anoofx-labs/anovx

AnovX is an intelligent AI assistant fine-tuned specifically for supporting Sri Lankan students in learning, knowledge exploration, and syllabus-based study. This model has been fine-tuned on 250 Q&A pairs covering:

• General knowledge about AnovX and its capabilities
• Sri Lankan history, geography, culture, and leaders
• Science, Math, ICT, English, and religion-related syllabus questions
• Educational topics tailored for Grades 6–11 students

The model is designed to provide educational, clear, and concise answers in a friendly, student-focused tone.

---

Intended Use

This model can be used to:

• Answer student questions about AnovX, Sri Lanka, and syllabus topics
• Generate explanations and educational content for study purposes
• Support AI-assisted tutoring applications

Note: This model does not provide personal advice, medical guidance, or deep technical AI explanations. It is purely educational.

---

Model Details

• Base Model: [Insert your base model name here, e.g., Qwen-2.5-Omni]
• Fine-Tuning: LoRA/PEFT adapter, trained on 250 question-answer pairs
• Training Epochs: 7
• Dataset: Custom JSONL file containing questions about AnovX, Sri Lanka, and school syllabus topics

---

Files in This Repository

• adapter_model.safetensors – Fine-tuned LoRA weights
• adapter_config.json – LoRA configuration
• tokenizer.json – Tokenizer JSON
• tokenizer_config.json – Tokenizer configuration
• vocab.json – Vocabulary
• special_tokens_map.json – Special token mappings
• added_tokens.json – Custom tokens added during fine-tuning
• README.md – This file
• chat_template.jinja – Optional chat formatting template
• merges.txt – Tokenizer merge rules

---

How to Use

1. Load the model in Python

from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
from peft import PeftModel

# Load base model
base_model_name = "your-base-model-name"
tokenizer = AutoTokenizer.from_pretrained(base_model_name)
base_model = AutoModelForCausalLM.from_pretrained(base_model_name, device_map="auto")

# Load fine-tuned AnovX LoRA adapter
model = PeftModel.from_pretrained(base_model, "anoof/anovx-finetuned")

# Quick test
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer, device=0)
print(pipe("Who is AnovX?", max_new_tokens=100)[0]['generated_text'])

2. Example Prompts

• "Who is AnovX?"
• "Explain photosynthesis."
• "What is the capital of Sri Lanka?"
• "Generate a quiz question for Grade 8 Science."

---

Intended Audience

• Sri Lankan students
• Educators and teachers looking for interactive study aids
• Developers building educational AI tools

---

Limitations

• Small dataset fine-tuning: answers are best for predefined syllabus topics and AnovX knowledge.
• May not answer questions outside its training scope accurately.
• Avoid using for medical, legal, or sensitive personal advice.

---

License

[Insert your license, e.g., MIT, CC-BY-SA, or other]