Users on Reddit are discussing difficulties in training new genres using Yue2, a LoRA+ model for StableDiffusion. Despite trying various datasets and training settings with tools like Fill's trainer, Aitoolkit, and Yue2 Studio, users report issues ranging from garbled noise to outputs that do not effectively incorporate the trained LoRA. In contrast, training with Acestep appears to be successful, although it has its own default audio quality issues. AI
IMPACT Users are encountering challenges in effectively training custom models with Yue2, indicating potential limitations or complexities in its current implementation for genre-specific LoRA training.
RANK_REASON User-generated content discussing difficulties with a specific tool/model.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
