diff --git a/nerfstudio/configs/method_configs.py b/nerfstudio/configs/method_configs.py
index e2e3707f2..ac1d64cbb 100644
--- a/nerfstudio/configs/method_configs.py
+++ b/nerfstudio/configs/method_configs.py
@@ -107,8 +107,8 @@ method_configs["lerf"] = TrainerConfig(
     ),
     optimizers={
         "proposal_networks": {
-            "optimizer": RAdamOptimizerConfig(lr=1e-2, eps=1e-15),
-            "scheduler": ExponentialDecaySchedulerConfig(lr_final=1e-3, max_steps=30000),
+            "optimizer": AdamOptimizerConfig(lr=1e-2, eps=1e-15),
+            "scheduler": None,
         },
         "fields": {
             "optimizer": RAdamOptimizerConfig(lr=1e-2, eps=1e-15),
diff --git a/nerfstudio/fields/lerf_field.py b/nerfstudio/fields/lerf_field.py
index 705344574..bd4d3cf02 100644
--- a/nerfstudio/fields/lerf_field.py
+++ b/nerfstudio/fields/lerf_field.py
@@ -27,10 +27,13 @@ class LERFField(Field):
         super().__init__()
 
         self.spatial_distortion = spatial_distortion
-        self.clip_encs = torch.nn.ModuleList([
-            LERFField._get_encoding( 16, 128, 12, indim=3), 
-            LERFField._get_encoding(128, 512, 12, indim=3)
-            ])
+        self.clip_encs = torch.nn.ModuleList(
+            [
+                # LERFField._get_encoding(16, 512, 16, indim=3)
+                LERFField._get_encoding(16, 128, 10, indim=3),
+                LERFField._get_encoding(128, 1024, 13, indim=3),
+            ]
+        )
         tot_out_dims = sum([e.n_output_dims for e in self.clip_encs])
 
         self.clip_net = tcnn.Network(
diff --git a/nerfstudio/models/lerf.py b/nerfstudio/models/lerf.py
index dd7a17ac6..7dcd77646 100644
--- a/nerfstudio/models/lerf.py
+++ b/nerfstudio/models/lerf.py
@@ -42,7 +42,6 @@ class LERFModel(NerfactoModel):
 
         self.network = self.kwargs["network"]
         self.lerf_field = LERFField(clip_n_dims=self.network.clip_n_dims)
-        # self.lerf_sampler = PDFSampler(num_samples=self.config.num_lerf_samples, histogram_padding=0.00)
 
     def get_max_across(self, ray_samples, weights, hashgrid_field, scales_shape, preset_scales=None):
         # TODO smoothen this out
@@ -106,8 +105,12 @@ class LERFModel(NerfactoModel):
         lerf_field_outputs = self.lerf_field.get_outputs(lerf_samples, clip_scales)
 
         if self.training:
-            outputs['clip'] = self.renderer_clip(embeds=lerf_field_outputs[FieldHeadNames.CLIP], weights=lerf_weights.detach())
-            outputs['dino'] = self.renderer_mean(embeds=lerf_field_outputs[FieldHeadNames.DINO], weights=lerf_weights.detach())
+            outputs["clip"] = self.renderer_clip(
+                embeds=lerf_field_outputs[FieldHeadNames.CLIP], weights=lerf_weights.detach()
+            )
+            outputs["dino"] = self.renderer_mean(
+                embeds=lerf_field_outputs[FieldHeadNames.DINO], weights=lerf_weights.detach()
+            )
 
         if not self.training:
             max_across, best_scales = self.get_max_across(