名前の通り、RAdamScheduleFree オプティマイザとNoobAI-XL V-predモデルのテストの為に、Hなゲームcg1340枚×2(左右反転)を試しに学習させたLyCORISです。 過学習気味？の為、weight 0.3-0.55位で使用してください。As the name suggests, this is LyCORIS, which was trained with 2 sets of H-game CG images (1340 images each, flipped left and right) for testing the RAdamScheduleFree optimizer and the NoobAI-XL V-pred model. Please use it with a weight of around 0.3 to 0.55, as it may be slightly overfitted.

Files with SHA256: c674fd95033d0104cfc092c1bc52857581b0c8ba0874b6137b7ea99710a7cbc5