzro = new Tenzro();

// List available models
const models = await tenzro.hub.list({
  category: 'text-generation',
  runtime: 'onnx',
});

// Get model details
const model = await tenzro.hub.get('tenzro/llama-3.2-1b-instruct-onnx');
console.log(model.name, model.size, model.downloads);

// Download model
const downloaded = await tenzro.hub.download('tenzro/llama-3.2-1b-instruct-onnx', {
  progress: (percent) => console.log(`${percent}% downloaded`),
});

import { TenzroEdge } from '@tenzro/edge';

const edge = new TenzroEdge({
  clientId: 'client_xxx',
});

// Load model for browser inference
const model = await edge.hub.load('tenzro/phi-3-mini-onnx', {
  runtime: 'webgpu', // or 'webgl', 'wasm'
  cache: true,       // Cache in IndexedDB
});

// Run inference
const result = await model.generate({
  prompt: 'Explain quantum computing',
  maxTokens: 100,
});

console.log(result.text);

// Upload a model
const upload = await tenzro.hub.upload({
  name: 'my-custom-model',
  description: 'Fine-tuned model for customer support',
  format: 'onnx',
  files: [
    { path: 'model.onnx', content: modelBuffer },
    { path: 'tokenizer.json', content: tokenizerBuffer },
  ],
  metadata: {
    baseModel: 'llama-3.2-1b',
    task: 'text-generation',
    quantization: 'int8',
  },
  visibility: 'private', // or 'public'
});

console.log('Model uploaded:', upload.id);

// Create a new version
await tenzro.hub.createVersion('my-org/my-model', {
  version: '1.1.0',
  files: updatedFiles,
  changelog: 'Improved accuracy on edge cases',
});

// List versions
const versions = await tenzro.hub.listVersions('my-org/my-model');

// Load specific version
const model = await edge.hub.load('my-org/my-model@1.0.0');

// Set model visibility
await tenzro.hub.updateVisibility('my-org/my-model', {
  visibility: 'private',
});

// Grant access to specific users/organizations
await tenzro.hub.grantAccess('my-org/my-model', {
  users: ['user-id-1', 'user-id-2'],
  organizations: ['partner-org'],
});

// Revoke access
await tenzro.hub.revokeAccess('my-org/my-model', {
  users: ['user-id-1'],
});

const model = await edge.hub.load('tenzro/llama-3.2-3b-onnx', {
  // Load only needed chunks (streaming inference)
  streaming: true,

  // Progress callback
  onProgress: ({ loaded, total, chunk }) => {
    console.log(`Loading chunk ${chunk}: ${loaded}/${total}`);
  },
});

await tenzro.hub.updateModelCard('my-org/my-model', {
  description: 'A fine-tuned model for customer support',
  intendedUse: 'Customer service chatbots',
  limitations: 'May not handle technical queries well',
  trainingData: 'Customer support transcripts (anonymized)',
  evaluation: {
    accuracy: 0.92,
    f1Score: 0.89,
  },
  license: 'Apache-2.0',
});