---
license: apache-2.0
language:
- fa
metrics:
- f1
base_model:
- google/embeddinggemma-300m
pipeline_tag: text-classification
---

## Zirneshan
Zirneshane model is designed to predict whether to add kasre (ِ) at the end of each word in a sentence in Persian language.

This is a `token` classification problem and the focus is only on word-end position.


![image](/static-proxy?url=https%3A%2F%2Fcdn-uploads.huggingface.co%2Fproduction%2Fuploads%2F63ecdb2460ff4b318ad24653%2F5cDoIE-zGA_bmBbJ-IfII.png)


### Usage
Install according to the repo [Zirneshane](https://github.com/SadeghKrmi/zirneshane)
```
pip3 install git+ssh://git@github.com/SadeghKrmi/zirneshane
```

load model and predict
```
from src.model import HybridKasreModel
model = HybridKasreModel.load("./models/zirneshan-word-char-embedding-classifier-v1.0.pt", device="cpu")
model.eval()

result = model.predict("بعد از دیدن این صحنه‌های دلخراش، با عجله فراوان، همه شهر، مجبور به ترک مجلس شدند.", device="cpu")

print(result)
```