Download
1 variant available
54
42
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 far
0.65-1.25 of 2.0, appears fairly consistent, though weighting seems more sensitive here. (1.15 or there about's may be best?), up in the higher weight mixing lora had more troubles.
Hair length doesn't always follow prompt, hair color seems okay, a few miss-steps
need to explore facial gesture/expression some, I've added some to prompts and they have somewhat held; they weren't set in the trigger words, but captions may have caught the expressiveness.
Mixed bag with playing niceĀ with other lora, results are passable or mostly passable if with oneĀ and neither is set too high,Ā things start to fall apart with two or more,Ā though it is probably problematic images in the other sets.
Problems seen with multi-mixing: (fur-ish blobs, unusual textures, but mainly resolution reductions and disfiguring of face and hands/feet, arms and torsos less so.
Interesting to see if adding with other racial or nationality lora work and the weightings.
As for steps I'm thinkingĀ 30+, it probably is fine down to 25, over 42/44 things over-processed. I haven't tested using the same seeds to verify.