The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
feat(minimax): support prompt embeddings (#15697)
Silvere5a38e3f7b91
diff --git a/comfy/text_encoders/minimax.py b/comfy/text_encoders/minimax.pyindex c2dc47f7f..d79ccf0ea 100644--- a/comfy/text_encoders/minimax.py+++ b/comfy/text_encoders/minimax.py@@ -127,10 +127,6 @@ class MiniMaxH3Tokenizer(comfy.sd1_clip.SD1Tokenizer):tokenizer = lambda *a, **kw: Qwen3VLSDTokenizer(*a, **kw, embedding_size=5120, embedding_key="qwen3vl_32b")super().__init__(embedding_directory=embedding_directory, tokenizer_data=tokenizer_data, name="qwen3vl_32b", tokenizer=tokenizer)- def _text_ids(self, text):- tok = self.qwen3vl_32b.tokenizer- return tok(text, add_special_tokens=False)["input_ids"]-@staticmethoddef _vision_entry(data, video_block=False):emb = {"type": "image", "data": data, "original_type": "image"}@@ -143,7 +139,16 @@ class MiniMaxH3Tokenizer(comfy.sd1_clip.SD1Tokenizer):entries = []def add_text(s):- entries.extend((tid, 1.0) for tid in self._text_ids(s))+ if not s:+ return+ token_batches = self.qwen3vl_32b.tokenize_with_weights(+ s,+ return_word_ids=False,+ disable_weights=True,+ )+ if len(token_batches) != 1:+ raise ValueError("MiniMax H3 text segment exceeds the supported prompt length.")+ entries.extend(token_batches[0])def add_vision(data, video_block=False):entries.append((VISION_START, 1.0))