sweatSmile commited on
Commit
4adbde8
·
verified ·
1 Parent(s): 8a70a20

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +85 -3
README.md CHANGED
@@ -1,3 +1,85 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: microsoft/DialoGPT-small
3
+ pipeline_tag: text-generation
4
+ library_name: transformers
5
+ tags:
6
+ - conversational-ai
7
+ - finance
8
+ - fintech
9
+ - wealth-management
10
+ - financial-advisor
11
+ - investment-advisory
12
+ - financial-planning
13
+ - lora
14
+ - private-banking
15
+ - portfolio-management
16
+ - financial-qa
17
+ - client-advisory
18
+ - robo-advisor
19
+ - financial-consultation
20
+ language:
21
+ - en
22
+ license: mit
23
+ datasets:
24
+ - ChanceFocus/flare-finqa
25
+ metrics:
26
+ - perplexity
27
+ - accuracy
28
+ widget:
29
+ - text: "<|user|> As my financial advisor, please help me understand: What is the impact of interest rate changes on my bond portfolio? <|bot|>"
30
+ example_title: "Interest Rate Risk Advisory"
31
+ - text: "<|user|> As my financial advisor, please help me understand: How should I diversify my investment portfolio for retirement planning? <|bot|>"
32
+ example_title: "Portfolio Diversification"
33
+ - text: "<|user|> As my financial advisor, please help me understand: What are the tax implications of selling my stocks this year? <|bot|>"
34
+ example_title: "Tax Planning Consultation"
35
+ ---
36
+
37
+ # DialoGPT-Financial-Wealth-Management-Advisor
38
+
39
+ Fine-tuned DialoGPT-small for financial advisory conversations, wealth management guidance, and comprehensive investment consultation services.
40
+
41
+ ## Overview
42
+ - **Base Model:** microsoft/DialoGPT-small (117M parameters)
43
+ - **Fine-tuning Method:** LoRA (4-bit quantization)
44
+ - **Dataset:** Financial Q&A dataset (1K expert-level samples)
45
+ - **Training:** 3 epochs with optimized hyperparameters
46
+
47
+ ## Key Features
48
+ - Comprehensive financial advisory consultations
49
+ - Investment portfolio analysis and recommendations
50
+ - Risk assessment and management strategies
51
+ - Tax planning and wealth optimization advice
52
+ - Retirement and financial planning guidance
53
+ - Client-focused conversational interface
54
+
55
+ ## Usage
56
+
57
+ ```python
58
+ from transformers import AutoTokenizer, AutoModelForCausalLM
59
+
60
+ model = AutoModelForCausalLM.from_pretrained("sweatSmile/DialoGPT-Financial-Wealth-Management-Advisor")
61
+ tokenizer = AutoTokenizer.from_pretrained("sweatSmile/DialoGPT-Financial-Wealth-Management-Advisor")
62
+
63
+ # Financial advisory consultation example
64
+ prompt = "<|user|> As my financial advisor, please help me understand: How do foreign currency fluctuations affect my international investments? <|bot|>"
65
+ inputs = tokenizer(prompt, return_tensors="pt")
66
+ outputs = model.generate(**inputs, max_new_tokens=200, pad_token_id=tokenizer.eos_token_id)
67
+ print(tokenizer.decode(outputs[0], skip_special_tokens=True))
68
+ ```
69
+
70
+ ## Applications
71
+ - Wealth management client consultations
72
+ - Investment advisory services automation
73
+ - Financial planning and retirement guidance
74
+ - Private banking client support
75
+ - Robo-advisor conversation engines
76
+ - Financial education and client onboarding
77
+
78
+ ## Training Details
79
+ - LoRA rank: 8, alpha: 16
80
+ - 4-bit NF4 quantization with fp16 precision
81
+ - Learning rate: 2e-4 with linear scheduling
82
+ - Batch size: 8, Max length: 320 tokens
83
+ - 3 epochs on curated financial advisory dataset
84
+
85
+ Optimized for sophisticated wealth management and investment advisory conversations in professional financial services environments.