<p>This is a Gragas LoRA trained on a diverse set of images, including various angles of his Wild Rift model, official art, as well as both safe and nsfw fanart. It struggles with back shots because most of the training data is of frontal perspective images, but with some negatives like <code>frontal</code>, <code>in front</code>, etc. you should be able to get some back perspective images. Biggest trigger words are <code>gragas</code>, <code>beard</code>, <code>orange hair</code> and <code>fat man</code></p>

Gragas - League of Legends

<p>10 epochs using 4 repeats per image</p>

Files with SHA256: c081942e3370346299b1893378e9c7212710a452da430c5831c8da312456a748