davidit17/minilm-grocery-hebrew-v2

Hebrew Grocery Item Classifier v2

Fine-tuned multilingual-MiniLMv2-L6-mnli-xnli for classifying Hebrew grocery shopping list items into 12 supermarket categories. This is an improved version over v1 with a larger, more diverse training set and targeted augmentation for weak classes. finetune code at https://github.com/davidit17/shopping_list_bot

What's new in v2

• Larger dataset — 3043 total items (2434 train / 609 test), up from ~600 in v1
• 4 training sources — manual lists, Excel grocery data, Gemini-labelled web data, and a ChatGPT-labelled set
• Synthetic augmentation — rule-based Hebrew item generator to balance under-represented categories
• Weak-class booster — extra hand-curated examples for משקאות, קפואים וסלטים, and רטבים וממרחים
• Per-class F1 tracking — evaluation now reports per-category F1 to surface weak spots

Categories

• מוצרי חלב וביצים
• ירקות ופירות
• לחם מאפים ודגנים
• חטיפים ומתוקים
• ניקיון טיפוח וחד פעמי
• משקאות
• מוצרים לאפייה
• בשר ודגים
• קפואים וסלטים
• יבשים
• רטבים וממרחים
• אחר

Usage

from transformers import pipeline

classifier = pipeline(
    "text-classification",
    model=f"davidit17/minilm-grocery-hebrew-v2",
)

items = ["גבינה צהובה", "עגבניות שרי", "שמפו", "קוקה קולה", "חזה עוף"]
for item in items:
    result = classifier(item)[0]
    print(item, "→", result["label"], f"({result['score']:.2f})")

With confidence threshold (recommended)

CONFIDENCE_THRESHOLD = 0.5

def predict(text: str) -> str:
    result = classifier(text)[0]
    return result["label"] if result["score"] >= CONFIDENCE_THRESHOLD else "אחר"

Training

Parameter	Value
Base model	multilingual-MiniLMv2-L6-mnli-xnli
Max epochs	50 (early stopping, patience=8)
Learning rate	2e-5
Weight decay	0.01
Warmup ratio	0.1
Batch size	16
Max sequence length	32
Train / test split	80% / 20% stratified
Hardware	GPU (CUDA)

Data sources

Source	Description
Source A	Manual Hebrew grocery lists from Israeli forum
Source B	Excel-format grocery list, cleaned and melted
Source C	Web grocery list labelled with Gemini
Source D	ChatGPT-labelled items across all 12 categories
Source E	Synthetic items generated with rule-based Hebrew augmentor
Source F	Hand-curated booster set for weak categories

Evaluation

Overall

Metric	Base (zero-shot)	Fine-tuned v2
Accuracy	0.1494	0.8818
Weighted F1	0.1363	0.8808

Per-class F1 (fine-tuned v2)

Category	F1
מוצרי חלב וביצים	0.9057
ירקות ופירות	0.8696
לחם מאפים ודגנים	0.8767
חטיפים ומתוקים	0.9091
ניקיון טיפוח וחד פעמי	0.9020
משקאות	0.8750
מוצרים לאפייה	0.8687
בשר ודגים	0.9333
קפואים וסלטים	0.8632
יבשים	0.8000
רטבים וממרחים	0.8913
אחר	0.9024

Limitations

• Primarily covers Israeli supermarket conventions and Hebrew product naming.
• Brand-specific or very niche items may fall back to אחר.
• Low-confidence predictions (score < 0.5) should be treated as אחר.
• Synthetic training examples may not fully reflect natural shopping list variation.
• The אחר category is a catch-all and may have lower precision than domain-specific categories.