Sign In

Terragen

12

65

0

3

Verified:

SafeTensor

Type

LoRA

Stats

65

0

Reviews

Published

Sep 26, 2024

Base Model

Flux.1 D

Training

Steps: 3,000

Trigger Words

TRGN
3D

Hash

AutoV2
1287090C6D
default creator card background decoration
kklors's Avatar

kklors

The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.

IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

So this lora is aiming for an old-school 3D render look. I started with Terragen 0.9 from 2005, but it was nearly unusable and only 32bit. Ended up with Terragen 2.0, which is still old enough. For that reason I only rendered training data in 2K, because even trying to import high res modern shaders it either didn't work at all or was not much better due to lack of tools.

Because Terragen is primarily made for landscapes this lora is mostly made for those too. I was too lazy to download whacky human models and put them in there. Might be an idea for version two. So having any kind of nature in your prompt will always trigger it.

For topics other than landscapes play with lowering the strength.

However there is a major difference on how to render it.

1 pass: Will give you that old 3D render look out of the box.

Multiple passes with de-noising: Will still look like a 3D render but heavily enhance it into a more modern look with better lighting.

I will include examples of both 1 and multiple passes in the gallery, you will notice which one is which.

Dirty self promotion. Combining it with https://civitai.com/models/703894/mandelbulb3d might give interesting results.