Furkan Gözükara
patreon
1 Click Download 160 Very Best Stable Diffusion 1.5 Based (SD 1.5) Models
🕑 Added 2024-01-17 13:38:28 +0000 UTCJoin discord and tell me your discord username to get a special rank : SECourses Discord
Login or Register RunPod : https://runpod.io?ref=1aka98lq
23 January 2024 Update
- Test results written text file model_test_findings.txt added to the attachments
- Full tutorial video published : https://youtu.be/G-oZn4H-aHQ
- If you could upvote this Reddit Thread I would appreciate very much
18 January 2024 Update:
- Comparison links added to the post check them out
- Preparing those comparison files took a full day
I have spent over 1 day to code a special python script that will download the very best 160 models from CivitAI and Hugging Face.
Total size is 409 GB.
You can edit the download_top_160_SD_1_5_models .py file and disable download of any model you want.
Why I prepared this?
Currently I am doing testing on each model. Once I find the most realistic and advanced one, hopefully I will do academic experiment on that model to find out very best DreamBooth training configuration as I did for SDXL.
On Windows download_top_160_SD_1_5_models. bat file and download_top_160_SD_1_5_models. py file and double click bat file to start downloading.
If any error occurs during download, it will move to the next model. After entire download ended, run the script again and it will download only error occurred models. Such errors occurs since this is a network based script.
For runpod just upload download_top_160_SD_1_5_models .py file into workspace folder and run as below
- python download_top_160_SD_1_5_models.py
When you run the script, it will ask you how many threads to download. 1 works best but based on your network speed, you can enter more such as 5.
You can also edit py file and set the download folder as you wish.
The files will be downloaded into the stable-diffusion-webui/models/Stable-diffusion folder by default wherever the script is run. You can edit it.
Automatic Installers for Automatic1111 SD Web UI
- Windows : https://www.patreon.com/posts/86307255
- RunPod : https://www.patreon.com/posts/86438018
Comparison x/y/z Plot Results PNG Images
Click To Download More Comparisons (10 X/Y/Z plots - 1.1 GB)
- Click links and save as to download and open.
- 40 steps and DPM++ 2M SDE Karras used to generate images.
- No negative prompts or anything else.
- 768x768 grids are : 1205.82 Megapixels (1.2 gigapixel)
- 768x1024 grid is : 1491.64 Megapixels (1.5 gigapixel)
- 1024x1024 grid is : 2131.01 Megapixels (2.1 gigapixel)
- Image Of A Man - 768x768 - full resolution - 1.1 GB - right click save as
- Image Of A Man - 768x768 - 50% resolution - 300 MB - right click save as
- Image Of A Man - 768x768 - 25% resolution - 77 MB - right click save as
- Photo Of A Man - 768x768 - full resolution - 1.1 GB - right click save as
- Photo Of A Man - 768x768 - 50% resolution - 300 MB - right click save as
- Photo Of A Man - 768x768 - 25% resolution - 77 MB - right click save as
- Picture Of A Man - 768x768 - full resolution - 1.2 GB - right click save as
- Picture Of A Man - 768x768 - 50% resolution - 300 MB - right click save as
- Picture Of A Man - 768x768 - 25% resolution - 77 MB - right click save as
- Photo Of A Man - 768x1024 - full resolution - 1.5 GB - right click save as
- Photo Of A Man - 768x1024 - 50% resolution - 378 MB - right click save as
- Photo Of A Man - 768x1024 - 25% resolution - 100 MB - right click save as
- Photo Of A Man - 768x1024 - full resolution - 2 GB - right click save as
- Photo Of A Man - 768x1024 - 50% resolution - 500 MB - right click save as
- Photo Of A Man - 768x1024 - 25% resolution - 128 MB - right click save as
Comments
Furkan Gözükara
yes it downloads all models as .safetensors model
Gpr0mpt
Is it possible to download the files in the safetensor model? Because I don't have Stable Diffusion installed on the machine, I use it on websites...
Furkan Gözükara
yes i am planning hopefully
OctonionPrime
Thank you very much for this! Do you plan to do the same for SDXL models? :)
Furkan Gözükara
Oh I see. yes that can be the reason. My auto installers are expecting CUDA and NVIDIA drivers to be installed.
Robert White
I think I might have found the issue. I'm using an EC2 instance on AWS and by default it didn't have the nvidia drivers installed on it. I'll try installing them and then see how that goes.
Furkan Gözükara
hello. what is your GPU? if AMD you need to add --skip-torch-cuda-test and some other parameters.
Robert White
This is for the auto1111 install
Robert White
I've got your automatic downloader but I'm running into this error when running it. "RuntimeError: Torch is not able to use GPU; add --skip-torch-cuda-test to COMMANDLINE_ARGS variable to disable this check"
Furkan Gözükara
thanks a lot for sharing. i checked a lot of models but must have missed
Nenad Kuzmanovic
Oo nice, but you have missed one really interesting model :-) https://civitai.com/models/148165/puredarkness Just kidding, but you have to check it out - i'v Finetuned it with almost 1000 "cheery-awesome-picked" images that i created in Illuminati Diffusion... I think it is kinda unique model, you can do a lot of awesome imgs with it. I find out it is really great for merging with other models - so you can try to merge it with your Dreambooth model trained on your face etc...
