Supertron-embedding-300M is a high-performance, compact embedding model fine-tuned from the google/embeddinggemma-300m architecture. It is specifically designed to provide state-of-the-art semantic representations for Retrieval-Augmented Generation (RAG), semantic search, and document clustering applications while maintaining a low computational footprint suitable for production environments.
Supertron-embedding-300M demonstrates competitive performance across the Massive Text Embedding Benchmark (MTEB). It is particularly effective in Semantic Textual Similarity (STS) tasks, outperforming many larger models in its weight class.
1from sentence_transformers import SentenceTransformer
2
3model_id = "surpem/Supertron-embedding-300M"
4
5# Load the model
6model = SentenceTransformer(model_id)
7
8# Define target text
9sentences = [
10 "The financial results exceeded market expectations.",
11 "The company reported better than expected quarterly earnings."
12]
13
14# Compute embeddings
15embeddings = model.encode(sentences)
16
17# Calculate cosine similarity
18similarity = model.similarity(embeddings[0], embeddings[1])
19print(f"Semantic Similarity: {similarity.item():.4f}")
20Training Procedure
21Hyperparameters
22Precision: bfloat16
23
24Max Sequence Length: 256 tokens
25
26Optimizer: AdamW
27
28Batch Size: 256
29
30Learning Rate: 2e-5
31
32Citation
33Code-Snippet
34@misc{surpem2026supertron,
35 title={Supertron-embedding-300M: High-Efficiency Semantic Representation Model},
36 author={Surpem},
37 year={2026},
38 url={[https://huggingface.co/surpem/Supertron-embedding-300M](https://huggingface.co/surpem/Supertron-embedding-300M)},
39}
