Supernova TeraiLLM Photon-1 V5
A mathematical and scientific equation solver developed by Supernova.
Architecture
- Architecture: Multi-equation mathematical solver
- Equations: 6
- Solving directions: 18
- Trainable research parameters: 1
- Research component: Scalar calibration
- Pretrained models: None
- Language generation: None
Supported equations
| Equation | Formula | Directions |
|---|---|---|
| Einstein | E = mcยฒ | 3 |
| Speed | v = d/t | 3 |
| Density | ฯ = m/V | 3 |
| Newton | F = ma | 3 |
| Work | W = Fd | 3 |
| Power | P = W/t | 3 |
Total: 18 solving directions.
Features
- Automatic equation selection
- Variable recognition
- SI unit conversion
- Unit compatibility validation
- Invalid input rejection
- Numerical consistency checking
Development benchmarks
- Direction benchmark: 18,000/18,000
- Automatic routing: 18,000/18,000
- Invalid-input tests: 11/11
- Unit conversion tests: 6/6
- Functional tests: 7/7
Reported mean relative error: 8.57e-16 percent.
These are synthetic development benchmarks, not independent scientific certification.
Technical clarification
Photon-1 V5 uses an exact Python mathematical core. Its single trainable research parameter belongs to the scalar calibration component. The checkpoint stores metadata and calibration. Arithmetic is performed by the Python implementation.
Example usage
pip install supernova-photon
from Supernova_photon import v5
photon = v5.photon()
result = photon.solve(
"speed",
{
"distance": (100, "m"),
"time": (10, "s")
}
)
print(result)
Files
- photon1_v5.pt: PyTorch checkpoint
- config.json: Model configuration
- requirements.txt: Dependencies
Developer
Supernova
Project: Supernova TeraiLLM Photon-1 Version: V5
- Downloads last month
- 75
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support