Reverse cowgirl pose. Trained on images with the faces blanked out, so there is no face data present to corrupt your existing LORAsThis means it can be mixed with other LORAs, but you need to adjust the weights between them to get better results (and the CLIP weights if you can)The blank faces don't appear in created images as they were mentioned in the captioning.If using with no other LORA, use at 0.75-1 strength (LORA clip strength between 0.6-1.0)Results will not be perfect due to LORAs competing, but...With a second "person" LORA:Person: 1.0, (low clip strength 0.2-0.5 preferably)This LORA: 0.65-1.0 (higher clip strength 0.5-0.8)Clip strength can be adjusted in ComfyUI. Not sure about other services...

Files with SHA256: ecd2200fb78cae4823568e8f03fa543b06ec2cd02fee58dee16190a76f3f8c84