Wan2.2 based Transitions and Frames-2-video

🎬 Morphic Frames to Video

High-quality video generation from image frames using Wan2.2

GitHub Hugging Face Website

Video Input Images & Prompt
β†’

Prompt: "A clown, slowly transforms into a poster."

Type: Two-frame transition
β†’ β†’ β†’

Prompt: "The animated girl rises up from her chair and waves hi to the camera as the camera zooms in."

Type: Multi-frame interpolation

Setting up the repository

First clone the Morphic Interpolation repo:

git clone https://github.com/morphicfilms/frames-to-video.git

To install the environment, we recommend following the Wan2.2 installation guide.

Or you could alternatively run : bash setup_env.sh -> we recommend using the flash-attn version listed in the bash file for hassle free install.

Downloading the weights

First: download Wan2.2 I2V weights:

huggingface-cli download Wan-AI/Wan2.2-I2V-A14B --local-dir ./Wan2.2-I2V-A14B

Second : download the Morphic Frames to Video lora weights :

huggingface-cli download morphic/Wan2.2-frames-to-video --local-dir ./morphic-frames-lora-weights

Running Frames to Video

For multi node run for 2 frame interpolation :

torchrun --nproc_per_node=8 generate.py \
    --task i2v-A14B \
    --size 1280*720 \
    --frame_num 81 \
    --ckpt_dir ./Wan2.2-I2V-A14B-Interpolation \
    --high_noise_lora_weights_path ./morphic-frames-lora-weights/lora_interpolation_high_noise_final.safetensors \
    --dit_fsdp \
    --t5_fsdp \
    --ulysses_size 8 \
    --image examples/transition9_1.png \
    --prompt "Aa clown, slowly transforms into a poster." \
    --img_end examples/transition9_2.png \

For multi node run for multi frame interpolation :

torchrun --nproc_per_node=8 generate.py \
    --task i2v-A14B \
    --size 1280*720 \
    --frame_num 81 \
    --ckpt_dir ./Wan2.2-I2V-A14B-Interpolation \
    --high_noise_lora_weights_path ./morphic-frames-lora-weights/lora_interpolation_high_noise_final.safetensors \
    --dit_fsdp \
    --t5_fsdp \
    --ulysses_size 8 \
    --image examples/pink_1.png \
    --prompt "The animated girl rises up from her chair and waves hi to the camera as the camer zooms in." \
    --img_end examples/pink_4.png \
    --middle_images examples/pink_2.png examples/pink_3.png \
    --middle_images_timestamps 0.4 0.7

Note:

  1. --middle_images_timestamps : should be used if multiple intermediate frames are provided, the numbers indicate the the location the intermediate frame is provided (0.5 -> midway, 0.33, 0.66 -> 2 equally spaced intermediate frames, 0.25, 0.5, 0.75 -> 3 equally spaced intermediate frames)
  2. Number of middle_images must be equal to number of middle_images_timestamps
Downloads last month
-
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ 3 Ask for provider support

Model tree for morphic/Wan2.2-frames-to-video

Adapter
(82)
this model