Sign In

DATA MÖ%$"sH1N)G

19

83

128

7

Verified:

SafeTensor

Type

LoRA

Stats

83

128

337

Reviews

Published

Oct 22, 2024

Base Model

Flux.1 D

Training

Steps: 3,000

Trigger Words

MOSH

Hash

AutoV2
4C1207236E
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.

A super fun effect that was a thing about 14 years ago in music videos.

Trained on my own footage and some random music videos

Why such a high strength though? 1.5-2.0? You must be a such a noob! Tripadvisor 1/5 rating!

TLDR:

It turns out that footage shot on modern cameras use modern codecs that can't replicate this effect that easily. (Stuff that Flux was likely trained on)

Took me whooping 15 minutes to get close to it in Nuke, but in the end went with an Xvid version from 2010 and Avidemux from 2011 deleting random I and P frames and finally it worked as expected <3

Training data is compressed with 8x8, 16x16 and 32x32 blocks in Xvid.

Play with the strength for sure. Downloading this lora you're likely not that type of person who renders their pixel perfect waifu at 16K. (I hope so)