Upload README.md with huggingface_hub

README.md

@@ -1,32 +1,39 @@
 ---
-license: apache-2.0
+base_model: mistralai/Mistral-7B-v0.1
+library_name: peft
 tags:
-- finance
-- tax
-- banking
-- investment
-- indian-finance
-- gst
-- conversational
+  - finance
+  - tax
+  - banking
+  - investment
+  - qlora
+  - mistral
+  - indian-finance
+  - gst
+license: apache-2.0
 ---
 
-# 🏦 Finxan
+# Finxan — Finance Domain LLM
 
-Finxan is an AI assistant specialized in Indian finance — covering GST, tax compliance, B2B invoicing, banking, and investments.
+Fine-tuned **Mistral-7B-v0.1** using QLoRA (4-bit) on Indian finance domain examples:
+- 🧾 GST / Tax compliance & calculations
+- 🏦 B2B Invoicing & banking
+- 📈 Investment advice
 
-## 💬 What can Finxan do?
+**Training details:**
+- Hardware: Google Colab T4 GPU (free)
+- Method: QLoRA 4-bit, LoRA r=16
+- Epochs: 3
+- Dataset: 110 finance instruction examples
 
-- 🧾 Calculate GST and verify tax compliance
-- 🏦 Create and manage B2B invoices
-- 📈 Answer investment and banking queries
-- ✅ Check financial document compliance
+## How to use
 
-## 🚀 Try it
 ```python
 from peft import PeftModel
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
 
+# Load base + adapter
 base = AutoModelForCausalLM.from_pretrained(
     "mistralai/Mistral-7B-v0.1",
     torch_dtype=torch.float16,
@@ -35,21 +42,16 @@ base = AutoModelForCausalLM.from_pretrained(
 model = PeftModel.from_pretrained(base, "webdpro/finxan")
 tokenizer = AutoTokenizer.from_pretrained("webdpro/finxan")
 
-def ask_finxan(instruction, input_text=""):
-    prompt = f"""### Instruction:\n{instruction}\n\n### Input:\n{input_text}\n\n### Response:"""
-    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
-    outputs = model.generate(**inputs, max_new_tokens=150, temperature=0.7, do_sample=True)
-    full = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    return full.split("### Response:")[-1].strip()
+# Run inference
+prompt = """### Instruction:
+Calculate GST
 
-print(ask_finxan("What is your name?"))
-print(ask_finxan("Calculate GST", "Invoice Rs 50000, GST 18%"))
-```
+### Input:
+Invoice amount Rs 50000, GST rate 18%
 
-## 📌 Example
+### Response:"""
 
-**You:** What is your name?
-**Finxan:** I am Finxan, your AI finance assistant. I can help you with GST, invoicing, tax compliance, and more.
-
-**You:** Calculate GST for invoice Rs 75000 at 18%
-**Finxan:** GST amount: Rs 13500. Total payable: Rs 88500.
+inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+outputs = model.generate(**inputs, max_new_tokens=100, temperature=0.7, do_sample=True)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+```