Downloads · 30 days
14
14% of all-time downloads
gemma-student-gemma_emb_mrl_distilled_128d
yrrhall/gemma-student-gemma_emb_mrl_distilled_128d
gemma-student-gemma_emb_mrl_distilled_128d is a sentence similarity model from yrrhall. Use it when you need a score for how close two texts are. It is set up for sentence-transformers. The card lists the license as gemma.
Compact multilingual sentence encoder compressed from google/embeddinggemma-300m (27x compression).
All-time downloads
99
Public
Parameters
11.2M
49.6 MB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors44.9 MB · 79%
At a glance
- Task
- Sentence Similarity
- Library
- sentence-transformers
- License
- gemma
- Model type
- gemma3_text
- Access
- Public
- Created
- Jul 17, 2026
- Updated
- Jul 17, 2026
- SHA
- 8c97442e
Model card
From the Hugging Face model README
gemma_emb_mrl_distilled_128d (Distilled)
Compact multilingual sentence encoder compressed from google/embeddinggemma-300m (27x compression).
Model Details
| Property | Value |
|---|---|
| Base model | google/embeddinggemma-300m |
| Architecture | gemma3_text (decoder) |
| Hidden dim | 128 (from 768) |
| Layers | 6 (from 24) |
| Intermediate | 192 |
| Attention heads | 1 |
| KV heads | 1 |
| Vocab size | 81,052 (from 262,144) |
| Parameters | ~11.2M |
| Model size (FP32) | 42.8MB |
| Compression | 27x |
| Distilled | Yes (2-stage) |
Quick Start
from sentence_transformers import SentenceTransformer
model = SentenceTransformer("gemma_emb_mrl_distilled_128d", trust_remote_code=True)
sentences = [
"Hello, how are you?",
"안녕하세요, 잘 지내세요?",
"こんにちは、元気ですか？",
"你好，你好吗？",
]
embeddings = model.encode(sentences)
print(embeddings.shape) # (4, 128)
MTEB Evaluation Results
Overall Average: 23.8%
| Task Group | Average |
|---|---|
| Classification | 31.32% |
| Clustering | 27.55% |
| STS | 13.78% |
Classification
| Task | Average | Details |
|---|---|---|
| AmazonCounterfactualClassification | 52.78% | en: 54.49%, ja: 52.73%, en-ext: 52.64%, de: 51.27% |
| Banking77Classification | 8.31% | default: 8.31% |
| ImdbClassification | 50.36% | default: 50.36% |
| MTOPDomainClassification | 27.22% | th: 36.78%, de: 29.73%, es: 29.22%, en: 26.76%, fr: 24.91% |
| MassiveIntentClassification | 13.73% | th: 19.26%, cy: 18.33%, pl: 18.0%, sq: 17.67%, it: 17.43% |
| MassiveScenarioClassification | 15.97% | cy: 20.38%, sl: 20.23%, sv: 19.6%, km: 19.52%, ka: 19.4% |
| ToxicConversationsClassification | 47.43% | default: 47.43% |
| TweetSentimentExtractionClassification | 34.78% | default: 34.78% |
Clustering
| Task | Average | Details |
|---|---|---|
| ArXivHierarchicalClusteringP2P | 48.99% | default: 48.99% |
| ArXivHierarchicalClusteringS2S | 46.35% | default: 46.35% |
| BiorxivClusteringP2P.v2 | 6.42% | default: 6.42% |
| MedrxivClusteringP2P.v2 | 18.95% | default: 18.95% |
| MedrxivClusteringS2S.v2 | 18.15% | default: 18.15% |
| StackExchangeClustering.v2 | 40.92% | default: 40.92% |
| StackExchangeClusteringP2P.v2 | 32.65% | default: 32.65% |
| TwentyNewsgroupsClustering.v2 | 7.97% | default: 7.97% |
STS
| Task | Average | Details |
|---|---|---|
| BIOSSES | 8.94% | default: 8.94% |
| SICK-R | 20.63% | default: 20.63% |
| STS12 | 12.23% | default: 12.23% |
| STS13 | 15.73% | default: 15.73% |
| STS14 | 8.76% | default: 8.76% |
| STS15 | 27.92% | default: 27.92% |
| STS17 | 17.1% | es-es: 34.2%, en-en: 33.26%, ar-ar: 30.05%, ko-ko: 21.98%, nl-en: 13.79% |
| STS22.v2 | 5.47% | es: 18.55%, fr: 18.16%, en: 17.61%, fr-pl: 16.9%, zh: 14.38% |
| STSBenchmark | 7.27% | default: 7.27% |
Training
Stage 1: Model Compression
- Teacher:
google/embeddinggemma-300m(24L, 768d, 303M params) - Compression: Layer pruning → Hidden dim reduction → Vocab pruning
- Result: 6L / 128d / 81,052 vocab
Stage 2: Two-Stage Knowledge Distillation
Compression ratio 27x requires progressive distillation:
- Stage 1: Teacher (303M) → Intermediate (~61M)
- MSE + Cosine Similarity loss
- MTEB task datasets (Classification/Clustering/STS)
- Stage 2: Intermediate → Final Student (11.2M)
- Same training objective
- AdamW (lr=2e-5, weight_decay=0.01), Cosine annealing
License
This model is a derivative of Google's Gemma. Gemma is provided under and subject to the Gemma Terms of Use found at ai.google.dev/gemma/terms. Use of this model must comply with the Gemma Prohibited Use Policy.
Supported Languages (18)
ko, en, ja, zh, es, fr, de, pt, it, ru, ar, hi, th, vi, id, tr, nl, pl
- Task
- Sentence Similarity
- Library
- sentence-transformers
- Type
- gemma3_text
- License
- gemma
- Languages
- ko, en, ja, zh, es, fr
- Created
- Jul 17, 2026
- Updated
- Jul 17, 2026