<h2 id="heading-93">If you like this model and are interested in generating more visit me <a target="_blank" rel="ugc" href="http://ianolmstead-com.vercel.app/">on my website</a>!</h2><h2 id="heading-37">Bio</h2><p>Avril Lavigne is a Canadian singer-songwriter. She is considered a key musician in the development of pop-punk music, as she paved the way for female-driven, punk-influenced pop music in the early 2000s. Her accolades include eight Grammy Award nominations, among others.</p><h2 id="heading-38">Pre Training</h2><p>I gathered 50 images of Avril Lavigne. I trained on wallpapers mostly. I used Blip captioning to generate the filewords and edited each individually to reduce potential hallucinations. I added a 1 to the end of Avril on the embedding name to ensure there is no mixing with previously trained images of her.</p><h2 id="heading-39">Training</h2><p>I used 0.005:100:0,0.0025:250,0.001:500,0.0005:1000,0.00025 for my learning rate. I am going for 2.5K training steps total. I am using a batch size of 1 with Gradient Accumulation Steps set to 3. I am running on a RTX 4090 on the cloud. I am using 12.5 out of 24 GB. The estimated time of completion is .5 hours. For the embedding I am using 5 vectors per token. I switched to SD 1.5 EMA Only model for training.</p>

Files with SHA256: ae4d9218a4d97d8c4b7402303fa540694a23595ebd4c52063a7051e56a9997d9