---
license: apache-2.0
datasets:
- TrainingDataPro/customers-reviews-on-banks
language:
- en
base_model:
- unsloth/Qwen2.5-7B
pipeline_tag: text-generation
---
## Base Model
Unsloth implementation of Qwen2.5-7B: [unsloth/Qwen2.5-7B](https://huggingface.co/unsloth/Qwen2.5-7B)

## Finetune Method
Supervised fine tuning (SFT)

## Prompt Template

```
prompt_tmpl = """Below is a customer comment relating to their banking experience. \
Please output the banking aspects and their related sentiments expressed by the customer. \
Banking aspects must be short nouns or noun-phrases containing no more than 2 words that appear in the comment. \
Sentiments must be either positive, negative or neutral.

Output must follow the following format with NO explanations:
(credit card, positive)
(long queue, negative)
(app experience, neutral)

### Comment:
{comment}

### Response:
"""
```