|
--- |
|
base_model: unsloth/llama-3.2-1b-bnb-4bit |
|
tags: |
|
- text-generation |
|
- qlora |
|
- llama |
|
- gguf |
|
- lyrics |
|
- taylor-swift |
|
license: apache-2.0 |
|
language: en |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
# Taylor Swift Lyrics Llama Model (3.2, GGUF Format) |
|
|
|
- **Base Model**: unsloth/llama-3.2-1b-bnb-4bit |
|
- **Fine-tuned on**: Taylor Swift lyrics using QLoRA |
|
- **Format**: GGUF (Not compatible with the `transformers` library; requires `llama-cpp-python`) |
|
- **License**: Apache-2.0 |
|
- **Developed by**: Covvenheimer and Team |
|
|
|
This model, fine-tuned on Taylor Swift lyrics, is tailored for generating text in the style of her songs. It was trained with a 2x speed improvement using [Unsloth](https://github.com/unslothai/unsloth) and Hugging Face’s TRL library, specifically optimized for GGUF format. |
|
|
|
> **Important:** This model requires `llama-cpp-python` to run. It is incompatible with the `transformers` library due to its GGUF format. |
|
|
|
## Installation and Setup |
|
|
|
To load and run this model, install the `llama-cpp-python` library and download the model file from the Hugging Face Hub. |
|
|
|
### Step-by-Step Code Example |
|
|
|
1. **Install llama-cpp-python** |
|
|
|
```python |
|
!pip install llama-cpp-python |
|
``` |
|
|
|
2. **Load the Model with llama-cpp** |
|
|
|
```python |
|
from huggingface_hub import hf_hub_download |
|
from llama_cpp import Llama |
|
|
|
# Define your model repository and file name |
|
repo_id = "Covvenheimer/taylor_swift_model" |
|
filename = "unsloth.Q4_K_M.gguf" |
|
|
|
# Download the GGUF model file from Hugging Face |
|
model_path = hf_hub_download(repo_id=repo_id, filename=filename) |
|
|
|
# Load the model using llama-cpp-python |
|
llm = Llama(model_path=model_path) |
|
``` |
|
|
|
3. **Generate Text Using a Prompt** |
|
|
|
```python |
|
# Define a prompt for generating lyrics |
|
prompt = """You are a songwriter composing a song in the style of Taylor Swift. |
|
Write lyrics that reflect her themes and musical style, focusing on Love, Heartbreak, Romance.""" |
|
|
|
# Generate lyrics |
|
output = llm(prompt, max_tokens=512, temperature=0.8) |
|
print(output["choices"][0]["text"]) |
|
``` |
|
|
|
> **Tip:** To customize the themes, change the last line of the prompt. For example, you can modify the focus to other themes like "Friendship, Growth, Resilience." Replace `"Love, Heartbreak, Romance"` with your chosen themes. |
|
|
|
This setup will allow you to use the model efficiently and generate lyrics in the style of Taylor Swift. |
|
|
|
--- |
|
|
|
# Uploaded model |
|
|
|
- **Developed by:** Covvenheimer |
|
- **License:** apache-2.0 |
|
- **Finetuned from model :** unsloth/llama-3.2-1b-bnb-4bit |
|
|
|
This llama model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library. |
|
|
|
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth) |
|
|
