| # How to make | |
| 1. pretrained model: [epiCRealism](https://civitai.com/models/25694?modelVersionId=134065) + [hyper CFG lora 12steps](https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SD15-12steps-CFG-lora.safetensors) | |
| -> merge with lora weight 0.3 | |
| 2. model merged at step1 + lora model: [AnimateLCM_sd15_t2v_lora.safetensors](https://huggingface.co/wangfuyun/AnimateLCM/blob/main/AnimateLCM_sd15_t2v_lora.safetensors)-> merge with lora weight 0.8 | |
| ```python | |
| # Load the motion adapter | |
| adapter = MotionAdapter.from_pretrained("guoyww/animatediff-motion-adapter-v1-5-3", torch_dtype=torch.float16) | |
| # load SD 1.5 based finetuned model | |
| model_id = "/home/hyejin2/test/models/epiCRealism-hyper-LCM-8.safetensors" | |
| pipe = AnimateDiffVideoToVideoPipeline.from_single_file(model_id, motion_adapter=adapter, torch_dtype=torch.float16) | |
| pipe.save_pretrained("models/hello") | |
| ``` | |
| # How to use | |
| ```python | |
| model_id = "jstep750/animatediff_v2v" | |
| pipe = AnimateDiffVideoToVideoPipeline.from_pretrained(model_id, torch_dtype=torch.float16) | |
| # enable memory savings | |
| pipe.enable_vae_slicing() | |
| pipe.enable_model_cpu_offload() | |
| ``` |