Please post the first picture directly here in the Model Gallery. This is the best way to support creators :)
Hello ♥
for whatever reason you want to show me appreciation, you can: ❤️Ko-Fi❤️
_________________
This was my first images back in 2022 with Midjourney with the prompt:
green spongebob squarepants with tons of oreo
Imagine, this was my first ever generated image with AI and this was the first “prompt” I had in my head 🤣
Now, with this Lora and ForRealXL this is the result
This Lora was a test with a new captioning method that delivers much better results than expected.
So you can try it out too.
This Lora was trained with 34 Midjourney V6.1 images. (Version 0.1)
Settings
0.3 - 1.0
Trigger (v0.2+)
aidmaMJ6.1Trigger (v0.1)
midjourneyv6.1Hope you get a lot of beautiful results ♥
Description
v0.5
FAQ
Comments (24)
Looking forward to when this gets released!
Doesn't work well together with Pony (Cyberrealistic). Creates boring backgrounds, ignores my prompt very often. Had 70% of failure results. When I get a good result, it looks awesome.
Thank you very much for this Lora ☺ the quality of the new IL version is exceptional ~ gives the pictures a touch of elegance
thank you so much, so glad you like it ♥
Is the dataset still 34 Midjourney images? Because I see the description is virtually unchanged despite new versions being released.
oh no, the v0.5 has 160 images. The 34 images in the description was just for v0.1
@AIDigitalMediaAgency heh, you should definitely update the description then ^^
Which trainer did you use? I assume either One Trainer or XLabs? (Assuming this is not a Kohya or Diffusers trained LoRa.)
I train all my Flux and IL loras with the Civitai trainer.
This work stands out for its detail and unique style. The use of the LeRA model adds a special charm and realism to the image. I particularly liked how the textures and colors were conveyed, making the image vibrant and atmospheric. Great job that showcases skill and creativity!
Thank you so much. Glad you like it ♥️
Could you please not use a dot in the triggers? This is a very bad decision.
I'm not the one who created this model, but I'm guessing that the reason you're thinking it's a bad decision is because the dot is used at the end of the sentences.
I can fix it in the next version :)
In the mean time, I used "Adjust LoRa Filename" extension in Automatic1111 to change it (first rename the lora file to whatever you want the trigger word to be, then run the extension on it. I simply just removed the dots, to "aidmaMJ61SDXL-v05.safetensors" etc. Ofc not the file extension dot).
https://github.com/zentrocdot/sd-webui-adjust_lora_filename
You'll get some error warning in termonal about invalid header deserialization, but it'll still work (afaik). I'm not sure what that is, probably some kind of checksum doesn't add up or something. I've done this on a few to remove double underscores on some Lora names or triggers, since another extension (dynamic prompts) use that for its syntax, and it'll get confused otherwise.
But I guess most of the world have moved away from Auto1111, this probably doesn't help much... just FYI just in case.
Ugh, half disregard my previous comment, I meant it changes the "output filename", not the trigger(s).
which flux model works best with this lora?
Classic Flux dev.
Whats the difference in this model to the FLUX Image Upgrader from same of yours? I noticed that the Image Upgrader gives more details and this model make smooth(or I didn't notice a big difference). Love both models, good work!
It's not well explained what this Lora is supposed to do, can you explain it to me properly and how to best use it?
Or even just fucking-- hey: this is what you're looking at, this what it does.
Hi! Can you tell me how it works? I downloaded lora flux.2 v 0.5 But the model doesn't load with the error lora key not loaded: base_model.model.single_blocks.43.linear1.lora_B.weight and so on. What am I doing wrong? in the standard flux 2 dev workflow, comfyui
same problem, lots of errors "lora key not loaded:"
same
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.