Computational Lithography

Model Benchmarks

Reference numbers for the OPC/ILT baselines that ship with OpenLithoHub, evaluated on the synthetic-8 reference set. Submit your own model via the CLI — rankings are by Edge Placement Error (lower is better).

upload_file Submit via CLI

openlithohub leaderboard submit

bar_chart

At a Glance

EPE lower-is-better · DRC: pass / fail / n/a

EPE Mean by Model (nm)

0.00 0.13 0.27 0.40 0.53 dummy-identity: EPE 0.00 nm · 45nm · DRC fail 0.00 dummy-identity OpenILT: EPE 0.00 nm · 45nm · DRC fail 0.00 OpenILT Neural-ILT: EPE 0.00 nm · 45nm · DRC fail 0.00 Neural-ILT LevelSet-ILT: EPE 0.04 nm · 45nm · DRC fail 0.04 LevelSet-ILT rule-based-opc: EPE 0.53 nm · 45nm · DRC fail 0.53 rule-based-opc EPE Mean (nm)

Pareto Front · EPE Mean vs PV Band

0.00 0.14 0.28 0.42 0.56 0.00 1.30 2.59 3.89 5.19 dummy-identity: EPE 0.00 nm · PV 4.28 nm · 45nm dummy-identity OpenILT: EPE 0.00 nm · PV 4.28 nm · 45nm OpenILT Neural-ILT: EPE 0.00 nm · PV 4.28 nm · 45nm Neural-ILT LevelSet-ILT: EPE 0.04 nm · PV 4.25 nm · 45nm LevelSet-ILT rule-based-opc: EPE 0.53 nm · PV 4.94 nm · 45nm rule-based-opc EPE Mean (nm) PV Band (nm) Pareto front

Data source: src/data/leaderboard.json, regenerated from submissions/ in the OpenLithoHub repo. Charts render server-side as static SVG — no JS bundle, no third-party chart library.

Filters

Sort By

analytics Rankings
5 models
# Model Dataset Node EPE Mean (nm) EPE Max (nm) PV Band (nm) DRC Submitted
workspace_premium 01
dummy-identity manhattan
synthetic-8 45nm 0.00 0.00 4.28 Fail May 21, 2026
workspace_premium 02
OpenILT manhattan
synthetic-8 45nm 0.00 0.00 4.28 Fail May 21, 2026
workspace_premium 03
Neural-ILT manhattan
synthetic-8 45nm 0.00 0.00 4.28 Fail May 21, 2026
04
LevelSet-ILT manhattan
synthetic-8 45nm 0.04 0.25 4.25 Fail May 21, 2026
05
rule-based-opc manhattan
synthetic-8 45nm 0.53 1.41 4.94 Fail May 21, 2026

Submit Results

Use the CLI to submit your model's benchmark results to the public leaderboard.

openlithohub leaderboard submit --file results.json
terminal

Schema Reference

All submissions are validated against the BenchmarkResult schema with Pydantic.

VIEW DOCS arrow_forward