Update README.md

README.md

@@ -9,4 +9,42 @@ metrics:
 base_model:
 - FacebookAI/xlm-roberta-large
 pipeline_tag: text-classification
----
+---
+
+# Model Card for Persian-EmoRoBERTa-BiGRU
+
+## Model Details
+
+### Model Description
+
+This model is designed for emotion recognition in Persian text. It predicts the presence of six emotions: anger, disgust, fear, sadness, happiness, and surprise, as well as the primary emotion within these categories, including an "other" category for cases when none of the specified emotions are present. The model leverages XLM-RoBERTa, a pre-trained transformer-based language model, fine-tuned on two datasets: EmoPars and ArmanEmo. It includes a Bidirectional Gated Recurrent Unit (BiGRU) layer to better capture contextual dependencies, improving performance on emotion classification tasks.
+
+- **Developed by:** Faezeh Sarlakifar, Morteza Mahdavi Mortazavi, Mehrnoush Shamsfard
+- **Model type:** Text Emotion Classification (Transformer + BiGRU)
+- **Language(s):** Persian
+- **License:** MIT
+- **Finetuned from model:** XLM-RoBERTa (a pre-trained transformer model)
+
+### Model Sources
+
+- **Repository:** [GitHub Repository](https://github.com/faezesarlakifar/text-emotion-recognition)
+- **Paper:** [EmoRecBiGRU: Emotion Recognition in Persian Tweets with a Transformer-based Model, Enhanced by Bidirectional GRU](http://journal.itrc.ac.ir/article-1-653-en.html)
+
+## Uses
+
+### Direct Use
+
+This model can be directly used for emotion recognition in Persian text. It can predict the presence of six emotions and the primary emotion from those six or an "other" category. The model checkpoints can be downloaded for local usage or integrated into existing systems.
+
+### Downstream Use
+
+The model can be further fine-tuned for specific tasks or integrated into larger applications such as sentiment analysis systems, chatbots, and customer service systems where emotion recognition is required.
+
+### Out-of-Scope Use
+
+This model should not be used for tasks involving languages other than Persian or general-purpose sentiment analysis without further adaptation.
+
+### Recommendations
+
+Users should be aware of the model's limitations and biases, especially in high-stakes applications such as mental health or sensitive customer feedback. For critical applications, it is recommended to combine this model with other validation tools.
+