tmmdev/codellama-pattern-analysis

Pattern Analysis Model

A specialized financial pattern recognition model optimized for technical analysis in trading charts.

Model Description

This model identifies technical patterns in financial market data with high precision, focusing on:

• —Price action patterns
• —Chart formations
• —Support/Resistance levels
• —Trend analysis
• —Volume analysis

Model Details

• —Architecture: CodeLlama Base
• —Training Focus: Technical pattern recognition
• —Input: OHLCV data
• —Output: Pattern analysis with confidence scores
• —Context Window: 4096 tokens
• —Quantization: 8-bit
• —Memory Requirements: 8GB RAM minimum

Pattern Recognition Capabilities

• —Channel Patterns (>= 4 points)
• —Triangle Formations (>= 3 points)
• —Support/Resistance (>= 2 touches)
• —Double Top/Bottom (2% max deviation)
• —Trend Line Analysis
• —Volume Profile Analysis

Performance Metrics

• —Minimum Confidence Threshold: 0.8
• —Pattern Accuracy: >85%
• —False Signal Rejection: >90%

License

Apache-2.0