stable-diffusion icon indicating copy to clipboard operation
stable-diffusion copied to clipboard

Error running scripts/inptaint.py (no streamlit)

Open ArielReplicate opened this issue 3 years ago • 0 comments

Hi, I'm trying to inpaint without streamlit using the scripts/inpaint.py but i get this error

Traceback (most recent call last):
  File "scripts/inpaint.py", line 83, in <module>
    c = model.cond_stage_model.encode(batch["masked_image"])
  File "/home/ariel/repos/stable_inpaint/ldm/modules/encoders/modules.py", line 162, in encode
    return self(text)
  File "/opt/conda/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/ariel/repos/stable_inpaint/ldm/modules/encoders/modules.py", line 154, in forward
    return_overflowing_tokens=False, padding="max_length", return_tensors="pt")
  File "/opt/conda/lib/python3.7/site-packages/transformers/tokenization_utils_base.py", line 2452, in __call__
    "text input must of type `str` (single example), `List[str]` (batch or single pretokenized example) "
ValueError: text input must of type `str` (single example), `List[str]` (batch or single pretokenized example) or `List[List[str]]` (batch of pretokenized examples).

Looking at the scripts/inpaint_st.pt i see some differences (For instance the inpatinting.py have no prompt used). I think the line

c = model.cond_stage_model.encode(batch["masked_image"]) should be c = model.get_first_stage_encoding(model.encode_first_stage(batch["masked_image"]))

but it gives other errors. Can you check if scripts/inpaint.pt works as should be?

ArielReplicate avatar Oct 29 '22 08:10 ArielReplicate