---
license: apache-2.0
base_model: Qwen/Qwen2.5-7B-Instruct
tags:
  - fine-tuned
  - qwen2.5
  - customer-support
  - unsloth
  - lora
pipeline_tag: text-generation
---

# maya-qwen-7b

Fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct)
for customer support conversations.

## Training Details

- **Method**: LoRA fine-tuning with Unsloth + TRL SFTTrainer
- **Base model**: Qwen/Qwen2.5-7B-Instruct
- **LoRA rank**: 16
- **Format**: ChatML

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("vivekmdrift/maya-qwen-7b")
tokenizer = AutoTokenizer.from_pretrained("vivekmdrift/maya-qwen-7b")

messages = [
    {"role": "system", "content": "You are a helpful customer support agent."},
    {"role": "user", "content": "How can I track my order?"},
]

text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
```