You can deploy this model and start earning money today!

gomyk/
minilm-student-L6_top

Name: minilm-student-L6_top
Rating: 2.5 (45 reviews)

sentence-transformers

Views

No views yet

L6_top

Lightweight sentence encoder created from sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 via layer pruning + vocabulary pruning.

Model Details

Property	Value
Teacher	sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
Architecture	MiniLM-L12 (pruned)
Hidden dim	384
Layers	6 / 12
Layer indices	[6, 7, 8, 9, 10, 11]
Strategy	6 layers, top half (semantic-focused)
Parameters	106,825,344
Model size (FP32)	98.1MB
Distilled	No

Architecture

==============================================================
  TEACHER: MiniLM-L12  →  STUDENT: 6L / 38,775 vocab
==============================================================

            TEACHER                        STUDENT          
  ───────────────────────────    ───────────────────────────

  ┌─────────────────────────┐    ┌─────────────────────────┐
  │   Input Tokens          │    │   Input Tokens          │
  └────────────┬────────────┘    └────────────┬────────────┘
               │                              │
  ┌────────────┴────────────┐    ┌────────────┴────────────┐
  │  Embeddings             │    │  Embeddings (pruned)    │
  │  vocab: 250,002         │    │  vocab:  38,775         │
  │  dim:  384              │    │  dim:  384              │
  └────────────┬────────────┘    └────────────┬────────────┘
               │                              │
  ┌─────────────────────────┐    ┌─────────────────────────┐
  │  Layer  0               │  ╳   │                         │
  ├ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─┤    │                         │
  │  Layer  1               │  ╳   │                         │
  ├ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─┤    │                         │
  │  Layer  2               │  ╳   │                         │
  ├ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─┤    │                         │
  │  Layer  3               │  ╳   │                         │
  ├ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─┤    │                         │
  │  Layer  4               │  ╳   │                         │
  ├ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─┤    │                         │
  │  Layer  5               │  ╳   │                         │
  ├─────────────────────────┤    ├─────────────────────────┤
  │  Layer  6               │ ──►  │  Layer  0 ← L6         │
  ├─────────────────────────┤    ├─────────────────────────┤
  │  Layer  7               │ ──►  │  Layer  1 ← L7         │
  ├─────────────────────────┤    ├─────────────────────────┤
  │  Layer  8               │ ──►  │  Layer  2 ← L8         │
  ├─────────────────────────┤    ├─────────────────────────┤
  │  Layer  9               │ ──►  │  Layer  3 ← L9         │
  ├─────────────────────────┤    ├─────────────────────────┤
  │  Layer 10               │ ──►  │  Layer  4 ← L10        │
  ├─────────────────────────┤    ├─────────────────────────┤
  │  Layer 11               │ ──►  │  Layer  5 ← L11        │
  └────────────┬────────────┘    └────────────┬────────────┘
               │                              │
  ┌────────────┴────────────┐    ┌────────────┴────────────┐
  │  Mean Pooling           │    │  Mean Pooling           │
  │  → 384d embedding       │    │  → 384d embedding       │
  └─────────────────────────┘    └─────────────────────────┘

  Size: 448.0MB (FP32)           →  98.1MB (FP32)
  Params: 117,451,392        →  25,714,176
  Reduction: 78.1%
==============================================================

Quick Start

python

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("L6_top", trust_remote_code=True)

sentences = [
    "Hello, how are you?",
    "안녕하세요",
    "Bonjour, comment allez-vous?",
]

embeddings = model.encode(sentences)
print(embeddings.shape)  # (3, 384)

Training

Created via layer pruning + vocabulary pruning (no additional training):

Teacher: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 (12 layers, 384d)
Layer selection: [6, 7, 8, 9, 10, 11] - 6 layers, top half (semantic-focused)
Vocab pruning: Corpus-based filtering for target languages

Supported Languages (18)

ko, en, ja, zh, es, fr, de, pt, it, ru, ar, hi, th, vi, id, tr, nl, pl

gomyk/minilm-student-L6_top