For this assignment, I played around with the BigGAN for video and image generating. In detail, I played with the truncation and noise_seed for image generation. From my understanding, the truncation is to limit the latent space and resample from a refined space so that the generated image will be more vivid as the resampled latent space is denser. Besides, the noise_seed adds noise to the image generation process so that we are able to see different images sampled from the latent space.
As we can see from the above screenshots, the higher truncation value we set, the more vivid images we can get. However, from the paper, the author also states that if the model is weak, a higher truncation will introduce some side-effect to the generation process.