| license: gemma | |
| base_model: | |
| - litert-community/embeddinggemma-300m | |
| tags: | |
| - embeddinggemma | |
| - google-tensor | |
| - tensor-g4 | |
| - npu | |
| - edgetpu | |
| - litert | |
| - aot | |
| pipeline_tag: feature-extraction | |
| # EmbeddingGemma-300M — Tensor G4 NPU (Android 17 / Beta-SDK recompile) | |
| AOT-compiled **EmbeddingGemma-300M** for the **Google Tensor G4** NPU (darwinn EdgeTPU), built with the **official Google Tensor ML SDK (Beta)**. | |
| > ⚠️ **Android firmware note — why this repo exists.** The G4 NPU bytecode (DGC) is compiled against a specific Tensor NPU firmware. Builds compiled against **Android 16** firmware fail to load on **Android 17** (newer NPU runtime → "Failed to get Darwinn graph" / SB-invocation error). This repo holds an **A17-targeted recompile** on the current Beta SDK. The older **Android 16** build: [xThr45hx/EmbeddingGemma-300M-Tensor-G4-NPU](https://huggingface.co/xThr45hx/EmbeddingGemma-300M-Tensor-G4-NPU). | |
| > | |
| > 🚧 **Status: on-device A17 load verification in progress.** This build compiles clean (2265/2265 ops, single partition, DGC0 + rio_a0); confirming it loads + runs on a real Android 17 device is the next step. Provisional until this note is updated. | |
| ## File | |
| - `embeddinggemma-300M_seq256_Google_Tensor_G4.tflite` — seq256 (max 256 tokens in one pass), 768-d output. The efficient RAG workhorse for short chunks/queries. (A seq512 long-form variant may follow.) | |
| ## How it was compiled | |
| - **Input:** `embeddinggemma-300M_seq256_mixed-precision.tflite` from [litert-community/embeddinggemma-300m](https://huggingface.co/litert-community/embeddinggemma-300m) (the plain, non-device-compiled mixed-precision file). | |
| - **SDK:** `ai-edge-litert-nightly` + `ai-edge-litert-sdk-google-tensor==2.1.5`; official `aot_compile(target=[TENSOR_G4])`, **no flags** (mixed-precision path); mandatory `google_tensor_backend` import. | |
| - **Result:** 2265 / 2265 ops offloaded to **1 partition** (fully fused, no fallback). Output 196,993,056 bytes, markers **DGC0 + rio_a0 + tfl3**. SHA-256 `eec2daf64f07f8cc84a92080c5e2afb00fc6bdf0cb688e00638c0229620b0b4a`. | |
| ## License | |
| Gemma — inherits from EmbeddingGemma. See the base model card for terms. | |