EmbeddingGemma-300M

Text embeddings using Google's GEMMA model. 768-dimensional vectors (Matryoshka: 128/256/512/768) optimized for search, classification, and retrieval.

Usage

import replicate, base64, numpy as np

# Returns a base64 string by default
b64 = replicate.run(
    "zsxkib/embedding-gemma-300m",
    input={"text": "Your text here"}
)

# Decode base64 -> float32 vector
embedding = np.frombuffer(base64.b64decode(b64), dtype=np.float32)
print(embedding.shape)  # (768,)

Parameters

• text: Input text to embed
• task: Task type (retrieval_query, retrieval_document, classification)
• output_format: Return format (base64 or array)

Output

Returns a base64 string by default (efficient). Set output_format: "array" to get a list of floats.

Local Development

cog predict -i text="Hello world"

Model Details

• Architecture: Google GEMMA embedding model (300M parameters)
• Dimensions: 768
• Max tokens: 2048 (auto-truncated)
• Tasks: Retrieval queries/documents, classification

---

Built by @zsakib_ • GitHub