What is the best training method to train model certain object that is the same shape and similar fo
What is the best training method to train model certain object that is the same shape and similar form but it differs in decorations and filling. I have tried Dreambooth and Textual Inversion. With Dreambooth I used LORA for training and main problem is that when Im generating new things I can't specify how the object should look like because it is generated randomly and my prompts in most cases doesn't work, or work poorly. With textual inversion i feel like generated objects are poorer quality than loras. I have more control with the prompts but objects are generated in different perspecitve and in most cases not as a whole but zommed in.
