How to save trace in between sampling and then resume sampling later on?

Questions version agnostic
, ,
1

Hello everyone,

I would like to save the trace every n samples or so. Is there a way to do this in pymc? Specifically, I would like to do something like the following pseudo code:

with pm.Model() as model:

  ## priors 

  trace1 = pm.sample(3000, tune = 1000, chains = 3, random_seed=rng, 
                                            return_inferencedata=True)
  az.to_netcdf(trace1, 'TRACE1.nc')


  trace2 = pm.sample(4000, tune = 0, chains = 3, random_seed=rng, 
                                           return_inferencedata=True)
  az.to_netcdf(trace2, 'TRACE2.nc')

Following this, I would like to resume sampling in a separate script while being consistent with the previous definition of priors and model:

with pm.Model() as model:
  # Exact same priors and model

  # somehow resume from where trace2 stopped
  trace3 = pm.sample(5000, tune = 0, chains = 3, random_seed=rng, 
                                           return_inferencedata=True)  
  az.to_netcdf(trace3, 'TRACE3.nc')

Is this possible to do? If not, would it at least be possible to implement the first block?
Thank you for your time and help. I would really appreciate any guidance or comments.