We currently set `pymbar<4` to avoid issues with poor free energy estimates (see https://github.com/choderalab/pymbar/issues/544). We should switch to `pymbar>=4` and switch to the robust solver to solve these issues -- see https://github.com/Psivant/femto/pull/34.
We currently set
pymbar<4to avoid issues with poor free energy estimates (see choderalab/pymbar#544).We should switch to
pymbar>=4and switch to the robust solver to solve these issues -- see Psivant/femto#34.