Magic Model 🪄

Fine-tuned language model for MMLU-style question answering.

Developed by Likhon Sheikh 🚀

Features

• ✅ Multi-safetensor support
• ✅ Fast tokenizer with tokenizer.json
• ✅ LoRA fine-tuning for efficiency
• ✅ MMLU-optimized responses
• ✅ Production-ready deployment

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("fariasultanacodes/magic")
tokenizer = AutoTokenizer.from_pretrained("fariasultanacodes/magic")

prompt = "Question: What is AI?\n\nAnswer:"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_length=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Pipeline Usage

from transformers import pipeline

generator = pipeline("text-generation", model="fariasultanacodes/magic")
result = generator("Question: Explain machine learning.\n\nAnswer:")
print(result[0]['generated_text'])

Model Details

• Base Model: Qwen/Qwen2.5-1.5B
• Fine-tuning: LoRA adapters
• Dataset: MMLU-style questions
• Format: Safetensors (multi-file support)
• Tokenizer: Fast tokenizer with JSON

Citation

@misc{magic-model-2025,
  title={Magic: MMLU-Optimized Language Model},
  author={Likhon Sheikh},
  year={2025},
  url={https://huggingface.co/fariasultanacodes/magic}
}

License

Apache-2.0

---

🚀 Developed by Likhon Sheikh