agentsclimarketplace

Run3 calibrate glm

Skill cxcscmu/SkillLearnBench/skills/b3-teacher-feedback-gemini-3.1-flash-lite-preview/temperature-simulation/run3_calibrate_glm

[COLM'26] SkillLearnBench is the first benchmark for evaluating continual learning methods that automatically generate agent skills.

Install
npx -y skills add cxcscmu/SkillLearnBench --skill run3_calibrate_glm

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Performs efficient parameter optimization using a heuristic search or scipy.optimize to satisfy RMSE constraints within fixed parameter ranges.

SKILL.md

0.9 KB, 262 tokens by cl100k_base, as published. Nobody here has run it

Since a full grid search is computationally expensive, implement an optimization routine (e.g., scipy.optimize.minimize with bounds) to minimize the overall_rmse. Ensure all parameters stay within defined constraints: Kw [0.1, 0.5], coef_mix_hyp [0.3, 0.7], wind_factor [0.7, 1.3], lw_factor [0.7, 1.3], ch [0.0005, 0.002].

from scipy.optimize import minimize

def objective_function(x):
    params = {'Kw': x[0], 'coef_mix_hyp': x[1], 'wind_factor': x[2], 'lw_factor': x[3], 'ch': x[4]}
    update_glm_parameters(params)
    run_simulation()
    metrics = calculate_metrics(...)
    return metrics['overall_rmse']

# Use Bounds to enforce calibration ranges
bounds = [(0.1, 0.5), (0.3, 0.7), (0.7, 1.3), (0.7, 1.3), (0.0005, 0.002)]

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.