Stable Diffusion’s Animated Collaboration with Poe and Kills: A Must-See for AI Enthusiasts
Hello again, my little neural enthusiasts! Today you will see an amazing and terrifying collaboration between the Stable Diffusion neural network and Edgar Allan Poe and Natali Kills.
My new work, a recording of an actress reciting Poe’s “The Raven“ once came to me. But the recording had been altered, edited, and overlaid with music. I split the poem into two-line segments and provided timestamps for each reading. This is how the Stable Diffusion neural network sees Edgar Allan Poe’s poetry. I also added stylistic prompts to mimic Poe’s style, which I think suits the project perfectly.
After 16 hours of rendering for the 1024x512 version, and another 10 hours of upsampling by 4x, we have this piece of beauty in our hands. Of course, when you use such complex connections from 3rd party systems, something always goes wrong. But I think we managed to make a cool piece of art.
Sure, something went wrong when you use such complex connections from 3rd party systems, something always goes wrong. The ESRGAN_4x and R-ESRGAN 4x upsamplers work great, but they blur the texture of a pencil drawing, which is unacceptable to us. Lanczos is a great upsampler, it works fast and keeps the texture intact, but it runs in memory, literally within an hour Python was already taking up 25GB of my RAM (joking about the snake that eats too much, ha-ha, it’s Python, you know it). I had to run the process 4 or 5 times, restarting when the memory ran out.
As usual, I used the keyframe generator in a pretty simple way. I would be very grateful if you could provide me with a complete guide on how to create mathematical functions specifically for animation. If you have such a link, please let me know, I would be grateful and able to make more interesting things. Alternatively, you can just write to me and give me good advice on how to make a universal formula for 2D perspective rotation, or for dynamic zoom in both directions according to the BPM of my track.
Хотите научиться создавать такие же нейро-шедевры? Участвуйте в онлайн-интенсиве НЕЙРОБУСТ и изучите все основные инструменты и технологии для работы с нейросетями всего за два дня!
Посмотрите программу и зарегистрируйтесь
Вот список полезных ресурсов:
• Арт-нейросети от Nerual Dreming -
• Nerual Dreming на Boosty -
• ДЕНЬГИ - Lcq3XlPTFLE3NWIy
• ЧЕЛЛЕНДЖИ -
• СКЛАД -
• АДМИНСКАЯ -
• Чат нейро-операторов -
Подписывайтесь на эти ресурсы, чтобы получать новые знания и быть в курсе всех событий в мире нейросетей!
4 views
432
96
2 months ago 00:04:00 1
Я усталым таким ещё не был... (Сергей Есенин)
2 months ago 00:02:37 1
Заметался пожар голубой (Сергей Есенин) - [ver.2]
2 months ago 00:46:14 1
😎 Создание последовательного персонажа с помощью LoRA для FLUX - лучшей модели генерации изображений
2 months ago 00:05:31 1
Сектор газа - Частушки (НЕЙРОСЕТЬ)
2 months ago 00:05:19 1
АЛИСА
2 months ago 00:04:31 1
Ты меня не любишь, не жалеешь (Сергей Есенин)
2 months ago 00:05:18 1
LUMEN - Гореть (2023) | official video
2 months ago 00:02:51 1
Мне грустно на тебя смотреть... (С. Есенин) - [Ver.1]
2 months ago 00:00:19 1
Yuki Suou 周防 有希 - Crazy - LE SSERAFIM
2 months ago 00:01:27 1
Auto Painter AI - Texture Generation for 3D Models with AI in Blender
2 months ago 00:03:01 1
Песня солнечного луча (Саша Чёрный) - [Песни для детей]
2 months ago 00:03:19 1
4K, Laura’s Ai Art - Pink and Fluffy (Lookbook)
2 months ago 01:56:31 1
Как рисовать в казуальном стиле?
2 months ago 00:08:38 2
Ракурсы для генерации изображений на примере миньонов во FLUX.1 [dev]. Краткая инструкция #нейросети
2 months ago 00:03:42 1
Дорогая, сядем рядом...(Сергей Есенин) - [ver.1]
2 months ago 03:11:07 1
Как войти в топ 1% по использованию ChatGPT и других ИИ-инструментов? - Руслан Сыздыков, Higgsfield
2 months ago 03:51:22 1
Как мыслит нейросеть? Теория, кейсы «ЦЕХа» и «сломанный нейротелефон»