v1.0 uses the same training method and steps, but the factor/dim settings are incorrect, resulting in very small values.
v2.0 uses the normal factor/dim settings.
Trigger word: artist@yumenouchichiharu,
The two graphics styles are quite different, each with its own unique characteristics.
Even without a trigger word, a lighter style can be achieved.
Weight: 0.5-0.8
A single Lora effect is good.
However, when using multiple Lora, a high combined weight is not recommended; <2.0 is recommended.
Description
v1.0 uses the same training method and steps, but the factor/dim settings are incorrect, resulting in very small values.
v2.0 uses the normal factor/dim settings.
Trigger word: artist@yumenouchichiharu,
Even without a trigger word, a lighter style can be achieved.
Weight: 0.5-0.8
A single Lora effect is good.
However, when using multiple Lora, a high combined weight is not recommended; <2.0 is recommended.