/vt/ - Virtual Youtubers

>>38601455
>>38601378
Ill do a prompt matrix of my quality tags hang on

Using a low token amounts for me is to personally know what the embedding knows. Every embedding is different even if they are trained exactly the same. There is no way to know beforehand if one is good. A real use case scenario like picrel tells me that there is almost no real difference in the quality of the embedding.