optics
Datasets
All datasets matching “optics”geometric_optics_physical_consistency_eval
Geometric Optics – Physical Consistency Evaluation Dataset
This dataset contains visual failure cases in geometric optics for multimodal image generation models.
Scope
The dataset focuses on physical and geometric inconsistencies related to:
mirror reflections (law of reflection)
refraction and dispersion in prisms
light ray direction consistency
shadow direction vs light source
camera–object–light spatial coherence
Motivation
Current image generation models… See the full description on the dataset page: https://huggingface.co/datasets/8Planetterraforming/geometric_optics_physical_consistency_eval.euv-projection-optics-overlay-failure-horizon-and-thermal-control-routing-v0.1
What this dataset tests
Predict when overlay will breach specand route the minimal thermal control response.
This is the third layer of projection optics thermo-mechanical decoupling:
overlay drift scoreoverlay error growthfailure horizon forecastingthermal control routing
Task
Given current tool state, output:
failure_horizon_wafersoverlay_risk_score (0..1)recommended_routing_actionminimal_fix_set
Output format:
float,float,string,string
Example:
55,0.67,coolant retune… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/euv-projection-optics-overlay-failure-horizon-and-thermal-control-routing-v0.1.euv-plasma-optics-coherence-drift-detection-v0.1
What this dataset tests
Detection of coherence drift between plasma source output and downstream EUV optical elements.
It focuses on early-stage decoupling between plasma energy, collector mirror behavior, and beam stability.
Core idea
Lithography yield collapses before alarms trigger.
The first signal is coherence decay between:
plasma powermirror reflectivitythermal loadbeam uniformityfocus stability
Required outputs
coherence state estimate
drift… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/euv-plasma-optics-coherence-drift-detection-v0.1.euv-projection-optics-deformation-overlay-coherence-drift-detection-v0.1
What this dataset tests
When thermo-mechanical deformation stops predicting overlay error cleanly.
This is the second layer of projection optics thermo-mechanical decoupling:
deformation modesoverlay error growthcause-axis attribution
Task
Given deformation modes and overlay state, predict:
overlay_drift_score (0..1)drift_flag (0/1)dominant_cause_axis
Output format:
float,int,string
Example:
0.58,1,cooling
Fields
Inputs
thermal_mode1_nm
thermal_mode2_nm… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/euv-projection-optics-deformation-overlay-coherence-drift-detection-v0.1.optics-syntheticeuv-projection-optics-absorption-deformation-coherence-baseline-mapping-v0.1
What this dataset tests
Baseline coherence between:
absorbed EUV power per projection mirrorcooling and thermal environmentresulting thermo-mechanical deformation modes
This is the first layer needed to forecast overlay drift.
Task
Given power and environment inputs, predict:
coherence_score
A scalar 0..1 describing how predictable deformation is from absorbed power and thermal conditions.
Inputs
m1_absorbed_power_w
m2_absorbed_power_w
m3_absorbed_power_w… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/euv-projection-optics-absorption-deformation-coherence-baseline-mapping-v0.1.