<h2 id="lora-based-on-agm-:-https:x.comagm86997980">LoRA based on AGM : <a target="_blank" rel="ugc" href="https://x.com/hachifukuDZ">https://x.com/AGM86997980</a></h2><h2 id="trained-using-wai-il-v15">Trained using WAI IL V15</h2><h2 id="trigger-words:-agm-stylerx"><span style="color:rgb(250, 82, 82)">Trigger words:</span> agm stylerx</h2><h2 id="recommended-weight:-0.7-0.9"><strong>Recommended weight: 0.7-0.9</strong></h2><p>Note:</p><p>This is trained with mixed sfw and nsfw dataset, excels in nsfw AGM style</p><p>This style work best with plump body(Style bias)</p><p>Lower weight tend to have better background, you might lose the style the lower you go.</p>

[Style | illustrious-XL] AGM

<p>Miles better background performance even in weight 1</p><p>Out of frame in 2:3 should've been fixed</p><p>*patched with less aggressive training revision</p><p>This version have shown way less lora compatibly, if your character lora have a strong style by their own, it may pollute this lora causing it to lose the style</p><p>Therefore, i heavily recommend going higher weight for this version (0.8-0.9)</p>

Files with SHA256: b6fbd9bd5b81c0af4945850ca2540ee3f2c0933064226978990f22b52eb4dbdc