Pernah bikin foto yang realistis pakai Stable Diffusion Pipeline. Kali ini aku pake Google Colab buat running Stable Diffusion Pipeline nya. Dan model yang digunakan yaitu CompVis/stable-diffusion-v1-4.
Ini tampilan Stable Diffusion Pipeline di Google Colabnya:
Ini dia hasilnya:
prompt = "Lifestyle photography photo of a young pretty Japanese woman,
with long, black, straight hair, pouring tea in a traditional tea ceremony,
looking at viewer, upper body framing, in a Japanese tea room, symmetrical background,
under (natural lighting:1.3), shot at eye level, on a Sony A7111,
with a (soft focus filter:1.3),(in the style of Wes Anderson:1.3)"
image = pipe(prompt).images[0]
image
prompt = "Candid photo of a beautiful woman, with captivating eyes,
smiling amidst a bustling city, upper body framing, in a street photography setting,
golden hour lighting:1.3), shot at eye level, on a Fujifilm X-T4 with a 50mm lens,
in the style of Alfred Stieglitz"
image = pipe(prompt).images[0]
image
prompt = "(Glamour photography:1.3) photo of an Asian woman, with long,
sleek, black hair, looking dreamily into the distance, upper body framing from waist,
in a sparkling ballroom, with soft light, (from a low angle:1.2), shot on a Lumix GH5,
with a( dreamy haze filter:1.3), in the style of (David LaChapelle:1.2)"
image = pipe(prompt).images[0]
image