README.md

---
language: 
  - en
tags:
  - text-generation
  - transformers
  - finetuned
  - phi-4
  - lora
  - causal-lm
license: apache-2.0
datasets: custom
model-index:
  - name: mibera-v1-merged
    results: []
---

# 🏆 `mibera-v1-merged` 🏆
🚀 **Fine-tuned model based on `microsoft/phi-4` using LoRA adapters**

## 🔹 Model Details
- **Base Model**: `microsoft/phi-4`
- **Fine-tuned on**: Custom dataset
- **Architecture**: Transformer-based Causal LM
- **LoRA Adapter Merging**: ✅ Yes
- **Merged Model**: ✅ Ready for inference without adapters

## 📚 Training & Fine-tuning Details
- **Training Method**: Fine-tuning with **LoRA (Low-Rank Adaptation)**
- **LoRA Rank**: 32
- **Dataset**: Custom curated dataset (details not publicly available)
- **Training Library**: 🤗 Hugging Face `transformers` + `peft`

## 🚀 How to Use the Model
```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "ivxxdegen/mibera-v1-merged"

# Load tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_name)

# Load model
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")

print("✅ Model loaded successfully!")