<s>This model was trained 13 times. I need to consider multi-person generation, Joey Jones's "heroman gloves", and model style and make a balance. Of course, I can also go to colab to train, enable individual block training or add regularized datasets. However, the training set has more pictures, and the free time of colab may not be enough.</s>The sample image was generated using incursiosMemeDiffusion_v16PDXL, with a lora weight of 1joey jones<pre><code>joey jones,black jacket,hood,hood down,yellow-white striped shirt,belt,torn pants,</code></pre>simon kaina<pre><code>simon kaina,vest,orange vest,black shirt,grey pants,knee pads,wristband,</code></pre><pre><code>simon kaina,vest,orange vest,long hair,black shirt,grey pants,knee pads,helmet,spade \(shape\),fanny pack,</code></pre>Although Simon Kaina uses a crutch in the anime, the model may not be able to generate a crutch.<pre><code>crutch，holding crutch</code></pre><pre><code>crutch on back</code></pre>Style<pre><code>anime screenshot</code></pre>

Files with SHA256: e893e435de06bc61761180cfc8a65ed3c63675ab149e4159875b474765e9b0a6