AsrorAsr/qwen2.5-english-tutor-lora

Qwen2.5 English Tutor LoRA

A LoRA-fine-tuned version of Qwen2.5-0.5B designed to act as an English language tutor. Built especially with non-native English speakers in mind, including Uzbek learners.

📋 Model Description

This is a Parameter-Efficient Fine-Tuning (PEFT) adapter trained on top of Qwen2.5-0.5B using LoRA. It's optimized to provide friendly, patient, and educational responses for English learners.

🎯 Intended Use Cases

• Personalized English tutoring chatbot
• Grammar explanation and correction
• Vocabulary practice and examples
• Conversation practice for English learners
• Educational apps for language learning
• Foundation for further fine-tuning on language tasks

🚀 Quick Start

Install required libraries:

pip install transformers peft torch

Load the model with PEFT:

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base_model = "Qwen/Qwen2.5-0.5B"
adapter = "AsrorAsr/qwen2.5-english-tutor-lora"

tokenizer = AutoTokenizer.from_pretrained(base_model)
model = AutoModelForCausalLM.from_pretrained(base_model)
model = PeftModel.from_pretrained(model, adapter)

prompt = "Explain the difference between 'a' and 'the' for a beginner."
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=200)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

📊 Training Details

• Base model: Qwen2.5-0.5B (494M parameters)
• Method: LoRA (Low-Rank Adaptation)
• Framework: PEFT + Transformers + PyTorch
• Domain: English language education
• Target audience: Non-native English learners

🌐 Why This Model

English education is a global need, especially in Central Asia where access to native English speakers is limited. Small, efficient models like this can power offline or affordable tutoring tools.

⚠️ Limitations & Bias

• Small base model (0.5B) — limited reasoning depth
• Best for beginner to intermediate learners
• Not a replacement for human tutors or certified courses
• May occasionally produce incorrect grammar explanations
• Verify outputs for educational use

📜 License

Apache 2.0 — free for commercial and personal use.

🔗 Related Models

• AsrorAsr/qwen2.5-0.5b-uzbek-legal-lora — Uzbek legal domain LoRA

👤 Author

Asror Akmalov | AI/ML Engineer
🔗 GitHub
🔗 Hugging Face

Open to freelance projects and collaboration on LLM fine-tuning, RAG, and language education AI.

📬 Citation

@misc{akmalov2025englishtutor,
  author = {Akmalov, Asror},
  title = {Qwen2.5 English Tutor LoRA},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/AsrorAsr/qwen2.5-english-tutor-lora}
}