ensure backwards compat with optional args

Author: BlenderNeko

comfy/sd.py

@@ -372,12 +372,16 @@ def add_patches(self, patches, strength=1.0):
     def clip_layer(self, layer_idx):
         self.layer_idx = layer_idx
 
-    def tokenize(self, text):
-        return self.tokenizer.tokenize_with_weights(text)
+    def tokenize(self, text, return_word_ids=False):
+        return self.tokenizer.tokenize_with_weights(text, return_word_ids)
 
-    def encode(self, tokens):
+    def encode(self, text, from_tokens=False):
         if self.layer_idx is not None:
             self.cond_stage_model.clip_layer(self.layer_idx)
+        if from_tokens:
+            tokens = text
+        else:
+            tokens = self.tokenizer.tokenize_with_weights(text)
         try:
             self.patcher.patch_model()
             cond = self.cond_stage_model.encode_token_weights(tokens)

comfy/sd1_clip.py

@@ -240,7 +240,7 @@ def _try_get_embedding(self, embedding_name:str):
         return (embed, "")
 
 
-    def tokenize_with_weights(self, text:str):
+    def tokenize_with_weights(self, text:str, return_word_ids=False):
         '''
         Takes a prompt and converts it to a list of (token, weight, word id) elements.
         Tokens can both be integer tokens and pre computed CLIP tensors.
@@ -301,6 +301,10 @@ def tokenize_with_weights(self, text:str):
 
         #add start and end tokens
         batched_tokens = [[(self.start_token, 1.0, 0)] + x + [(self.end_token, 1.0, 0)] for x in batched_tokens]
+
+        if not return_word_ids:
+            batched_tokens = [[(t, w) for t, w,_ in x] for x in batched_tokens]
+
         return batched_tokens
 
 

nodes.py

@@ -44,8 +44,7 @@ def INPUT_TYPES(s):
     CATEGORY = "conditioning"
 
     def encode(self, clip, text):
-        tokens = clip.tokenize(text)
-        return ([[clip.encode(tokens), {}]], )
+        return ([[clip.encode(text), {}]], )
 
 class ConditioningCombine:
     @classmethod