This is my second lora trained on larger network dimension. Combined with maximum resolution, it seems to give more details and more variety in the results. But I do see that it seems to capture more biases so I release it with this notice. Terms like "long/large nipples" aren’t used in the captions, but the mention of erect nipples should work to have large nipples. So far, the results are more random, I will try to change this in following tests.The dataset for this lora has been prepared like my typical lora. It is captioned in natural language without trigger words. It is mostly trained on close-ups with some wider angle shots in there to prevent view biases. All faces in dataset have been censored in various ways to prevent the result from looking like the persons in the dataset. If you want to see a face, put details of it in your prompt. Random edits were used to prevent the introduction of bias caused by these edits.

Naked long and large nipples for FLUX

<pre><code>{
 "engine": "kohya",
 "unetLR": 0.0005,
 "clipSkip": 1,
 "loraType": "lora",
 "keepTokens": 0,
 "networkDim": 4,
 "numRepeats": 4,
 "resolution": 1024,
 "lrScheduler": "cosine_with_restarts",
 "minSnrGamma": 5,
 "noiseOffset": 0.1,
 "targetSteps": 2614,
 "enableBucket": true,
 "networkAlpha": 8,
 "optimizerType": "AdamW8Bit",
 "textEncoderLR": 0,
 "maxTrainEpochs": 40,
 "shuffleCaption": false,
 "trainBatchSize": 3,
 "flipAugmentation": false,
 "lrSchedulerNumCycles": 3
}</code></pre>

Files with SHA256: 5a0b6132f63f895d5797cbe83bc6d215572625d90589d7b8c16551e0b8fc75cd