YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
🌊 Torrential Attention — vanilla
Results
- Eval Loss: 5.6815
- Eval Perplexity: 293.39
- Parameters: 14,624,024
- Training Time: 986s
Architecture
Torrential Attention Mechanism with five pillars:
- Sigmoid Flood Attention (unnormalized, pressure-driven)
- Position-Decay Budgets (cascading energy)
- Cross-Layer Flood Thresholds (dam-break phase transitions)
- Cross-Head Spillover (tributary merging)
- Differential Noise Cancellation
- Downloads last month
- 2
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support