ComfyUI/comfy/text_encoders/long_clipl.py

31 lines
1.5 KiB
Python
Raw Normal View History

from comfy import sd1_clip
import os
class LongClipTokenizer_(sd1_clip.SDTokenizer):
def __init__(self, embedding_directory=None, tokenizer_data={}):
super().__init__(max_length=248, embedding_directory=embedding_directory, tokenizer_data=tokenizer_data)
class LongClipModel_(sd1_clip.SDClipModel):
def __init__(self, *args, **kwargs):
textmodel_json_config = os.path.join(os.path.dirname(os.path.realpath(__file__)), "long_clipl.json")
super().__init__(*args, textmodel_json_config=textmodel_json_config, **kwargs)
class LongClipTokenizer(sd1_clip.SD1Tokenizer):
def __init__(self, embedding_directory=None, tokenizer_data={}):
super().__init__(embedding_directory=embedding_directory, tokenizer_data=tokenizer_data, tokenizer=LongClipTokenizer_)
class LongClipModel(sd1_clip.SD1ClipModel):
def __init__(self, device="cpu", dtype=None, model_options={}, **kwargs):
super().__init__(device=device, dtype=dtype, model_options=model_options, clip_model=LongClipModel_, **kwargs)
def model_options_long_clip(sd, tokenizer_data, model_options):
w = sd.get("clip_l.text_model.embeddings.position_embedding.weight", None)
if w is None:
w = sd.get("text_model.embeddings.position_embedding.weight", None)
if w is not None and w.shape[0] == 248:
tokenizer_data = tokenizer_data.copy()
model_options = model_options.copy()
tokenizer_data["clip_l_tokenizer_class"] = LongClipTokenizer_
model_options["clip_l_class"] = LongClipModel_
return tokenizer_data, model_options