{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreid5rl5bzghvjxe35mo3p5z3kkhn5ra7w4rl2wocw36b5pqctqcuui",
"uri": "at://did:plc:nfto3lv2rcs5s7h7digotzlu/app.bsky.feed.post/3mojy4gawa3c2"
},
"coverImage": {
"$type": "blob",
"ref": {
"$link": "bafkreig7saeenw6c3prrms24e2bpcmq7mskq26az4ajg73ufx33cf4gf2a"
},
"mimeType": "image/png",
"size": 32236
},
"path": "/packages/betto_inferencing",
"publishedAt": "2026-06-18T04:22:05.234Z",
"site": "https://pub.dev",
"textContent": "ONNX Runtime inference and embedding models for dense text retrieval. Changelog excerpt: Initial development release providing ONNX Runtime inference and embedding models for dense text retrieval on native platforms (macOS, Linux, Windows, Android, iOS). ### Features - **`EmbeddingModel`**— abstract interface for text-to-vector embedding, decoupling consumers from any specific inference backend. - **`OnnxEmbeddingModel`**— ONNX Runtime implementation backed by BGE Small En v1.5, delivering dense embeddings suitable for semantic search and retrieval tasks. - **`BertTokenizer`/ `Tok[...]",
"title": "v0.1.0-dev.1 of betto_inferencing",
"updatedAt": "2026-06-18T03:36:12.096Z"
}