diff --git a/src/ae.py b/src/ae.py
index 479d1e6..577d765 100644
--- a/src/ae.py
+++ b/src/ae.py
@@ -40,7 +40,10 @@ def from_pretrained(cls, model_uri: str, **kwargs):
         from safetensors.torch import load_file
         from .ae_nn import AutoEncoder
 
-        base = pathlib.Path(huggingface_hub.snapshot_download(model_uri))
+        try:
+            base = pathlib.Path(huggingface_hub.snapshot_download(model_uri))
+        except Exception:
+            base = pathlib.Path(model_uri)
 
         enc_cfg = OmegaConf.load(base / "encoder_conf.yml").model
         dec_cfg = OmegaConf.load(base / "decoder_conf.yml").model
diff --git a/src/world_engine.py b/src/world_engine.py
index 38eea42..77b6969 100644
--- a/src/world_engine.py
+++ b/src/world_engine.py
@@ -36,6 +36,7 @@ def __init__(
         """
         model_uri: HF URI or local folder containing model.safetensors and config.yaml
         quant: None | w8a8 | nvfp4
+        model_config_overrides: Dict to override model config values
         """
         self.device, self.dtype = device, dtype
 
@@ -49,7 +50,8 @@ def __init__(
 
         self.prompt_encoder = None
         if self.model_cfg.prompt_conditioning is not None:
-            self.prompt_encoder = PromptEncoder("google/umt5-xl", dtype=dtype).to(device).eval()  # TODO: dont hardcode
+            pe_uri = getattr(self.model_cfg, "prompt_encoder_uri", "google/umt5-xl")
+            self.prompt_encoder = PromptEncoder(pe_uri, dtype=dtype).to(device).eval()
 
         self.model = WorldModel.from_pretrained(model_uri, cfg=self.model_cfg).to(device=device, dtype=dtype).eval()
         apply_inference_patches(self.model)