<h3 id="foreword-6z6zfjvx8"><strong><span style="color:rgb(76, 110, 245)">Foreword</span></strong></h3><p>This is the last <em>LoRA</em> model of several attempts to change the behavior of a bottle in an image. And, I am satisfied with this <em>LoRA</em> model so far.</p><h3 id="goal-1gedje7uf"><strong><span style="color:rgb(76, 110, 245)">Goal</span></strong></h3><p>The goal should be to get images with a laying bottle on a floor, on the ground, on a table, in the sand at the beach or wherever one wants, if one uses this <em>LoRA</em> model. The bottle should be at first glance transparent with a cap made of cork or other material.</p><h3 id="what-the-model-does-67n3pzyv6"><strong><span style="color:rgb(76, 110, 245)">What the Model Does</span></strong></h3><p>The LoRA model creates a laying transparent glass bottle with a cap. </p><p>The color of the bottle as well as the cap can be modified.</p><h3 id="good-prompt-9ix34trnd"><strong><span style="color:#4c6ef5">Good Prompt</span></strong></h3><p>A good working Prompt looks like</p><pre><code>laying on its side transparent red bottle on stony ground, background lava spewing volcano, &lt;lora:EmptyGlassBottleConcept:0.75&gt;</code></pre><p>First one describes the bottle a little bit and second one describes the background more or less in detail. Then one has a good base concept.</p><h3 id="weights-3lsqtlwif"><strong><span style="color:rgb(76, 110, 245)">Weights</span></strong></h3><p>With a weight of 1.0 one will get images which looks like the ones I prepared for the training.</p><p>Good results got I with weights between 0.5 and 0.7.</p><h3 id="training-lhsh830j9"><strong><span style="color:rgb(76, 110, 245)">Training</span></strong></h3><p>After a few failures in creating a good LoRA model, I made suitable images and hoped that these would work better than the previous created images. </p><p>In the end I used a reduced set of images. In the end there were only 4 images remaining. Using buckets and mirroring I used for the training 32 images in total. </p><p>Training iteration were 3500 in total at the end.</p><p>I used as image file names</p><pre><code>lying down bottle [1...4].jpg</code></pre><p>In the case of this LoRA model I did not use a caption text file for the training.</p><h3 id="trigger-word-cb6psb1tj"><strong><span style="color:#4c6ef5">Trigger Word</span></strong></h3><p>The Trigger Word is</p><pre><code>glassbottle</code></pre><p>The LoRa model will work well without the use of the <em>Trigger Word</em>.</p><h3 id="further-reading-swfr8mfd7"><strong><span style="color:rgb(76, 110, 245)">Further Reading</span></strong></h3><p>In parallel to the LoRA model development I wrote an article about the reason and motivation for this model. This article can be found here <a target="_blank" rel="ugc" href="https://civarchive.com/articles/7841/pitfalls-of-a-prompt">https://civarchive.com/articles/7841/pitfalls-of-a-prompt</a></p><h3 id="final-words-oh7qxrpbc"><strong><span style="color:rgb(76, 110, 245)">Final Words</span></strong></h3><p>Have a nice day. Have fun with model. Be inspired!</p>

Files with SHA256: 97ffe0e1dd330c4dd90c02305e21c19196d7974d269a6e67e324a9716e92f77a