Well the specific example isn’t that important, I just wanted to show that PyMC can estimate parameters just fine after a highly non-linear transformation. As noted, you don’t typically want to use the smc sampler.
The pytensor.tensor.blas warning means you’re going to have performance issues. Based on what you posted the example is running fine, just slowly. How did you install PyMC? Did you follow the official installation guide? If you’re on a mac with an ARM64 chip, did you install accelerate BLAS?