metadata
language:
- en
tags:
- banking
- customer-support
- fintech
- small-language-model
- gpt2
license: apache-2.0
datasets:
- custom-banking-dataset
metrics:
- perplexity
- accuracy
- f1
base_model: microsoft/DialoGPT-medium
model_type: causal-lm
pipeline_tag: text-generation
library_name: transformers
widget:
- text: Hello! How can I help you with your credit card today?
🏦 Banking-SLM-v1
Small Language Model (SLM) for Banking Customer Support — fine-tuned from DialoGPT
This model is optimized for financial and banking domain conversations, including customer support automation, FAQs, and personalized query handling.
It can handle:
- Credit card and loan inquiries
- Account opening and transaction queries
- Policy explanation and chatbot-style interactions
🚀 Model Overview
| Property | Value |
|---|---|
| Base Model | microsoft/DialoGPT-medium |
| Fine-tuned Dataset | Custom Banking & FinTech dialogues |
| Framework | Hugging Face Transformers |
| Trained On | Azure GPU VM (NVIDIA T4) |
| Fine-tuning Epochs | ** |
| Learning Rate | 2e-4 |
| Tokenizer | AutoTokenizer |
| Total Parameters | ~355.61M |
| Trainable Parameters (LoRA) | ~** |
🧠 Use Cases
- Banking chatbots and conversational AI
- Automated customer support (loan, account, credit cards)
- FinTech knowledge assistants
- Smart reply systems for banking helpdesks
🧩 Example Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("techpro-saida/banking-slm-v1")
model = AutoModelForCausalLM.from_pretrained("techpro-saida/banking-slm-v1")
prompt = "Hi, can you tell me my credit card limit?"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100, temperature=0.7)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
#(or) use pipeline
from transformers import pipeline
slm = pipeline(
"text-generation",
model="./banking-slm-v1",
tokenizer=tokenizer,
max_new_tokens=60,
temperature=0.6,
top_p=0.9
)
prompt = "How do I reset my net-banking password?"
result = slm(prompt)
print(result[0]["generated_text"])