jayyun98/embeddinggemma-2-text-270m-bf16

EmbeddingGemma 2 — Text Only — BF16 — Transformers / SentenceTransformers

A modular deployment export of Google DeepMind's EmbeddingGemma 2. Only the text backbone is retained; the audio encoder and audio projection are absent, as are the vision encoder and projection. No training, distillation, or quantization was applied.

Retained weights preserve the upstream BF16 precision.

Property	Value
Runtime	Transformers / SentenceTransformers
Stored floating-point tensors	BF16 (all tensors checked)
Effective model parameters	271,002,624
Weight files	542.06 MB, decimal
Inputs	Text and code
Output	768 dimensions; MRL at 512, 256, 128
Context budget	8,192 tokens
License	Apache 2.0

Weight size is not total runtime memory. 270M/440M in repository names are rounded deployment sizes.

Install

pip install "transformers>=5.19.0" "sentence-transformers>=6.1.0"

Tested with Transformers 5.19.0, SentenceTransformers 6.1.0, and PyTorch 2.14.1.

Text and code

import torch
from sentence_transformers import SentenceTransformer

model = SentenceTransformer(
    "jayyun98/embeddinggemma-2-text-270m-bf16",
    device="mps",
    model_kwargs={"dtype": torch.bfloat16},
)
queries = model.encode(
    ["What causes the northern lights?", "로컬 코드 검색 모델을 찾고 싶어요."],
    prompt_name="SearchQuery",
    normalize_embeddings=True,
)
documents = model.encode(
    ["Charged particles from the sun cause the northern lights."],
    prompt_name="Document",
    normalize_embeddings=True,
)
print(model.similarity(queries, documents))

code_query = model.encode(
    "Find a Python function that sorts a list.",
    prompt_name="CodeRetrieval",
    truncate_dim=256,
    normalize_embeddings=True,
)

Use SearchQuery for search queries, CodeRetrieval for code-search queries, and Document for corpus items. For the MLX API, prepend the corresponding literal prefix from config_sentence_transformers.json, as shown above. For titled documents, use title: {title} | text: {content} without another prefix. Use matching dimensions for queries and documents, and normalize after truncation. Do not cast this model to FP16; use BF16 or FP32 inference.

Verification

MPS GPU BF16 inference was tested in the package's named precision. Outputs were exactly equal to the full source model running in the same precision. Checks cover English/Korean search and document text, code queries, and 128/256/512-dimensional normalized vectors.

Fixture	Minimum cosine vs source FP32	Maximum absolute difference
search_english_korean	0.999962518	0.00236937404
documents	0.999946502	0.00167062879
code	0.999964441	0.00194609165

Measurements compare with the pinned full Google checkpoint. They are small numerical and loading checks, not MTEB results, a retrieval-quality evaluation, or a speed benchmark. The original prompts, pooling behavior, tokenizer and processor assets are retained. Processor metadata does not restore the removed encoder weights. Image, audio and video encoders are unavailable.

• conversion.json: storage precision, removal and conversion checks.
• verification.json: runtime, numerical results and tested environment.
• Original source revision: 914f7f89142e33e77833254d9c9b90c3cef7303b.

Attribution and license

Original weights and tokenizer/processor assets: Google DeepMind.

This is an independent derivative deployment export, not an official Google release. The Apache 2.0 LICENSE and NOTICE are included. See the original model card for training, intended use and limitations.