A request of @GeneticPerfectionAI.
Tahlia Hall is an Australian model and social media personality.
200-step TI trained on a dataset of 15 images with these settings. I've modified them because I wanted to reuse the dataset @GeneticPerfectionAI had provided me with back in the day, which consisted of 25 images. Thus, I decided to reduce batch size from 3 to 2 and train until 210 steps, finally setting for step 200 after some testing. Gotta say I'm not totally sure how much of an improvement this version is, as the previous one still holds up perfectly today with current checkpoints, in my opinion, but you can definitely download both. :)
Curious about my work process? I have summarized it here.
Do you have a specific idea for a TI in mind? Visit my website and let me know.
You're obviously free to experiment, but bear in mind that my TIs are trained with a more or less fixed phrasing, that normally starts with:
"photo of EMBEDDING_NAME, a woman"
So I recommend always starting your prompt like that and then building the rest of the prompt from there. For instance, "photo of beautiful (t4hli4h4ll:0.99), a woman, perfect blonde hair, (closeup portrait), (modern photo, party strapless dress), 85mm, (analog, cinematic, film grain:1.3), (at a red carpet event), ((detailed eyes)), (color picture:1.1), (looking at viewer)".
Please also note that I'm using the "add detail" LoRA for my example pics. I recommend setting it around 0.5 for best results. CLIP Skip should always be set at 1 when generating pics with this TI.