setting a batch size greater than 1 causes the following error response
*** Error executing callback cfg_denoiser_callback for /home/djabacus/stable-diffusion-webui/extensions/sd-webui-refiner/scripts/refiner.py
Traceback (most recent call last):
File "/home/djabacus/stable-diffusion-webui/modules/script_callbacks.py", line 195, in cfg_denoiser_callback
c.callback(params)
File "/home/djabacus/stable-diffusion-webui/extensions/sd-webui-refiner/scripts/refiner.py", line 110, in denoiser_callback
params.text_cond['vector'] = torch.cat((params.text_cond['vector'][:, :2304], self.c_ae), 1)
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 3 but got size 1 for tensor number 1 in the list.
(for a batch size of 3, for example) on each refiner step, and the refiner is visibly not applied after comparing with a generation of the same seed simply with the refiner extension disabled
hi there :)
setting a batch size greater than 1 causes the following error response
(for a batch size of 3, for example) on each refiner step, and the refiner is visibly not applied after comparing with a generation of the same seed simply with the refiner extension disabled
the regression seems to have been introduced in https://github.com/wcde/sd-webui-refiner/commit/f0b9713218c9b99c01bb5b8a5bb2b3f15e1db392 but i genuinely don't know anything about the underlying tensor stuff :( i tried trivially changing the
1
at the end of the.cat()
toparams.text_cond['vector'].shape[0]
but that returnedthanks!