Can we use the . Pt model to generate images and the corresponding dlatents directly？

[wytcsuch]

Hello, can I use the. Pt model to generate dlatent and then generate the images as the official implementation:

    src_latents = np.stack(np.random.RandomState(seed).randn(Gs.input_shape[1]) for seed in src_seeds)
    dst_latents = np.stack(np.random.RandomState(seed).randn(Gs.input_shape[1]) for seed in dst_seeds)
    src_dlatents = Gs.components.mapping.run(src_latents, None) # [seed, layer, component]
    dst_dlatents = Gs.components.mapping.run(dst_latents, None) # [seed, layer, component]
    src_images = Gs.components.synthesis.run(src_dlatents, randomize_noise=False, **synthesis_kwargs)
    dst_images = Gs.components.synthesis.run(dst_dlatents, randomize_noise=False, **synthesis_kwargs)

[lucidrains]

@wytcsuch sure, I can probably offer a similar interface

what is dst_latents vs src_latents?

[wytcsuch]

@wytcsuch sure, I can probably offer a similar interface

what is dst_latents vs src_latents?

Maybe my problem is not clearly described. I hope that I can give Z randomly, and then convert Z to dlatent and finally generate the corresponding image： z- > dlalent - >image

    z = np.stack(np.random.RandomState(seed).randn(Gs.input_shape[1]) for seed in src_seeds) 
    dlatents = Gs.components.mapping.run(z,None) # [seed, layer, component]
    images = Gs.components.synthesis.run(dlatents, randomize_noise=False, **synthesis_kwargs)

You don't need to understand dst_latents and src_dlatents Specifically, they are both dlatents. I just need to record the dlatent and the corresponding image

[lucidrains]

@wytcsuch ok, it's much clearer to me now

yup, i'll provide you an API similar to the official repo soon

[lucidrains]

@wytcsuch https://github.com/lucidrains/stylegan2-pytorch#coding let me know if that works for you

[wytcsuch]

@lucidrains thanks a lot，I'll try it. I think it should work

[LeMagicienHuang]

@lucidrains thanks a lot，I'll try it. I think it should work

Hello bro, does this encoder work?