diff --git a/comfy/ldm/models/autoencoder.py b/comfy/ldm/models/autoencoder.py
index bd698621c..fb6576ec7 100644
--- a/comfy/ldm/models/autoencoder.py
+++ b/comfy/ldm/models/autoencoder.py
@@ -53,6 +53,8 @@ class AutoencoderKL(torch.nn.Module):
         if path.lower().endswith(".safetensors"):
             import safetensors.torch
             sd = safetensors.torch.load_file(path, device="cpu")
+        elif path.lower().endswith(".pth") or path.lower().endswith(".pt"):
+            sd = torch.load(path, map_location='cpu')
         else:
             sd = torch.load(path, map_location="cpu")["state_dict"]
         keys = list(sd.keys())
diff --git a/comfy/sd.py b/comfy/sd.py
index 2d7ff5ab0..bfca15f88 100644
--- a/comfy/sd.py
+++ b/comfy/sd.py
@@ -767,15 +767,23 @@ def load_style_model(ckpt_path):
     return StyleModel(model)
 
 
-def load_clip(ckpt_path, embedding_directory=None):
-    clip_data = utils.load_torch_file(ckpt_path)
-    config = {}
-    if "text_model.encoder.layers.22.mlp.fc1.weight" in clip_data:
-        config['target'] = 'ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder'
+def load_clip(ckpt_path, version = None, embedding_directory=None):
+    if version is not None:
+        assert version in ("openai/clip-vit-large-patch14", )
+        config = {}
+        if version == "openai/clip-vit-large-patch14":
+            config['target'] = 'ldm.modules.encoders.modules.FrozenCLIPEmbedder'
+            config["params"] = {"textmodel_path": version}
+            clip = CLIP(config=config, embedding_directory=embedding_directory)
     else:
-        config['target'] = 'ldm.modules.encoders.modules.FrozenCLIPEmbedder'
-    clip = CLIP(config=config, embedding_directory=embedding_directory)
-    clip.load_from_state_dict(clip_data)
+        clip_data = utils.load_torch_file(ckpt_path)
+        config = {}
+        if "text_model.encoder.layers.22.mlp.fc1.weight" in clip_data:
+            config['target'] = 'ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder'
+        else:
+            config['target'] = 'ldm.modules.encoders.modules.FrozenCLIPEmbedder'
+        clip = CLIP(config=config, embedding_directory=embedding_directory)
+        clip.load_from_state_dict(clip_data)
     return clip
 
 def load_checkpoint(config_path, ckpt_path, output_vae=True, output_clip=True, embedding_directory=None):
diff --git a/nodes.py b/nodes.py
index ffbba9f94..b1af0871b 100644
--- a/nodes.py
+++ b/nodes.py
@@ -373,6 +373,20 @@ class CLIPLoader:
         clip = comfy.sd.load_clip(ckpt_path=clip_path, embedding_directory=folder_paths.get_folder_paths("embeddings"))
         return (clip,)
 
+class CLIPVersionLoader:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {"required": { "clip_version": (["openai/clip-vit-large-patch14"], ),
+                             }}
+    RETURN_TYPES = ("CLIP",)
+    FUNCTION = "load_clip"
+
+    CATEGORY = "loaders"
+
+    def load_clip(self, clip_version):
+        clip = comfy.sd.load_clip(ckpt_path=None, version=clip_version, embedding_directory=folder_paths.get_folder_paths("embeddings"))
+        return (clip,)
+
 class CLIPVisionLoader:
     @classmethod
     def INPUT_TYPES(s):
@@ -1065,6 +1079,7 @@ NODE_CLASS_MAPPINGS = {
     "LatentCrop": LatentCrop,
     "LoraLoader": LoraLoader,
     "CLIPLoader": CLIPLoader,
+    "CLIPVersionLoader": CLIPVersionLoader,
     "CLIPVisionEncode": CLIPVisionEncode,
     "StyleModelApply": StyleModelApply,
     "unCLIPConditioning": unCLIPConditioning,