------------------------------------------------------------------

Version 5.2: Added dynamic poses on top of Version 4. Due to the complexity of the actions, there may be some issues, just draw a couple of more cards, hope everyone likes it.

------------------------------------------------------------------

------------------------------------------------------------------

Version 4: Re-optimized the training set, the mecha's texture is improved compared to v3.1 and wings are added.

-------------------------------------------------------------------

Version 3.1: This time, 60 out of 120 images were selected for retraining. Perhaps I found a more suitable large model for generating images, and the diversity in the generated images seems better than v2.

-------------------------------------------------------------------

Version 2: Reduced some glare relative to the first version, slightly increased the metal texture. I like both versions;

-------------------------------------------------------------------

- Model Introduction: Model of version SD1.5, Gundam model, feels more playable than the SDXL version;

- Generation suggestions: Lora weights above 0.7 give a transparent acrylic effect, the lower the weight, the less transparent it is; 2D, 2.5D, 3D, realistic large models are all supported.