Lora for lifted fake breasts, which are slightly elongated and larger.For larger breasts, 1.2 gives a good result in my tests, a lot of the dataset is naked breasts so if you don’t want nipples, try including something like « the dress cover her nipples, there is no nipple visible ».V2: Major overhaul, dataset update and training update. This was my first lora trained on flux, I learned a lot since then. This update removes a lot of biases from the dataset and I added a few images in the dataset. It is trained at max resolution with a bigger network dimension for better details. Results seems more reliable and way more flexible.The dataset for V2 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.

Lifted fake breasts for FLUX

<pre><code>{
 "engine": "kohya",
 "unetLR": 0.0005,
 "clipSkip": 1,
 "loraType": "lora",
 "keepTokens": 0,
 "networkDim": 4,
 "numRepeats": 5,
 "resolution": 1024,
 "lrScheduler": "cosine_with_restarts",
 "minSnrGamma": 5,
 "noiseOffset": 0.1,
 "targetSteps": 2025,
 "enableBucket": true,
 "networkAlpha": 8,
 "optimizerType": "AdamW8Bit",
 "textEncoderLR": 0,
 "maxTrainEpochs": 30,
 "shuffleCaption": false,
 "trainBatchSize": 2,
 "flipAugmentation": false,
 "lrSchedulerNumCycles": 3
}</code></pre>

Files with SHA256: b0d9bc621668b92617d94dc8baf8896cc770c344b3078b11d34bc07475bea18d