---
license: apache-2.0
language:
- en
tags:
- mistral
- transformers
- text-generation
- causal-lm
- conversational
pipeline_tag: text-generation
base_model: mistralai/Mistral-7B-Instruct-v0.2
widget:
- text: Hello, how can I help you today?
- text: Please explain how to reset my password.
datasets:
- Lakshan2003/customer_service_client_agent_conversations
model-index:
- name: mistral-customer-service-full
  results:
  - task:
      type: text-generation
      name: Text Generation
    dataset:
      name: Lakshan2003 Customer Service Conversations
      type: Lakshan2003/customer_service_client_agent_conversations
    metrics:
    - type: BLEU
      name: Validation BLEU
      value: 4.22
    - type: perplexity
      name: Validation Perplexity
      value: 8.94
---


# 🧠 Mistral-7B Customer Service Fine-tuned Model

This model is a fine-tuned version of `mistralai/Mistral-7B-Instruct-v0.2`
on a dataset of **customer service client-agent conversations** (Lakshan2003).

- Optimized for dialogue, polite response, and multi-turn coherence.
- Trained using LoRA (rank=16, alpha=32, dropout=0.05) with H2O LLM Studio.
- Dataset size: ~5k samples (balanced customer ↔ agent turns)

## Example Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "n0xgg04/mistral-customer-service-full"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype="bfloat16")

prompt = "Hi! Can you help me change my shipping address?"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=150)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))