Merge pull request #91 from YosefLab/jhong/zuprior

z_u_prior false should remove the prior

src/mrvi/_module.py

@@ -407,18 +407,12 @@ def loss(
                 inference_outputs["qu"], generative_outputs["pu"]
             ).sum(-1)
         inference_outputs["qeps"]
+
+        kl_z = 0.0
         eps = inference_outputs["z"] - inference_outputs["z_base"]
         if self.z_u_prior:
             peps = dist.Normal(0, jnp.exp(self.pz_scale))
             kl_z = -peps.log_prob(eps).sum(-1)
-        else:
-            kl_z = (
-                -dist.Normal(inference_outputs["z_base"], jnp.exp(self.z_u_prior_scale))
-                .log_prob(inference_outputs["z"])
-                .sum(-1)
-                if self.z_u_prior_scale is not None
-                else 0
-            )
 
         weighted_kl_local = kl_weight * (kl_u + kl_z)
         loss = reconstruction_loss + weighted_kl_local