The way forward is InferenceData so we didn’t offer the old MultiTrace for new samplers, to also reduce code complexity. Is it too much work to update the plotting functionality?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Specific model is slower in pymc that numpyro with same sampler | 1 | 506 | April 9, 2022 | |
| When can we just hit "the inference button"? | 1 | 697 | June 14, 2017 | |
| Need help for trace in jax_sampling | 12 | 794 | November 7, 2021 | |
| PyMC v4 not returning 'observed_data' in InferenceData object | 6 | 997 | March 30, 2022 | |
| Thoughts pipeline-style API for `sample()` | 2 | 739 | October 4, 2017 |