$ python app.py
Model directory '/home/zzz/ai/software/Pyramid-Flow/pyramid_flow_model' already exists. Skipping download.
using half precision
Using temporal causal attention
We interp the position embedding of condition latents
Traceback (most recent call last):
File "app.py", line 47, in <module>
model = PyramidDiTForVideoGeneration(
File "/home/zzz/ai/software/Pyramid-Flow/pyramid_dit/pyramid_dit_for_video_gen_pipeline.py", line 99, in __init__
self.text_encoder = SD3TextEncoderWithMask(model_path, torch_dtype=torch_dtype)
File "/home/zzz/ai/software/Pyramid-Flow/pyramid_dit/modeling_text_encoder.py", line 19, in __init__
self.tokenizer = CLIPTokenizer.from_pretrained(os.path.join(model_path, 'tokenizer'))
File "/home/zzz/ai/software/miniconda3/envs/pyramid/lib/python3.8/site-packages/transformers/tokenization_utils_base.py", line 2070, in from_pretrained
raise EnvironmentError(
OSError: Can't load tokenizer for '/home/zzz/ai/software/Pyramid-Flow/pyramid_flow_model/tokenizer'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure '/home/zzz/ai/software/Pyramid-Flow/pyramid_flow_model/tokenizer' is the correct path to a directory containing all relevant files for a CLIPTokenizer tokenizer.
And the contents of the folder in question:
$ ls -l pyramid_flow_model/tokenizer
total 1560
-rw-rw-r-- 1 zzz zzz 524619 Oct 11 22:15 tokenizer_merges.txt
-rw-rw-r-- 1 zzz zzz 588 Oct 11 22:15 tokenizer_special_tokens_map.json
-rw-rw-r-- 1 zzz zzz 705 Oct 11 22:15 tokenizer_tokenizer_config.json
-rw-rw-r-- 1 zzz zzz 1059962 Oct 11 22:15 tokenizer_vocab.json
Here's the full error:
And the contents of the folder in question: