Feature Idea
Currently, Anima's qwen3-06B tag handling method makes it difficult to effectively mix artist tags. Perhaps a clip loader should be provided as an optional enhancement to the Anima text encoder to compensate for the shortcomings of the qwen3-06B text encoder when handling multiple artist tags. Of course, LLM is an excellent text encoder, but as an evolution of the SDXL model, and with Prompt still maintaining a tag-based approach, the loading and compatibility of the clip model should be considered. In short, this is beneficial for improving the compliance of prompt words.
References：https://github.com/Anzhc/Anima-Mod-Guidance-ComfyUI-Node
Existing Solutions
No response
Other
No response
Feature Idea
Currently, Anima's qwen3-06B tag handling method makes it difficult to effectively mix artist tags. Perhaps a clip loader should be provided as an optional enhancement to the Anima text encoder to compensate for the shortcomings of the qwen3-06B text encoder when handling multiple artist tags. Of course, LLM is an excellent text encoder, but as an evolution of the SDXL model, and with Prompt still maintaining a tag-based approach, the loading and compatibility of the clip model should be considered. In short, this is beneficial for improving the compliance of prompt words.
References：https://github.com/Anzhc/Anima-Mod-Guidance-ComfyUI-Node
Existing Solutions
No response
Other
No response