Update README.md

README.md

@@ -1,74 +1,75 @@
 ---
+language: mk
+tags:
+  - macedonian
+  - mistral
+  - llm
+  - nlp
+  - text-generation
 license: apache-2.0
+datasets:
+  - macedonian-wikipedia
+  - news-articles
+  - books
+metrics:
+  - perplexity
+  - bleu
+  - rouge
+  - accuracy
 ---
-**MK-LLM-Mistral: Open Macedonian Language Model**
 
-🌍 About This Model
-MK-LLM-Mistral is the **first Macedonian Large Language Model (LLM)**, trained using a fine-tuned version of **Mistral-7B**.  
-This project is developed by **AI Now - Association for Artificial Intelligence in Macedonia**.
+ # MK-LLM-Mistral: Fine-Tuned Macedonian Language Model  
 
-📌 Website: [www.ainow.mk](https://www.ainow.mk)  
-📩 Contact: [contact@ainow.mk](mailto:[email protected])  
-🛠 GitHub Repository: [MK-LLM](https://github.com/AI-now-mk/MK-LLM)  
+## 🌍 Overview  
+**MK-LLM-Mistral** is a **fine-tuned Macedonian language model**, built to enhance **text generation, comprehension, and NLP capabilities** in the Macedonian language.  
+This model is developed by **AI Now - Association for Artificial Intelligence in Macedonia** as part of the **MK-LLM initiative**, Macedonia's first open-source LLM project.
 
----
-
-## 📌 Model Details
-- Model Name: MK-LLM-Mistral  
-- Base Model: [Mistral-7B](https://huggingface.co/mistralai/Mistral-7B)  
-- Language: Macedonian 🇲🇰  
-- Fine-tuned on: Wikipedia, news articles, government websites, Macedonian books  
-- Tasks: Chatbot, Text Completion, Q&A, Macedonian NLP  
+📌 **Website:** [www.ainow.mk](https://www.ainow.mk)  
+📩 **Contact:** [[email protected]](mailto:[email protected])  
+🛠 **GitHub Repository:** [MK-LLM](https://github.com/AI-now-mk/MK-LLM)  
 
 ---
 
-🛠 How to Use This Model
-### 1️⃣ Install Dependencies
-```bash
-pip install transformers torch huggingface_hub
+## 📌 Model Details  
+- **Architecture:** Fine-tuned **Mistral 7B**  
+- **Language:** Macedonian 🇲🇰  
+- **Training Data:** Macedonian Wikipedia, news articles, books, and open-source datasets  
+- **Tokenization:** Custom Macedonian tokenization  
+- **Framework:** [Hugging Face Transformers](https://huggingface.co/docs/transformers/index)  
+- **Model Type:** Causal Language Model (CLM)  
 
-2️⃣ Load the Model for Inference
-from transformers import AutoModelForCausalLM, AutoTokenizer
-import torch
+---
 
-# Load the fine-tuned model
-MODEL_NAME = "ainowmk/MK-LLM-Mistral"
-tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
-model = AutoModelForCausalLM.from_pretrained(MODEL_NAME)
+## 🎯 Intended Use  
+This model is optimized for **Macedonian NLP tasks**, including:  
+✅ **Text Generation** – Macedonian text continuation and creative writing  
+✅ **Summarization** – Extracting key points from Macedonian documents  
+✅ **Question Answering** – Responding to Macedonian-language queries  
+✅ **Chatbots & Virtual Assistants** – Enhancing automated Macedonian-language interactions  
 
-# Move model to GPU if available
-device = "cuda" if torch.cuda.is_available() else "cpu"
-model.to(device)
+---
 
-# Example prompt in Macedonian
-input_text = "Здраво, како си?"
-inputs = tokenizer(input_text, return_tensors="pt").to(device)
-outputs = model.generate(**inputs, max_length=100)
+## ⚠️ Limitations & Ethical Considerations  
+⚠️ This model **may not always be accurate** and could generate **biased or misleading** responses. It is recommended to:  
+- **Validate outputs** before using them in real-world applications.  
+- **Avoid using for critical decision-making** (e.g., legal, medical, financial).  
+- **Improve it further** with domain-specific fine-tuning.  
 
-# Decode and print the result
-print("\n🧠 Model Output:")
-print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+---
 
-📌 Model Files
-File Name	Description
-pytorch_model.bin	The fine-tuned model weights
-config.json	Configuration for the model architecture
-tokenizer.json	Tokenizer used for the Macedonian language
-README.md	Documentation for the model
-.gitattributes	Git LFS tracking for large files
+## 🚀 How to Use the Model  
+You can load and run the model using **Hugging Face Transformers** in Python:
 
-📌 Training Details
-Dataset: Collected Macedonian texts (Wikipedia, news, government websites)
-Training Compute: GPU-based training on NVIDIA A100
-Training Time: Estimated XX hours
-Fine-tuned using: Hugging Face Transformers & PyTorch
+### **🔹 Load the Model for Inference**
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
 
-📌 Contributing
-MK-LLM-Mistral is open-source, and contributions are welcome! 🎯
+model_name = "ainowmk/MK-LLM-Mistral"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name)
 
-Open issues on GitHub
-Submit pull requests for improvements
-Join discussions on Hugging Face Community
-📩 For collaboration, reach out at: [email protected]
+input_text = "Која е главната цел на вештачката интелигенција?"
+inputs = tokenizer(input_text, return_tensors="pt")
+output = model.generate(**inputs, max_length=50)
 
-🚀 Let’s build the future of Macedonian AI together! 🇲🇰
+print(tokenizer.decode(output[0], skip_special_tokens=True))