Download
1 variant available
6640 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
750 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
3130 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
(81)
Oct 13, 2024
Complete retrain with
full (JOY2) captions (v1 had only few tag words)
attention-mask (no backgrounds)
1024x1024 (v1 was 768x768)
640+ images (v1 was 450+)
Results:
follows prompts much better
should work better together with other Loras
is better on the closeup details (skin pores, lips, etc.)
Images look really different from V1.x
In general if the output looks 'too much'/overtrained you can simply lower the strength of the lora (down to 0.6)
Please see example images for prompts
Show more

4530 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
3.3K0 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
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.
Trained 1024x1024 on 600+ high resolution sharp images of women opening mouth wide, sticking out tongue, showing uvula, displaying teeth, roof of mouth etc.
Most realistic samplers are like DPM++ 2M or DEIS or SD Forge - Flux Realistic (slow) (Scheduler: Beta ), the latter gives amazing results!
Guidance / (distilled) CFG between 1,8 and 2,5, Keep it low for realism!
Lora strength between 0.7 and 1
V2.x has a different, extended, training set and configuration, following prompts better yet with different image results from V1.x, more variety.
V1.2 is the final from V1 training set and config, producing excellent image quality, but not following prompts as well and less variety in output

