AnimateDiff
AnimateDiff is an official software project that transforms text-to-image model from text developed by the community into animation manufacturers without the need for additional training. The project aims to learn transferable motion priors (motion priors), which can be applied to different Stable Diffusion variants. It mainly supports Stable Diffusion V1.5 and sdxl-beta branches, as well as offers Diffusers integration and sampling scripts.
guoyww · Local AI / Self-Hosted LLM
Common commercial products in this category: ChatGPT, Claude, Gemini
- Using community models, you run text-based animation production scripts.
- You provide motion control with MotionLoRA configuration files.
- By starting the Gradio interface on the local server, you manage the process through the visual interface.
- • Last update 2 years ago
- • NeutralGitHub Issue
EinopsError Shape mismatch
“Hi, I am encountered this issue: **EinopsError: Error while processing rearrange-reduction pattern "(b f) d c -> (b d) f c". Input tensor shape: torch.Size([2, 8040, 320]). Additional info: {'f': 16}. Shape mismatch, can't divide axis of length 2 in chunks of 16** **Please guide…”
- ↑ PositiveGitHub Issue
LCM sampler support
“Greetings! there is a new sampler named LCM, it can produce amazing images with very few steps as low as 1-4 steps and it is very fast compared to other samplers they are planning to bring it to automatic1111, i just wanted to say please support this sampler and their version of…”