/recipes/Pyros13/Wan2.2-VACE-Fun-A14B-Diffusers.md

Wan2.2 VACE-Fun Diffusers

All-in-one video creation and editing with Wan2.2 VACE-Fun A14B in diffusers format.

Summary

• Vendor: Pyros13
• Model: Pyros13/Wan2.2-VACE-Fun-A14B-Diffusers
• Base model: alibaba-pai/Wan2.2-VACE-Fun-A14B
• Task: VACE video generation and editing (T2V / I2V / R2V / V2V / inpainting)
• Mode: Offline inference and OpenAI-compatible online serving
• Maintainer: Community

When To Use This Recipe

Use this recipe to run Wan2.2-VACE-Fun-A14B through the standard diffusers-style Wan22VACEPipeline path in vLLM-Omni. The checkpoint is a diffusers conversion of alibaba-pai/Wan2.2-VACE-Fun-A14B, with model_index.json, transformer/, transformer_2/, text_encoder/, tokenizer/, vae/, and scheduler/ subfolders. Because the model already uses the expected diffusers layout, vLLM-Omni does not need any original-format key remapping or .pth component loaders.

The model is a two-expert architecture. The high-noise transformer and low-noise transformer_2 are switched by the Wan2.2 boundary during denoising.

References

• Model card: https://huggingface.co/Pyros13/Wan2.2-VACE-Fun-A14B-Diffusers
• Base model card: https://huggingface.co/alibaba-pai/Wan2.2-VACE-Fun-A14B
• Related example: examples/offline_inference/vace/vace_video_generation.md
• Related issue: vllm-project/vllm-omni#4206
• PR: vllm-project/vllm-omni#4667

Offline Inference

python examples/offline_inference/vace/vace_video_generation.py \
  --model Pyros13/Wan2.2-VACE-Fun-A14B-Diffusers \
  --mode i2v --image ./i2v_input.JPG \
  --prompt "Summer beach vacation style, a white cat wearing sunglasses sits on a surfboard" \
  --height 480 --width 832 --num-frames 81 --num-inference-steps 30 \
  --seed 42 --guidance-scale 5.0 \
  --vae-use-tiling --enforce-eager \
  --output vace_fun_i2v.mp4

See examples/offline_inference/vace/vace_video_generation.md for other VACE modes and their input flags.

Online Serving

Server

vllm serve Pyros13/Wan2.2-VACE-Fun-A14B-Diffusers --omni \
  --model-class-name Wan22VACEPipeline \
  --vae-use-tiling \
  --enforce-eager \
  --port 8091

For multi-GPU sequence-parallel serving, add the same parallelism flags you would use offline, for example --ulysses-degree 4. The explicit --model-class-name is optional for this diffusers checkpoint, but keeping it in examples makes the selected pipeline obvious.

Client

no_proxy=127.0.0.1 \
curl -X POST http://127.0.0.1:8091/v1/videos/sync \
  -F "prompt=Summer beach vacation style, a white cat wearing sunglasses sits on a surfboard" \
  -F "input_reference=@/absolute/path/to/i2v_input.JPG" \
  -F "width=832" \
  -F "height=480" \
  -F "num_frames=81" \
  -F "fps=16" \
  -F "num_inference_steps=30" \
  -F "guidance_scale=5.0" \
  --output vace_fun_i2v_serve.mp4

For video-to-video or inpainting-style requests, upload the reference video with input_reference=@/absolute/path/to/input.mp4 or use video_reference with a URL or JSON-safe data URI.

Cache-DiT / TaylorSeer Notes

Wan22VACEPipeline uses the shared Wan2.2 Cache-DiT enabler for both the high-noise and low-noise transformers. The adapter intentionally wraps only the main denoising blocks. VACE vace_blocks form a conditioning branch: each step they combine the current latent with the VACE context to produce hints that are then injected into selected main blocks. Recomputing that branch preserves the control signal; caching the main backbone still gives the acceleration target.

Validation

Validated on NVIDIA B300 with mkt_part_000.mp4, 736x1280, 61 frames, 30 inference steps, seed=1, and boundary_ratio=0.875.

The Pyros13 diffusers checkpoint produced a byte-identical MP4 to the alibaba-pai/Wan2.2-VACE-Fun-A14B original-format checkpoint under the same vLLM-Omni pipeline and inputs:

• SHA256: 2e2db74a612a009b0d6ad6d739e039883f0de84ce5cdf2676324daf78fb05ef8
• Frame shape: 61 x 736 x 1280
• MAE / RMSE / max pixel diff: 0 / 0 / 0
• PSNR: infinite
• Mean SSIM: 1.0