AttributeError in text_to_image.ipynb

[starkdg]

When running the text_to_image.ipynb, I get an Attribute Error.

AttributeError Traceback (most recent call last) in <cell line: 3>() 1 # SETUP MODELS & DATA 2 extras = core.setup_extras_pre() ----> 3 models = core.setup_models(extras) 4 models.generator.eval().requires_grad_(False) 5 print("STAGE C READY")

/gdrive/MyDrive/StableCascade/train/train_c.py in setup_models(self, extras) 161 generator.load_state_dict(load_or_fail(self.config.generator_checkpoint_path)) 162 else: --> 163 for param_name, param in load_or_fail(self.config.generator_checkpoint_path).items(): 164 set_module_tensor_to_device(generator, param_name, "cpu", value=param) 165 generator = generator.to(dtype).to(self.device)

AttributeError: 'NoneType' object has no attribute 'items'

[atomassoni]

this also happens in controlnet.ipynb

TypeError                                 Traceback (most recent call last)
[<ipython-input-12-af8515e91e5e>](https://localhost:8080/#) in <cell line: 11>()
      9 extras = core.setup_extras_pre()
     10 print(extras)
---> 11 models = core.setup_models(extras)
     12 models.generator.eval().requires_grad_(False)
     13 print("CONTROLNET READY")

[/content/StableCascade/train/train_c.py](https://localhost:8080/#) in setup_models(self, extras)
    126         effnet = EfficientNetEncoder()
    127         effnet_checkpoint = load_or_fail(self.config.effnet_checkpoint_path)
--> 128         effnet.load_state_dict(effnet_checkpoint if 'state_dict' not in effnet_checkpoint else effnet_checkpoint['state_dict'])
    129         effnet.eval().requires_grad_(False).to(self.device)
    130         del effnet_checkpoint

TypeError: argument of type 'NoneType' is not iterable

[danielstrenger]

I encounter the same problem and I guess, it arises from the model versions that we downloaded. I downloaded the float32 models (did you as well?), but in the file stage_3_3b.yaml, which is used in the notebook, there is the line generator_checkpoint_path: models/stage_c_bf16.safetensors which references a file that does not exist (because I downloaded the float32 version). I have tried to change the above line to generator_checkpoint_path: models/stage_c.safetensors which would now reference the correct file, but now the Jupyter kernel keeps dying when I try to execute the code cell for loading the models. (a similar problem should also arise for stage b)

[dome272]

This error indicates that the required model checkpoint is not existing. Have you downloaded all the checkpoints and put them to the correct path as indicated by the config files?

[Pranshu8694]

I tried downloading all checkpoints using the bash file, still I am facing the same issue.

[kenning]

for controlnet, it doesn't seem to work unless you download both of the big bfloat16 models. i hope that helps