Closed tterava closed 1 year ago
Could you try a single image instead of a batch? If this works, I might know how to fix it.
Yes, using batch size of 1 produces no errors
batch size 1.... `Loading refiner...
85%|██████████████████████████████████████████████████████████████████████ | 35/41 [00:08<00:01, 4.58it/s]*** Error executing callback cfg_denoiser_callback for C:\Projects\Stable_Diffusion\Automatic1111\stable-diffusion-webui\extensions\sd-webui-refiner\scripts\refiner.py Traceback (most recent call last): File "C:\Projects\Stable_Diffusion\Automatic1111\stable-diffusion-webui\modules\script_callbacks.py", line 195, in cfg_denoiser_callback c.callback(params) File "C:\Projects\Stable_Diffusion\Automatic1111\stable-diffusion-webui\extensions\sd-webui-refiner\scripts\refiner.py", line 113, in denoiser_callback params.text_cond['vector'] = torch.cat((params.text_cond['vector'][:, :2304], self.c_ae), 1) IndexError: too many indices for tensor of dimension 3 `
@XeonG Refiner works only with SDXL.
I keep having this error on every iteration when using a refiner:
Settings used: