GGUF files lack sd version metadata, required for stable-diffusion.cpp
[ERROR] stable-diffusion.cpp:238 - get sd version from file failed: 'sd3.5_large-Q4_1.gguf'
Would it be possible to add the required metadata to these GGUF files, please?
I downloaded model and tested it, working here
got prompt
ggml_sd_loader:
1 546
3 377
model weight dtype torch.float16, manual cast: None
model_type FLOW
Requested to load SD3
loaded completely 9.5367431640625e+25 4925.2955322265625 True
100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [00:24<00:00, 1.21it/s]
Requested to load AutoencodingEngine
loaded completely 9.5367431640625e+25 159.87335777282715 True
100%|██████████████████████████████████████████████████████████████████████████████████| 12/12 [00:06<00:00, 1.95it/s]
Prompt executed in 42.08 seconds
I'm having the same problem. Any fix?
what software are you using to generate images?
workflow in images
I am trying to use stable-diffusion.cpp./sd -m ~/AI/models/stable-diffusion-3.5/sd3.5_large-Q4_1.gguf -p "mona lisa with a hat on"
uftah, I have to guess that it doesn't like the quant format of Q4_1. If I do a Q4_K it would be like 14GB in size making it way harder to load.