---
language:
- en
library_name: transformers
license: apache-2.0
tags:
- mlx
- mlx
base_model: mlx-community/SmolLM-1.7B-Instruct-8bit
datasets:
- dattaraj/pc-insurance-cost-estimator
---

# dattaraj/smol-lora-insurance-estimates

The Model [dattaraj/smol-lora-insurance-estimates](https://huggingface.co/dattaraj/smol-lora-insurance-estimates) was converted to MLX format from [mlx-community/SmolLM-1.7B-Instruct-8bit](https://huggingface.co/mlx-community/SmolLM-1.7B-Instruct-8bit) using mlx-lm version **0.19.1**.
This is a test to demonstrate the power of small langauge models. We take a SmoLM 1.7B model and fine-tune it on insurance estimation dataset available at: https://huggingface.co/datasets/dattaraj/pc-insurance-cost-estimator
The fine-tuned language model is now expert at taking text description of damage and generating cost estimation.

## Use with mlx

```bash
pip install mlx-lm
```

```python
from mlx_lm import load, generate

model, tokenizer = load("dattaraj/smol-lora-insurance-estimates")

prompt="hello"

if hasattr(tokenizer, "apply_chat_template") and tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, tokenize=False, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)
```