AttributeError: 'Tensor' object has no attribute 'last_hidden_state'

[ParityError]

Describe the bug
When attempting to generate from the web interface I'm presented with an AttributeError. The issue seems similar to #474 and #13.

To Reproduce
Steps to reproduce the behavior:

1. ./stable-diffusion-webui/run_webui_mac.sh
2. Click on Generate
3. AttributeError shows in the terminal output and generate does nothing
4. See error

Expected behavior
No error.

Desktop (please complete the following information):

• OS: macOS 12.6 (Apple M1)
• Safari Version 16.0 (17614.1.25.9.10, 17614)
• Commit revision [db4ab47f09450cbeb454a6e02698cb636a456878]

Additional context
Complete output of the error below; I've also included the lines you asked to be included from issue #474, which show in the output.

% ./stable-diffusion-webui/run_webui_mac.sh
WARNING: overwriting environment variables set in the machine
overwriting variable PYTORCH_ENABLE_MPS_FALLBACK
WARNING: overwriting environment variables set in the machine
overwriting variable PYTORCH_ENABLE_MPS_FALLBACK
error: cannot pull with rebase: You have unstaged changes.
error: please commit or stash them.
Warning: Taming Transformers not found at path /Users/Me/taming-transformers/taming
Warning: CodeFormer not found at path /Users/Me/CodeFormer/inference_codeformer.py
Warning: BLIP not found at path /Users/Me/BLIP/models/blip.py
Warning: LDSR not found at path /Users/Me/latent-diffusion/LDSR.py
Warning: caught exception 'Torch not compiled with CUDA enabled', memory monitor disabled
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
making attention of type 'vanilla' with 512 in_channels
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
making attention of type 'vanilla' with 512 in_channels
Loading weights [7460a6fa] from /Users/Me/stable-diffusion/stable-diffusion-webui/models/model.ckpt
Global Step: 470000
Model loaded.
Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.

txt2img: a nice dog
<class 'ldm.modules.encoders.modules.FrozenCLIPEmbedder'>
<class 'transformers.models.clip.modeling_clip.CLIPTextModel'>
<class 'torch.Tensor'>
Error completing request
Arguments: ('a nice dog', '', 'None', 'None', 20, 0, False, False, 1, 1, 7, -1.0, -1.0, 0, 0, 0, False, 512, 512, False, False, 0.7, 0, 1, '', 4, '', True, False, False, False, None, '') {}
Traceback (most recent call last):
  File "/Users/Me/stable-diffusion/stable-diffusion-webui/modules/ui.py", line 136, in f
    res = list(func(*args, **kwargs))
  File "/Users/Me/stable-diffusion/stable-diffusion-webui/webui.py", line 60, in f
    res = func(*args, **kwargs)
  File "/Users/Me/stable-diffusion/stable-diffusion-webui/modules/txt2img.py", line 41, in txt2img
    processed = process_images(p)
  File "/Users/Me/stable-diffusion/stable-diffusion-webui/modules/processing.py", line 333, in process_images
    uc = prompt_parser.get_learned_conditioning(len(prompts) * [p.negative_prompt], p.steps)
  File "/Users/Me/stable-diffusion/stable-diffusion-webui/modules/prompt_parser.py", line 104, in get_learned_conditioning
    conds = shared.sd_model.get_learned_conditioning(texts)
  File "/Users/Me/stable-diffusion/ldm/models/diffusion/ddpm.py", line 834, in get_learned_conditioning
    c = self.cond_stage_model(c)
  File "/Users/Me/.pyenv/versions/anaconda3-2022.05/envs/web-ui/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
    return forward_call(*input, **kwargs)
  File "/Users/Me/stable-diffusion/stable-diffusion-webui/modules/sd_hijack.py", line 378, in forward
    z = outputs.last_hidden_state
AttributeError: 'Tensor' object has no attribute 'last_hidden_state'

^CInterrupted with signal 2 in <frame at 0x2eaae8ec0, file '/Users/Me/.pyenv/versions/anaconda3-2022.05/envs/web-ui/lib/python3.10/site-packages/gradio/blocks.py', line 1238, code block_thread>

[dylancl]

Have you tried a clean reinstall? Also, I see that you've got uncommitted changes causing the git pull --rebase to fail - restore your changes and try running again.

[dylancl]

Actually, all these warnings point to an incomplete/failed install

Warning: Taming Transformers not found at path /Users/Me/taming-transformers/taming Warning: CodeFormer not found at path /Users/Me/CodeFormer/inference_codeformer.py Warning: BLIP not found at path /Users/Me/BLIP/models/blip.py Warning: LDSR not found at path /Users/Me/latent-diffusion/LDSR.py

[ParityError]

Have you tried a clean reinstall? Also, I see that you've got uncommitted changes causing the git pull --rebase to fail - restore your changes and try running again.

I have tried re-installing multiple times and encountered the same error each time. I thought it might be a case where something was not installed correctly or in the correct location, which it seems. Why would the taming transformers want to be installing two directories above the directory where my stable-diffusion-webui installation resides?

I did the installation with the mac_setup.sh script within a conda env. If this is not the route to go then please let me know, thanks.

[telans]

/Users/Me/taming-transformers/taming

What directory path & filesystem was the stable-diffusion-webui repo downloaded to? That path appears like the script doesn't know where it's running from. The path for that directory in my case is $HOME/[...]/stable-diffusion-webui/repositories/taming-transformers/taming

[ParityError]

/Users/Me/taming-transformers/taming

What directory path & filesystem was the stable-diffusion-webui repo downloaded to? That path appears like the script doesn't know where it's running from. The path for that directory in my case is $HOME/[...]/stable-diffusion-webui/repositories/taming-transformers/taming

I executed git clone in the directory /Users/Me/stable-diffusion/, after that there was a directory /Users/Me/stable-diffusion/stable-diffusion/web-ui within it. I then executed:

% ./stable-diffusion-webui/setup_mac.sh

If you need a complete install log I can provide it.