Sign In

OB等轴3d房间 Isometric 3d room V2.0

174
1.1k
123
94
Verified:
SafeTensor
Type
LoRA
Stats
762
75
Reviews
Published
Sep 12, 2024
Base Model
Flux.1 D
Training
Steps: 11,000
Epochs: 20
Usage Tips
Clip Skip: 2
Trigger Words
OBdengzhou
C4D_style
Isometric Room Digital Illustr
Hash
AutoV2
8068818B9B
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.

A brand new version of FLUX is coming !!!

It improves the stability of image output, fixes the problem of blurred image quality, removes the close-up perspective, and fixes the angle of image output content to make it more dedicated. At the same time, it improves the expressive power of details.

Come and try it out. Maybe it's also good for commercial use!

To ensure the details of each object in the scene, please enable high-definition restoration. Magnification factor: 2 times. Restoration method: R-ESRGAN 4x + Anime6B.

After a lapse of two months, the 2.0 version of the isometric room is here! (Card drawing is required. The effect is not very stable yet. The repaired version will be launched tomorrow.)

This time, a brand new FLUX large model is used for training, which can well understand the prompt words and correct the spatial distribution of objects. At the same time, it brings better detail performance. The 2.0 version has strengthened the display of close-up scenes and lighting effects, enabling each object to be in an appropriate position in the picture, and enhancing the picture atmosphere.

V1 version isometric 3d room lora description:

This model to indoor 3d isometric perspective, C4D rendering of the model lora. need to open the high resolution fix (Hires. fix) to ensure the quality of the out of the picture, welcome to try to return more pictures ah. Questions can be commented on, see will reply, everyone more anti-picture.

Recommended Parameters

Positive quality word reference:

dengzhou,Octane Renderer,c4d,soft light,solid color background.

Translation: isometric, OC Renderer, C4D, soft light, clean background, these words affect the figure of light and shadow and texture, must be filled in, light can be changed according to the actual situation!

Negative tips word reference:

ntsf, deranged, urved lines, messy, overlapping, distorted, repetitive, proximity, etc.

Big model: sdvn53dcutewave_v10, 3d class all can try down. Model link: https://civitai.com/models/103178/sdvn5-3dcutewave

lora weights: 1.0

Sampling method (sampler): DPM++ 3M SED Karras (you can choose the sampler according to the actual situation)

Iteration Steps: 20

CFG Scale: 7

Resolution reference: 512*512 is recommended (square map)

High resolution fix (Hires. fix): 1. zoom algorithm: Latent, 2. redraw amplitude: 0.7, 3. zoom speed: 2 times