I decided to post my workflow. Maybe it’ll be useful to someone. I’ve tried a lot of different workflows, but they just didn’t work for me, and most of them simply didn’t run on my Predator Helios 300 laptop (1050ti, 4GB VRAM, 16GB RAM). And if anyone manages to improve it, I’d be more than happy.P.S.A request to those who downloaded these files: please post your results in the gallery.links:wan2_2-5b-ti2v-Turbo-Q4_K_M.gguf<a target="_blank" rel="ugc" href="https://huggingface.co/hum-ma/Wan2.2-TI2V-5B-Turbo-GGUF/tree/main">https://huggingface.co/hum-ma/Wan2.2-TI2V-5B-Turbo-GGUF/tree/main</a>nsfwWanUMT5XXLGGUF_q5AndQ4KM.gguf<a target="_blank" rel="ugc" href="https://civarchive.com/models/2352303/nsfw-wan-umt5-xxl-gguf">https://civarchive.com/models/2352303/nsfw-wan-umt5-xxl-gguf</a>NSFW-Anime-wan-5B_3.0Pro_lora_rank_128.safetensors<a target="_blank" rel="ugc" href="https://huggingface.co/Hobbidy/NSFW-Anime-wan-5B_3.0Pro_lora_rank_128/tree/main?not-for-all-audiences=true">https://huggingface.co/Hobbidy/NSFW-Anime-wan-5B_3.0Pro_lora_rank_128/tree/main?not-for-all-audiences=true</a>RealESRGAN_x2plus<a target="_blank" rel="ugc" href="https://huggingface.co/hlky/RealESRGAN_x2plus/tree/main">https://huggingface.co/hlky/RealESRGAN_x2plus/tree/main</a>TAEWput in to \models\vae_approx<a target="_blank" rel="ugc" href="https://huggingface.co/lightx2v/Autoencoders/blob/main/lighttaew2_2.safetensors">https://huggingface.co/lightx2v/Autoencoders/blob/main/lighttaew2_2.safetensors</a>

WAN 2.2 5b low vram workflow

Changes have been made to optimize the process, and a two-step decoding process has been added. First, the generated output passes through an ultra-fast TAEW to preview the result, and then through a full-fledged VAE. Upscaling has also been added.

Files with SHA256: 65033c4a634e9d7eac6de05e511dd3a255442ac30ba778eda7f641aee3b2974d