BloombergGPTOptions<T>
Options & Configuration · AiDotNet.Models.Options
Configuration options for BloombergGPT-style financial language model.
Inheritance: ModelOptions → BloombergGPTOptions<T>
For Beginners
BloombergGPT represents finance-specialized LLMs:
The Key Insight: General LLMs (GPT-3, GPT-4) lack deep financial knowledge. BloombergGPT-style models are trained on massive financial corpora to understand terminology, concepts, and relationships specific to finance.
What Problems Does BloombergGPT Solve?
- Financial sentiment analysis with high accuracy
- Named entity recognition for financial entities
- Financial question answering
- News classification and headline generation
- Understanding complex financial documents
Architecture Highlights:
- Large-scale decoder-only transformer
- Trained on mixed financial and general text
- Supports various financial NLP benchmarks
Key Benefits:
- Deep financial domain knowledge
- State-of-the-art on financial NLP tasks
- Understands complex financial terminology
- Can handle nuanced financial queries
How It Works
BloombergGPT-style models are large language models trained on extensive financial data including Bloomberg’s proprietary financial text corpus along with general text.
Reference: Wu et al., “BloombergGPT: A Large Language Model for Finance”, 2023.
Constructors
| Constructor | Summary |
|---|---|
BloombergGPTOptions() |
Initializes a new instance with default configuration. |
BloombergGPTOptions(BloombergGPTOptions<T> other) |
Initializes a new instance by copying from another instance. |
Properties
| Property | Summary |
|---|---|
double DropoutRate { get; set; } |
Dropout rate (default: 0.1). |
int HiddenDimension { get; set; } |
Hidden dimension (default: 1024 for medium model). |
int IntermediateDimension { get; set; } |
Intermediate feed-forward dimension (default: 4096). |
int MaxSequenceLength { get; set; } |
Maximum sequence length in tokens (default: 2048). |
int NumAttentionHeads { get; set; } |
Number of attention heads (default: 16). |
int NumClasses { get; set; } |
Number of output classes (default: 3). |
int NumLayers { get; set; } |
Number of transformer layers (default: 24). |
int VocabularySize { get; set; } |
Vocabulary size (default: 50257). |
string TaskType { get; set; } |
Task type: “classification”, “ner”, “qa”, “generation” (default: “classification”). |
Methods
| Method | Summary |
|---|---|
void Validate() |
Validates the BloombergGPT options. |