Wan video 2.2 text to video test with self-forcing LoRA
All the clips in this video were generated on an RTX 3090 GPU at 1024×576 resolution. I used GGUF Q8 Text-to-Video models with self forcing lora, so 72 frames could be generated on 24 GB of VRAM. I generated prompts on a wide variety of topics, aiming to make the selection as diverse as possible. If you’re curious about the possibilities of the latest models, this video is for you!