Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

Minimax-H3: Add missing special tokens (#15808)

Jukka Seppänen924743af083c
diff --git a/comfy/text_encoders/minimax.py b/comfy/text_encoders/minimax.py
index d79ccf0ea..b248e70c7 100644
--- a/comfy/text_encoders/minimax.py
+++ b/comfy/text_encoders/minimax.py
@@ -28,6 +28,10 @@ from .qwen3vl import Qwen3VL, Qwen3VLSDTokenizer
VISION_START = 151652
VISION_END = 151653
+# FL2VA/Ref2VA tokenizer_config extends Qwen with these, ids fixed by the released tokenizer
+MINIMAX_EXTRA_TOKENS = {"<d>": 151669, "</d>": 151670, "<|cutoff|>": 151671,
+ "<|lyrics_start|>": 151672, "<|lyrics_end|>": 151673,
+ "<|caption_start|>": 151674, "<|caption_end|>": 151675}
QWEN_IMAGE_MEAN = [0.5, 0.5, 0.5]
QWEN_IMAGE_STD = [0.5, 0.5, 0.5]
@@ -122,9 +126,16 @@ class MiniMaxH3TEModel(comfy.sd1_clip.SD1ClipModel):
clip_model=MiniMaxH3ClipModel, model_options=model_options)
+class MiniMaxQwenSDTokenizer(Qwen3VLSDTokenizer):
+ def __init__(self, *args, **kwargs):
+ super().__init__(*args, **kwargs)
+ self.tokenizer.add_special_tokens({"additional_special_tokens": list(MINIMAX_EXTRA_TOKENS)})
+ self.inv_vocab = {v: k for k, v in self.tokenizer.get_vocab().items()}
+
+
class MiniMaxH3Tokenizer(comfy.sd1_clip.SD1Tokenizer):
def __init__(self, embedding_directory=None, tokenizer_data={}):
- tokenizer = lambda *a, **kw: Qwen3VLSDTokenizer(*a, **kw, embedding_size=5120, embedding_key="qwen3vl_32b")
+ tokenizer = lambda *a, **kw: MiniMaxQwenSDTokenizer(*a, **kw, embedding_size=5120, embedding_key="qwen3vl_32b")
super().__init__(embedding_directory=embedding_directory, tokenizer_data=tokenizer_data, name="qwen3vl_32b", tokenizer=tokenizer)
@staticmethod