README.md

---
license: apache-2.0
datasets:
- leonvanbokhorst/tame-the-weights-personas
language:
- en
base_model:
- microsoft/Phi-4-mini-instruct
library_name: peft
---

# LoRA Adapter: captain_codebeard

This repository contains a LoRA (Low-Rank Adaptation) adapter for the base model `microsoft/Phi-4-mini-instruct`.

This adapter fine-tunes the base model to adopt the **captain_codebeard** persona.

Find the adapter files in this repository.

## Training Data

This adapter was fine-tuned on the `captain_codebeard` subset of the [leonvanbokhorst/tame-the-weights-personas](https://huggingface.co/datasets/leonvanbokhorst/tame-the-weights-personas) dataset.

## Usage (Example with PEFT)

```python
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base_model_id = "microsoft/Phi-4-mini-instruct"
adapter_repo_id = "leonvanbokhorst/microsoft-Phi-4-mini-instruct-captain_codebeard-adapter"

# Load the base model and tokenizer
model = AutoModelForCausalLM.from_pretrained(base_model_id)
tokenizer = AutoTokenizer.from_pretrained(base_model_id)

# Load the PEFT model
model = PeftModel.from_pretrained(model, adapter_repo_id)

# Now you can use the model for inference with the persona applied
# Example:
input_text = "Explain the concept of technical debt."
inputs = tokenizer(input_text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
```