All Hi! I consider, offers for the execution of private orders for the creation of LORA, please contact in PM (conditions, prices are discussed depending on the complexity of the order).

I present to your attention, LORA "Third biker 360 degrees" Third because two could not get there :)

The training input dataset used 360 degree perspective images,

LORA was familiarized with the directions: FRONT_VIEW, SIDE_VIEW, BACK_VIEW, on some models the hint works: AROUND_VIEW,

The direction hint works, not with all models, if it is stubborn, try changing the LORA force, it happens to help.

See the hint examples in the images I posted.

If you have many generations with black background, try using the hint: complex background. Doesn't help? reduce the strength of LORA.

trained on the basic model 1.5, works over the whole gain range 0.2...1, start with 0.5 but don't be afraid to do lower gain.

Don't be afraid to experiment and share your results, it will help in fine tuning your next projects.

Be Fun!

finished works based on LORA's submissions can be viewed here:

instagram: https://www.instagram.com/juanmari_art/

facebook: https://www.facebook.com/profile.php?id=61552126447484

p.s

Here is the Medium version of LORA with about 6000 iterations of training, All DEMO images are generated from it and not processed.

In the PATREON account, there are also LITE- 2000 iterations and HARD 12000 iterations.

Dear Friend! I invite you to subscribe to my PATREON account https://www.patreon.com/Juan_360degrees, you will not only support the creation of new LORA, but also get access to exclusive works.