arpitest
arpitest
I've got it working on OSX Mojave 10.14.6: makefile did not work at all, but compiling this way: $ cd src $ c++ -Dhidapi -Wall -O2 -std=gnu++11 -DVERSION=\"0.3.9\" main.cpp \*/\*.cpp...
File "/home/arpi/stable/GUI/stable-diffusion-webui/webui.py", line 40, in f res = func(*args, **kwargs) File "/home/arpi/stable/GUI/stable-diffusion-webui/webui.py", line 85, in shared.opts.onchange("sd_model_checkpoint", wrap_queued_call(lambda: modules.sd_models.reload_model_weights(shared.sd_model))) File "/home/arpi/stable/GUI/stable-diffusion-webui/modules/sd_models.py", line 262, in reload_model_weights load_model_weights(sd_model, checkpoint_info) File "/home/arpi/stable/GUI/stable-diffusion-webui/modules/sd_models.py", line 170,...
i'm still getting this size mismatch ERROR when switching checkpoint: RuntimeError: Error(s) in loading state_dict for LatentDiffusion: size mismatch for model.diffusion_model.input_blocks.0.0.weight: copying a param with shape torch.Size([320, 9, 3, 3])...
ok i've finaly found what is the problem: the filename does matter. after renaming inpaint-1.5.ckpt to sd-v1-5-inpainting.ckpt it loads. heh.
Undo at least the last step would be very helpful, now if it adds a wrong patch theis is no way back... also erasing (at least a rectangular block) would...
hmm, it's broken even if i set 512x512 as target size, or if i upload already upscaled 1024x1024 image...
> @arpitest is this still an issue? it is working now. even with xformers, it did not work a week ago yet.
i'm also seeing serious memory leaks on various systems, using pymilter (python 3.6). my application uses keras/tensorflow for neural network-based spam filtering, for long time i thought the leak comes...
I guess the memleak may be either in libmilter or the libmilter---python C connector. in theory pure python code cannot leak memory :) I could not reproduce memleaking by only...
> Running the milter-template.py under a testctx with 1000 iterations of test2 (a large message), there is no memory growth whatsoever. our production server processes 200000-500000 emails/day (big university), more...