Open-EmbeddingGemma: pure PyTorch reimplementation of Google's 740M multimodal embedding model
KyeGomezB · x · 2026-10-10
Kye Gomez (founder of Swarms) released Open-EmbeddingGemma, a pure PyTorch implementation of Google DeepMind's 740M-parameter multimodal embedding model EmbeddingGemma 2.
- Single-file implementation with no transformers dependency; module names mirror the official checkpoint, so released safetensors load directly.
- Text, code, images, video and audio all pass through one bidirectional Gemma-style encoder, producing a 768-dim L2-normalized vector.
- Supports Matryoshka truncation to 512/256/128 dims.
- Apache-2.0 licensed, with an example.py demo.
Related event: Open-EmbeddingGemma Reproduces Google's 740M Embedding Model in PyTorch(3 posts)→
More from coding & agent
- YC partner shares multi-agent workflow: agents write briefs for each other like middle managers — ycombinator · 2026-10-10
- AI writing skill boosts output to 4 posts a week, human editors become the bottleneck — danshipper · 2026-10-10
- AI agent digs through Azure billing to recover nearly $2,000 in lost credits — pswider · 2026-10-10
- HQ Agent Tool Ships Major Update: Per-App Postgres, Full-Stack Next.js, Interactive Pages — jacob_posel · 2026-10-10
- Agent swarm converts GTA footage to 3D scenes in just a few hours — Daniel_Farinax · 2026-10-10
- OpenAI's Huet notes DevDay Chromatic handheld has Wi-Fi, enabling Codex-powered hacks — romainhuet · 2026-10-10