Hidden Gem Alert: Stable Diffusion Model & LoRA of the Week - Episode 1
HTML-код
- Опубликовано: 22 апр 2024
- "Discover unique Stable Diffusion models and LoRAs each week! Expand your AI art toolkit with fresh styles and hidden gems. #stablediffusion #lora #aiart"
LORA Link : civitai.com/models/176555/har...
Dice's Dream Diffusion Model link : civitai.com/models/288400?mod...
Install Stable Diffusion Forge Web UI: • Use Forge UI for 60% F...
Inpainting Tutorials Playlist : • Stable diffusion inpai...
Stable Diffusion Tutorials Playlist : • Stable Diffusion tutor...
I have NVIDIA GTX 1650 with 4GB VRAM and have installed Forge UI. SD 1.5 works fine but SDXL won't run on it. What can I do?
What's your RAM size? And what error are you getting?
@@theinformativerealm 8.00 GB (7.84 GB usable) and 4GB VRAM.
Error on the UI - Connection errored out
On the CMD -
model_type EPS
UNet ADM Dimension 0
Running on local URL: 127.0.0.1:7860
To create a public link, set `share=True` in `launch()`.
Startup time: 34.4s (prepare environment: 9.8s, import torch: 9.2s, import gradio: 3.0s, setup paths: 1.6s, initialize shared: 0.2s, other imports: 2.4s, load scripts: 5.7s, create ui: 1.3s, gradio launch: 1.2s).
Using xformers attention in VAE
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
Using xformers attention in VAE
extra {'cond_stage_model.clip_l.logit_scale', 'cond_stage_model.clip_l.text_projection'}
left over keys: dict_keys(['alphas_cumprod', 'alphas_cumprod_prev', 'betas', 'log_one_minus_alphas_cumprod', 'model_ema.decay', 'model_ema.num_updates', 'posterior_log_variance_clipped', 'posterior_mean_coef1', 'posterior_mean_coef2', 'posterior_variance', 'sqrt_alphas_cumprod', 'sqrt_one_minus_alphas_cumprod', 'sqrt_recip_alphas_cumprod', 'sqrt_recipm1_alphas_cumprod'])
To load target model SD1ClipModel
Begin to load 1 model
Moving model(s) has taken 0.02 seconds
Model loaded in 32.1s (load weights from disk: 1.0s, forge instantiate config: 0.1s, forge load real models: 28.9s, load VAE: 0.1s, load textual inversion embeddings: 0.4s, calculate empty prompt: 1.5s).
Loading weights [d91d35736d] from E:\SD\stable-diffusion-webui\models\Stable-diffusion\Juggernaut_X_RunDiffusion.safetensors
model_type EPS
UNet ADM Dimension 2816
Using xformers attention in VAE
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
Using xformers attention in VAE
Press any key to continue . . .
Did you try deleting the whole folder and then re-installing again? Because it happened to me once on A1111 and reinstalling fixed the problem. If forge isn't working try A1111 with SDXL. Connection error happens when you close your CMD window. That's a weird error if you aren't closing your cmd window while running the software. I will try finding a solution if reinstalling doesn't fix your problem. Also 8GB RAM is too low for SDXL, my laptop used to freeze at 8GB RAM and now I am using 32 GB RAM and difference is significant
@@theinformativerealm Thanks for answering. I guess it must be the 8GB RAM. I think I'll stick to SD 1.5 for now, till I get a laptop with higher specs.
@@peacetoall1858 A cheeky little fix is to delete the enviroment file. re boot and the system will reinstall the file. Fixes pretty much any problem with webui
also looking at the model name your using there that isnt my lightning model your using the xl model . the lightning model says lightning in the model name lol
I am using lightning only. It downloaded with that name. It was under your lightning tab a while ago and thank you for an awesome model. I downloaded from the link i have given. Is that not your model?
Btw can i make a video on your SD3 likeness model. Just saw it while i visited the link to see if i made a mistake. But I didn't. The model downloaded with that name only. Seems like now the issue is fixed :)
@@theinformativerealm hi. yes just spoke to a few others and there model name is different to . maybe thats a civitai thing. but yes the sd3 likeness model is mine . its pretty epic in the text and listening to prompts exactly. youll notice in the description how precise it is with sd3 prompts to
@@theinformativerealm yes your welcome to make a video of the model. if you need any advice comment on the civitai model page im there a lot
I will soon. My exams are going on at the moment, then i will try to be consistent on the channel and even make better videos hopefully