nerdyrodent/VQGAN-CLIP

-conf and -ckpt argument question | use other models

felixbro opened this issue · 1 comments

Hello there, i have a issue with the -conf and -ckpt arguments:
i would like to use another model but how do i use them?
i type this in the Anaconda Promt:
python generate.py -p "man by Peter Paul Rubens" -ii samples/fb.jpg -i 1200 -se 10 -o output/output6.png -conf checkpoints/faceshq.yaml -ckpt checkpoints/faceshq.ckpt

facehq.yaml and facehq.ckpt are in the checkpoints folder. i only get this error:

(vqgan) d:\AI_Images\VQGAN-CLIP>python generate.py -p "man by Peter Paul Rubens" -ii samples/fb.jpg -i 1200 -se 10 -o output/output6.png -conf checkpoints/faceshq.yaml -ckpt checkpoints/faceshq.ckpt
Working with z of shape (1, 256, 16, 16) = 65536 dimensions.
Traceback (most recent call last):
File "d:\AI_Images\VQGAN-CLIP\generate.py", line 546, in
model = load_vqgan_model(args.vqgan_config, args.vqgan_checkpoint).to(device)
File "d:\AI_Images\VQGAN-CLIP\generate.py", line 527, in load_vqgan_model
parent_model = cond_transformer.Net2NetTransformer(**config.model.params)
File "d:\AI_Images\VQGAN-CLIP\taming-transformers\taming\models\cond_transformer.py", line 41, in init
self.transformer = instantiate_from_config(config=transformer_config)
File "d:\AI_Images\VQGAN-CLIP\taming-transformers\main.py", line 119, in instantiate_from_config
return get_obj_from_str(config["target"])(**config.get("params", dict()))
File "d:\AI_Images\VQGAN-CLIP\taming-transformers\main.py", line 22, in get_obj_from_str
return getattr(importlib.import_module(module, package=None), cls)
File "C:\Users\fb\Anaconda3\envs\vqgan\lib\importlib_init_.py", line 127, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "", line 1030, in _gcd_import
File "", line 1007, in _find_and_load
File "", line 986, in _find_and_load_unlocked
File "", line 680, in _load_unlocked
File "", line 850, in exec_module
File "", line 228, in _call_with_frames_removed
File "d:\AI_Images\VQGAN-CLIP\taming-transformers\taming\modules\transformer\mingpt.py", line 17, in
from transformers import top_k_top_p_filtering
ModuleNotFoundError: No module named 'transformers'

i only have a D:\AI_Images\VQGAN-CLIP\taming-transformers\taming\modules\transformer folder, not transformers
what can i do?

I had the same issue for a range of different models. It was driving me crazy. The issue is the difference between the official python pip package for taming-transformers and the github repo. It's solved by installing the github version.

pip install transformers
pip uninstall taming-transformers
pip install git+https://github.com/CompVis/taming-transformers

This should work for:
Ade20k
CelebAHQ
Cin
Coco
Drin
FacesHQ
FFHQ
Gumbel
Imagenet_1024
Imagenet_16384
SFlckr
Wikiart

Originally posted by @GucciFlipFlops1917 in #106 (comment)