---
base_model: unsloth/Llama-3.2-3B-Instruct
tags:
- text-generation-inference
- transformers
- unsloth
- llama
- trl
license: apache-2.0
language:
- en
---

# Job Skill Extractor - Fine-Tuned Llama Model


## Key Features

- **Fast Training and Inference**: Achieved 2x faster performance using Unsloth's techniques.
- **Base Model**: `unsloth/Llama-3.2-3B-Instruct`.
- **Language Support**: English.
- **License**: Apache 2.0.

---

## Intended Use

This model is designed to assist in:
- Extracting required job skills from job descriptions and titles.
- Automating job-skill matching for HR applications.
- Enabling intelligent job posting analysis in recruitment systems.

---

## Usage Example

Below is an example usage with the Unsloth library:

```python
from unsloth import FastLanguageModel

# Load model and tokenizer
model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "batuhanmtl/job-skill-extractor-llama3.2",
    max_seq_length = 8000,
    dtype = False,
    load_in_4bit = False,
)

# Enable faster inference
FastLanguageModel.for_inference(model)

# Define job description prompt
prompt_template = f"""
##### JOB TITLE #####
{job_title}

##### JOB DESCRIPTION #####
{job_description}
"""

# Tokenize input
messages = [
    {"role": "user", "content": prompt_template}
]
inputs = tokenizer.apply_chat_template(
    messages,
    tokenize = True,
    add_generation_prompt = True,
    return_tensors = "pt",
).to("cuda")

# Generate output
outputs = model.generate(
    input_ids=inputs,
    max_new_tokens=150,
    use_cache=True,
    temperature=1.5,
    min_p=0.1
)

start_token = "<|start_header_id|>assistant<|end_header_id|>"
end_token = "<|eot_id|>"

output = tokenizer.batch_decode(outputs)[0]

skill_list = output.split(start_token)[1].split(end_token)[0].strip()

print(skill_list)
```
---
## Model Overview

This fine-tuned Llama model, `batuhanmtl/job-skill-extractor-llama3.2`, is optimized for extracting relevant job skills from job titles and descriptions. It was trained 2x faster using [Unsloth](https://github.com/unslothai/unsloth) and Hugging Face's TRL (Transformers Reinforcement Learning) library. The model leverages the efficiency of the `unsloth/Llama-3.2-3B-Instruct` base model to provide fast and accurate text generation capabilities.

![Unsloth](https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png)

---