SDXL and shadows of the past
The outdated training mechanism holds the model back. Additionally, Qwen introduced tokenizer issues that were not found in the T5 model.
SD* models used to generate background from all text tokens, not just those which are related to the background.
The same is true for Anima, the attention modules are not weighted during training, which leads to concept bleeding and background issues.
The same text flows through at each inference step. It's still the unfiltered garbage in, garbage out, but with higher-resolution garbage from the training data.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support