<h2>How to Use This Model</h2>
<p><strong>USE THEM SIMULTANEOUSLY</strong>. In this case, you need to download both <code>ishtar_fgo.pt</code> and 
<code>ishtar_fgo.safetensors</code>, then <strong>use <code>ishtar_fgo.pt</code> as texture inversion embedding, and use
<code>ishtar_fgo.safetensors</code> as LoRA at the same time</strong>.</p>
<p><strong>それらを同時に使用してください</strong>。この場合、<code>ishtar_fgo.pt</code>と<code>ishtar_fgo.safetensors</code>の両方をダウンロード
する必要があります。<code>ishtar_fgo.pt</code>をテクスチャ反転埋め込みとして使用し、同時に<code>ishtar_fgo.safetensors</code>をLoRAとして使用してください。</p>
<p><strong>同时使用它们</strong>。在这种情况下，您需要下载<code>ishtar_fgo.pt</code>和<code>ishtar_fgo.safetensors</code>这两个文件，然后将<code>ishtar_fgo.pt</code>用作纹理反转嵌入，
同时使用<code>ishtar_fgo.safetensors</code>作为LoRA。</p>
<p>(Translated with ChatGPT)</p>
<p>The trigger word is <code>ishtar_fgo</code>, and the recommended tags are <code>masterpiece, best quality, highres, solo, {ishtar_fgo:1.10}, long_hair, black_hair, two_side_up, red_eyes, jewelry, earrings, breasts, hoop_earrings, ribbon, bangs, smile, single_thighhigh, medium_breasts, hair_ribbon, armlet, parted_bangs, navel</code>.</p>
<h2>How This Model Is Trained</h2>
<p>This model is trained with <a rel="ugc" href="https://github.com/7eu7d7/HCP-Diffusion">HCP-Diffusion</a>. 
And the auto-training framework is maintained by <a rel="ugc" href="https://huggingface.co/deepghs">DeepGHS Team</a>.</p>
<h2>Why Some Preview Images Not Look Like Ishtar Fgo</h2>
<p><strong>All the prompt texts</strong> used on the preview images (which can be viewed by clicking on the images) 
<strong>are automatically generated using clustering algorithms</strong> based on feature information extracted from the 
training dataset. The seed used during image generation is also randomly generated, and the images have 
not undergone any selection or modification. As a result, there is a possibility of the mentioned 
issues occurring.</p>
<p>In practice, based on our internal testing, most models that experience such issues perform better in 
actual usage than what is seen in the preview images. <strong>The only thing you may need to do is fine-tune 
the tags you use</strong>.</p>
<h2>I Felt This Model May Be Overfitting or Underfitting, What Shall I Do</h2>
<p>Our model has been published on <a rel="ugc" href="https://huggingface.co/CyberHarem/ishtar_fgo">huggingface repository - CyberHarem/ishtar_fgo</a>, where
models of all the steps are saved. Also, we published the training dataset on 
<a rel="ugc" href="https://huggingface.co/datasets/CyberHarem/ishtar_fgo">huggingface dataset - CyberHarem/ishtar_fgo</a>, which may be help to you.</p>
<h2>Why Not Just Using The Better-Selected Images</h2>
<p>Our model's entire process, from data crawling, training, to generating preview images and publishing, 
is <strong>100% automated without any human intervention</strong>. It's an interesting experiment conducted by our team, 
and for this purpose, we have developed a complete set of software infrastructure, including data filtering, 
automatic training, and automated publishing. Therefore, if possible, we would appreciate more feedback or 
suggestions as they are highly valuable to us.</p>

Files with SHA256: 3807cf37c5a8881c011770d45bc8c8a62a1e05355fce41de537dd3b2ffc76251