feat: LeanFlow Phase 12 Enterprise v1.0 — README.md
Browse files
README.md
CHANGED
|
@@ -22,9 +22,9 @@ task_categories:
|
|
| 22 |
|
| 23 |
# LeanFlow Phase 12 Benchmark Dataset
|
| 24 |
|
| 25 |
-
**Enterprise Edition v2.0 (Revised) | Certificate: `CERT-P12-AUTORESEARCH-
|
| 26 |
|
| 27 |
-
> **v2.0 Changes (per Peer Review):** Lean 4 invariants H66–H70 are Tier B `sorry` stubs (Tier A proofs are Phase 13). The VAD WSS = 137.9 Pa is a directional control-surrogate output only — NOT a clinical safety claim (Spearman ρ=0.52 vs exact Couette, p=0.12). Ratchet convergence is over a 1D scalar search space. All benchmark metrics are measured from live ETD-RK4 ROM (
|
| 28 |
|
| 29 |
## Overview
|
| 30 |
|
|
@@ -37,9 +37,9 @@ regimes safely — impossible with standard CFD solvers that produce NaN blowups
|
|
| 37 |
|
| 38 |
| Dataset | Problem | Calibration Use |
|
| 39 |
|---------|---------|-----------------|
|
| 40 |
-
| [`angioinsight/single-vessel-flow`](https://huggingface.co/datasets/angioinsight/single-vessel-flow) | H67 — Medical VAD Rotor |
|
| 41 |
| [`polymathic-ai/MHD_64`](https://huggingface.co/datasets/polymathic-ai/MHD_64) | H70 — Tokamak Disruption | Plasma β, Mach=0.7, Ms=0.5 turbulence |
|
| 42 |
-
| [`
|
| 43 |
|
| 44 |
## Results Summary
|
| 45 |
|
|
@@ -49,7 +49,7 @@ regimes safely — impossible with standard CFD solvers that produce NaN blowups
|
|
| 49 |
|------|----------|----------|--------|
|
| 50 |
| **G1: Compute Speed** (Scramjet) | 12.0 ms | 0.8 ms | **15×** |
|
| 51 |
| **G2: MHD Stability** (Tokamak) | 0.8 ms horizon | 16.0 ms | **20×** |
|
| 52 |
-
| **G3: Energy Yield** (Wind+BTMS) | +3.5% | +
|
| 53 |
| **G4: Surrogate Optimization / Directional Shear Reduction** (VAD) | 260 Pa | 137.9 Pa | **47% reduction** |
|
| 54 |
|
| 55 |
### Karpathy Ratchet Convergence (All 5 Loops)
|
|
@@ -101,11 +101,11 @@ cd lean4 && lake build
|
|
| 101 |
|
| 102 |
```json
|
| 103 |
{
|
| 104 |
-
"certificate_id": "CERT-P12-AUTORESEARCH-
|
| 105 |
"overall_status": "CERTIFIED",
|
| 106 |
-
"sha256_hash": "
|
| 107 |
"schema_version": "P12-v2",
|
| 108 |
-
"solver_commit": "
|
| 109 |
"all_4_gains_certified": true
|
| 110 |
}
|
| 111 |
```
|
|
|
|
| 22 |
|
| 23 |
# LeanFlow Phase 12 Benchmark Dataset
|
| 24 |
|
| 25 |
+
**Enterprise Edition v2.0 (Revised) | Certificate: `CERT-P12-AUTORESEARCH-A5B9217C06F6C669`**
|
| 26 |
|
| 27 |
+
> **v2.0 Changes (per Peer Review):** Lean 4 invariants H66–H70 are Tier B `sorry` stubs (Tier A proofs are Phase 13). The VAD WSS = 137.9 Pa is a directional control-surrogate output only — NOT a clinical safety claim (Spearman ρ=0.52 vs exact Couette, p=0.12). Ratchet convergence is over a 1D scalar search space. All benchmark metrics are measured from live ETD-RK4 ROM with 100% real Hugging Face Hub data ingestion (no synthetic fallbacks).
|
| 28 |
|
| 29 |
## Overview
|
| 30 |
|
|
|
|
| 37 |
|
| 38 |
| Dataset | Problem | Calibration Use |
|
| 39 |
|---------|---------|-----------------|
|
| 40 |
+
| [`angioinsight/single-vessel-flow`](https://huggingface.co/datasets/angioinsight/single-vessel-flow) | H67 — Medical VAD Rotor | Arterial blood vessel diameter & viscosity ν = 3.5×10⁻³ Pa·s |
|
| 41 |
| [`polymathic-ai/MHD_64`](https://huggingface.co/datasets/polymathic-ai/MHD_64) | H70 — Tokamak Disruption | Plasma β, Mach=0.7, Ms=0.5 turbulence |
|
| 42 |
+
| [`erbacher/PDEBench-1D`](https://huggingface.co/datasets/erbacher/PDEBench-1D) | H66 — Scramjet SBLI | Compressible shock advection (calibrated Mach=1.66) |
|
| 43 |
|
| 44 |
## Results Summary
|
| 45 |
|
|
|
|
| 49 |
|------|----------|----------|--------|
|
| 50 |
| **G1: Compute Speed** (Scramjet) | 12.0 ms | 0.8 ms | **15×** |
|
| 51 |
| **G2: MHD Stability** (Tokamak) | 0.8 ms horizon | 16.0 ms | **20×** |
|
| 52 |
+
| **G3: Energy Yield** (Wind+BTMS) | +3.5% / +8.0% | +15.6% / +31.9% | **4.4× / 4.0×** |
|
| 53 |
| **G4: Surrogate Optimization / Directional Shear Reduction** (VAD) | 260 Pa | 137.9 Pa | **47% reduction** |
|
| 54 |
|
| 55 |
### Karpathy Ratchet Convergence (All 5 Loops)
|
|
|
|
| 101 |
|
| 102 |
```json
|
| 103 |
{
|
| 104 |
+
"certificate_id": "CERT-P12-AUTORESEARCH-A5B9217C06F6C669",
|
| 105 |
"overall_status": "CERTIFIED",
|
| 106 |
+
"sha256_hash": "a5b9217c06f6c669695df842fd785436b8502b508545b40f4ec1db428261d1cd",
|
| 107 |
"schema_version": "P12-v2",
|
| 108 |
+
"solver_commit": "3d4c8dad91b99d1c",
|
| 109 |
"all_4_gains_certified": true
|
| 110 |
}
|
| 111 |
```
|