Parsbert Fa Sentiment Twitter

🚀 ParsBERT Digikala Sentiment Analysis Model

This model is a fine - tuned ParsBERT Digikala sentiment analysis model on approximately 600,000 Persian tweets.

🚀 Quick Start

System Requirements

To load the model, you need at least 650 megabytes of RAM and disk space. You also need to have the tensorflow, transformers, and numpy libraries installed.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

import numpy as np
from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
#loading model
tokenizer = AutoTokenizer.from_pretrained("nimaafshar/parsbert-fa-sentiment-twitter")
  
model = TFAutoModelForSequenceClassification.from_pretrained("nimaafshar/parsbert-fa-sentiment-twitter")
classes = ["negative","neutral","positive"]

Advanced Usage

#using model
sequences = [".غذا خیلی افتضاح بود متاسفم برای مدیریت رستورن خیلی بد بود.",
 "خیلی خوشمزده و عالی بود عالی",
"می‌تونم اسمتونو بپرسم؟"
]

for sequence in sequences:
  inputs = tokenizer(sequence, return_tensors="tf")
  classification_logits = model(inputs)[0]
  results = tf.nn.softmax(classification_logits, axis=1).numpy()[0]
  print(classes[np.argmax(results)])
  percentages = np.around(results*100)
  print(percentages)

⚠️ Important Note

This model is trained on a Persian corpus and is intended for use on Persian texts.