<h3 id="introduction:">Introduction:</h3>Ttis is the character Sloth from <span style="background-color:rgb(26, 27, 30);color:rgb(141, 141, 143);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">from Deadly 7 Inside Me by "Deruu Rio Ta"<hr /><h3 id="note:">Note:</h3>In terms of accuracy, the overall dress is 80% similar and both the "sleep mask" and "blue pocket" will sometimes break. As for the "doll" bear, it's basically trained in the datasets, but the model never learns, so you won't get the correct image of it when you create it. In this part, I have not planned to train more dolls to make lora.About the second version train plan as mentioned before in the section "<a target="_blank" rel="ugc" href="https://civarchive.com/models/1874283/nyanko-onlinepetclub">https://civarchive.com/models/1874283/nyanko-onlinepetclub</a>" in the section dated "Aug 9, 2026"Second, I had to spend many days to complete the dataset for the characters. before there was a model because the comic was too big to finish and then had to be arranged.<hr /><h3 id="images-test-from-frontsidebehind">Images Test From Front/Side/Behind</h3>combie_front_side_behind.png, 1941x1368, Taken from ComfyUI Outputs, Edited with Krita. Note : Each direction requires adjusting the prompt to get the correct image.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/cfb7aac4-bf61-43c7-b13f-2a1129a640fa/original=true/cfb7aac4-bf61-43c7-b13f-2a1129a640fa.jpeg" /><hr /><h3 id="prompt-trigger:">Prompt Trigger:</h3>Base Character:<pre><code>S1o1h, 1girl, twintails, purple eyes, purple hair, black fluffy hair ornaments, ahoge, medium hair, hair scrunchie,
sleep mask,</code></pre>Outfit Pajamas:<pre><code>dress, black dress, short sleeves, ruffled hem, pajamas, blue gingham, blue pocket, pink trim, frills dress,

slippers, black footwear, black slippers, </code></pre>

Sloth - Deadly 7 Inside Me

### Training Information (Anima DiT)+ Output Name: my_anima_lora+ Network Module: networks.lora_anima+ Epochs: 11+ Max Train Steps: 616+ Learning Rate: 0.0001+ Text Encoder LR: 5e-05+ Optimizer: AdamW8bit+ Mixed Precision: bf16+ LR Scheduler: constant+ LR Warmup Steps: 0+ Rank: 16/16 (Dim, Alpha)### Anima Arguments:+ timestep_sample_method: logit_normal+ discrete_flow_shift: 3+ weighting_scheme: logit_normal### Dataset Info (dataset.toml):+ Batch Size: 2+ Resolution: 1152x1152+ Steps (calculated): 0+ Bucketing: True### Datasets Parameter:+ V1 : images: 29 | repeats: 3 | keep_tokens: 1 | flip_aug: False | shuffle_caption: True

Files with SHA256: 7e45512d61d0770d786c69b891a2e9689c33622f95f4f711f419cd223e4d955f