I have published an updated set of parameters here: Updated Wan 2.2 T2V, T2I, I2V and Qwen + Wan Low Parameter
I consider the following parameters to be out-of-date. I highly suggest using the updated parameters linked above!
Since the release of Wan 2.2, many of us have been researching ideal parameter and lora combinations to produce high-quality generations quickly. Though many others have shared their parameters, I thought I would include my own.
I use SwarmUI for the most part, but these parameters can easily be repurposed for a Comfy workflow. I'll try to include notes where something might be different between SwarmUI and ComfyUI parameters.
For reference, these parameters should allow you to generate Wan 2.2 videos with decent quality very quickly. For high-quality videos, you'll want to alter these parameters. Some methods to improve quality include increasing resolution or increasing steps while decreasing the strength of lightx2v lora. Either of these methods will also increase generation time.
I've attached the presets themselves to this article so you can download and import them directly instead of trying to parse my instructions below. If you've got questions, comments, or suggestions, leave a comment below!
Wan 2.2 Image-to-Video Parameters
Average Gen Time: 3.25 sec per frame, 65 sec per 20 frames (~1 video second)
Display Advanced Options?: Checked
Sampling
Sampler: Euler
Scheduler: Simple
Resolution
720x720
Note: Lower resolutions may be blurry
Image To Video
Video Model: Wan 2.2 I2V A14B High Noise Q2_K
Video Swap Model: Wan 2.2 I2V A14B Low Noise Q4_K_M
Swap Percent: 0.5
Video Steps: 4
Video CFG: 1
Video Resolution: Image
ComfyUI Note: In a ComfyUI workflow, the Video Model is used in the first KSampler and the Video Swap Model is used in the second KSampler. Also, the first KSampler's end_at_step is (Video Steps x Swap Percent) and the second KSampler's start_at_step is also (Video Steps x Swap Percent) and the end_at_step is (Video Steps).
Advanced Video
FPS: 20-24
Advanced Sampling
Sigma Shift: 8
Other Fixes
Trim Video Start Frames: 4
Note: This is only necessary if you notice a flash of color at the start of your videos
Lora
Lightx2v I2V 14B Rank 256: 1.5 Strength
Note: I haven't noticed a difference in lora ranks
SwarmUI Note: To save this as a preset, include the lora in the preset's Prompt field as: <lora:lightx2v_i2v_14b_rank256:1.5> {value}
Wan 2.2 Text-to-Video Parameters
Average Gen Time: 4.3 sec per frame, 86 sec per 20 frames (~1 video second)
Display Advanced Options?: Checked
Model
Model: Wan 2.2 T2V High Noise 14B fp8 scaled
Core Parameters
Steps: 6
CFG: 5
Sampling
Sampler: Euler
Scheduler: Simple
Resolution
360x360 - 480x480
Note: Lower resolutions may be blurry. Greater resolutions may take much longer to generate.
Text to Video
FPS: 20-24
Note: For longer videos (10 sec+), you can use 10 FPS with 2 Interpolation. But the video may appear somewhat slomo.
Refine/Upscale
Refiner Control Percentage: 0.5
Refiner Method: Step-Swap (SDXL Refiner Model Original)
Refiner Upscale: 2
Refiner Model: Wan 2.2 T2V Low Noise 14B fp8 scaled
Refiner CFG: 1
ComfyUI Note: In a ComfyUI workflow, the Model is used in the first KSampler and the Refiner Model is used in the second KSampler. Also, the first KSampler's end_at_step is (Steps x Refiner Control Percent) and the second KSampler's start_at_step is also (Steps x Refiner Control Percent) and the end_at_step is (Steps).
REFINER-ONLY Lora
Lightx2v T2V 14B Rank 256: 1.5 Strength
Note: I haven't noticed a difference in lora ranks
SwarmUI Note: To apply this to only the refiner, you must not select the lora but instead must include the <refiner> tag with the lora in your Prompt, such as: Some prompt <refiner> <lora:lightx2v_t2v_14b_rank256:1.5
To save this as a preset, include the lora in the preset's Prompt field as: {value} <refiner> <lora:lightx2v_t2v_14b_rank256:1.5>
Advanced Sampling
Sigma Shift: 8
Other Fixes
Trim Video Start Frames: 4
Note: This is only necessary if you notice a flash of color at the start of your videos
Wan 2.2 Text-to-Image Parameters
Average Gen Time: 53 sec
Display Advanced Options?: Checked
Model
Model: Wan 2.2 T2V High Noise 14B fp8 scaled
Core Parameters
Steps: 6
CFG: 1
Sampling
Sampler: Euler
Scheduler: Simple
Resolution
480x480 - 512x512
Note: Lower resolutions may be blurry. Greater resolutions may take much longer to generate.
Text to Video
Frames: 1
Refine/Upscale
Refiner Control Percentage: 0.5
Refiner Method: Step-Swap (SDXL Refiner Model Original)
Refiner Upscale: 2
Refiner Model: Wan 2.2 T2V Low Noise 14B fp8 scaled
Refiner CFG: 1
ComfyUI Note: In a ComfyUI workflow, the Model is used in the first KSampler and the Refiner Model is used in the second KSampler. Also, the first KSampler's end_at_step is (Steps x Refiner Control Percent) and the second KSampler's start_at_step is also (Steps x Refiner Control Percent) and the end_at_step is (Steps).
Lora
Lightx2v T2V 14B Rank 256: 1.5 Strength
Note: I haven't noticed a difference in lora ranks
SwarmUI Note: To save this as a preset, include the lora in the preset's Prompt field as: {value} <lora:lightx2v_t2v_14b_rank256:1.5>
Advanced Sampling
Sigma Shift: 1.5
