This is an experiment to see if I can make a TI embedding that gets the flavour of konyconi’s BohoAI LORA.<a target="_blank" rel="ugc" href="https://civarchive.com/models/51966/bohoai">https://civarchive.com/models/51966/bohoai</a>Thank you to @konyconi for sharing his dataset for the excellent BohoAI LORA.<a target="_blank" rel="ugc" href="https://civarchive.com/models/52697/tutorial-konyconi-style-lora">https://civarchive.com/models/52697/tutorial-konyconi-style-lora</a>The showcase uses 2 models:revAnimated_v122.safetensors [4199bcdd14] with clip skip = 2avalonTruvision_v2.safetensors [a4df55d292] with clip skip = 1This TI can produce some decent Boho pix but it gets confused sometimes... eg asking for a spaceship and getting a truck. Perhaps for this sort of TI you need to use a lot more pictures in the training dataset, with more subject variation?---------------------------Update 09 May 2023Continued the training to step 4000, and then 5000.kcboho07-4000 produces a stronger Boho style.kcboho07-5000 is stronger again but has increased duplication/repetition. e.g. more fingers, more hands, duplicate cities floating in the sky.Tried 6000 steps but it’s even worse - overcooked.I’ve uploaded the 4000 step version as, probably, the best result for this experiment.Also uploaded the 5000 step version since it can produce nice results with careful object prompts.---------------------------I’ve been struggling to work out how to make a style TI...what makes a good training dataset?what training settings should I use in automatic1111?how long to cook the TI for?For my training dataset I copied konyconi’s 76 1024x1024 images to a new folder without the associated TXT files, and reduced them all to 512x512. Then I renamed them “01 aeroplane.png”, “02 city.png”, “03 tank.png” etc.Why? Because I was trying to match what I’ve done in the past for TIs that ended up usable. The reduced images dataset folder is what I used in the settings below.The wikipage for automatic1111 Textual Inversion is here:<a target="_blank" rel="ugc" href="https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Textual-Inversion">https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Textual-Inversion</a>but way out of date. Last rev Jan 5, I’m writing this May 8.I found this thread useful in parts. It’s a long read!<a target="_blank" rel="ugc" href="https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/1528">https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/1528</a>Training model: v1-5-pruned.ckpt [e1441589a6]I used this because I don’t know any better, and it’s been useful in the past. Should I be using a different model for training, or is the base SD15 the best thing to use? No idea.Create embedding:name: kcboho07initialization text: boho style photonumber of vectors per token: 4Train embedding:Embedding name: kcboho07Embedding Learning Rate: 0.001:250, 0.0005:500, 0.00075:1000, 0.001Gradient Clipping: disabledBatch size: 1Dataset directory: wherever you’ve put it on your computerLog directory: textual_inversionPrompt template: minimum_style_2.txtThe template has 3 lines:&lt;&lt;&lt;[name] style, [filewords][name] style, a photo of [filewords][name] style, an illustration of [filewords]&gt;&gt;&gt;Width = Height = 512Do not resize images: OFFMax steps: 3000Save image steps: 25Save embedding steps: 25Use PNG alpha channel: OFFSave images with embedding in PNG chunks: ONRead parameters from txt2img tab: OFFShuffle tags: OFFDrop out tags: 0Latent sampling method: deterministicTraining time: about 50mins per 1000 steps on a 2060/6GB.The TI at 3000 steps does produce a Boho style, although I think it’s a bit hit-and-miss compared to the BohoAI LORA.If anyone has suggestions about what I should be doing differently please add a comment. Or if I’m doing anything obviously stupid! :-)

BohoTI

Initial version (6 fails before I got this v07). This is a TI so you can change the trigger word by renaming the PT file.

Files with SHA256: eb6d3ef84bfb0d68e9652688ff234d2e6634f8e471e6bc7fc0cf5766a613995e