AttributeError when using Wan2.2_LowNoise for Text-to-Image Processing with umt5_xxl

[30SoD]

When the formatter.py gets the value from WanT5Tokenizer, attribute error pops up.

The reason behind it is because WanT5Tokenizer doesnt have clip_l attribute.
DualClipLoader will not work when using Wan Model to generate the image.

Since WanT5Tokenizer uses the base SD1Tokenizer as the schematic, i edited the "formatter.py" file

if clip_ is not None:
    tokenizer = clip_.tokenizer
    if isinstance(tokenizer, SD1Tokenizer):

# Edited Code Section Start
        try:
            clip = tokenizer.clip_l
        except AttributeError:
            clip = None
# Edited Code Section End

    elif isinstance(tokenizer, SD2Tokenizer):
        clip = tokenizer.clip_h
    elif isinstance(tokenizer, SDXLTokenizer):
        clip = tokenizer.clip_l
    elif isinstance(tokenizer, SD3Tokenizer):
        clip = tokenizer.clip_l
    elif isinstance(tokenizer, FluxTokenizer):
        clip = tokenizer.clip_l

I could have done the following alternate fix too
Anyways both work on same logic

if clip_ is not None:
    tokenizer = clip_.tokenizer

# Edited Code Section Start
    if isinstance(tokenizer, SD1Tokenizer) and hasattr(tokenizer, "clip_l"):
        clip = tokenizer.clip_l
    elif isinstance(tokenizer, SD1Tokenizer) and not hasattr(tokenizer, "clip_l"):
        clip = None
# Edited Code Section End

    elif isinstance(tokenizer, SD2Tokenizer):
        clip = tokenizer.clip_h
    elif isinstance(tokenizer, SDXLTokenizer):
        clip = tokenizer.clip_l
    elif isinstance(tokenizer, SD3Tokenizer):
        clip = tokenizer.clip_l
    elif isinstance(tokenizer, FluxTokenizer):
        clip = tokenizer.clip_l

I did this temporary fix for the things to work, Hoping for an earlier update on this issue.

Thanks & Regards