<p><span style="color:rgb(34, 139, 230)">Please contact me here on civitAi or on Twitter if you want to provide the LoRa in a generation service.</span></p><p></p><h3 id="heading-19">TLDR (Because I don't care)</h3><p>Use <code>piper bird &lt;lora:piper_v1:1&gt;</code> in positiv promt.</p><p><em>If you like the LoRa, please feel free to share your generated images here and leave a review (</em><strong><em>or at least a rating</em></strong><em>) ;D</em></p><p></p><h2 id="heading-20">About</h2><p>I wanted to challenge myself with something new and I was wondering how well animals can be trained as LoRa, after watching the short Pixar movie "Piper" (just 5 minutes of wohlsomness ^^) I decided to use this little bird as a data source...</p><p></p><h2 id="heading-21">Training</h2><p>Even though it's a really short film, I was able to get 47 shots to train on, simply because there's a lot of movement going on, which is the best way to get training data :D</p><h2 id="heading-22">Usage</h2><p>No weight adjustments required:</p><p><code>piper bird, animal &lt;lora:tinkerBell_v1:1&gt;</code></p><p>I recommend the addition of <code>animal</code> to the positive prompt... most of the checkpoints I use were trying to get a human person into the image without it (well understandable ^^).</p><p>I also recommend not using latent upscaling, unless you want a bird with human ears... (uuurg so creepy, but that depends on the checkpoint again...) simply using another upscaler sloved the problem for me ^^</p>

Files with SHA256: f9ec930a8d4d01d6d1f8bc97961ecf2fe529f1a2ead5cd691019041878a67506