7. Troubleshooting

This chapter covers common issues and their solutions.

7.3. Quick Fixes

Fit doesn’t converge:

  • Improve initial guess p0

  • Add bounds to constrain parameters

  • Use workflow='auto_global' for global search

Wrong results:

  • Check model function uses jax.numpy

  • Verify data is correct (plot it first)

  • Try different initial guess

Memory errors:

  • Reduce memory_limit_gb

  • Use streaming optimizer for large datasets

  • Close other applications

Slow performance:

  • Use GPU acceleration

  • Loosen tolerances (ftol=1e-6)

  • Start with data subset

7.4. Error Messages

Error

Solution

“p0 is required”

Provide initial parameter guess

“bounds required for auto_global”

Add bounds=([lower], [upper])

“Model must use jax.numpy”

Replace np with jnp in model

“Covariance could not be estimated”

Check model/data fit, improve p0

“Out of memory”

Reduce memory_limit_gb or data size