Wan2.1 parameter sweep

# Wan2.1 parameter sweep – Replicate blog

[Replicate has joined Cloudflare](https://replicate.com/blog/replicate-cloudflare)

[](https://replicate.com/)

cmd+k

[Explore](https://replicate.com/explore)[Pricing](https://replicate.com/pricing)[Enterprise](https://replicate.com/enterprise)[Docs](https://replicate.com/docs)[Blog](https://replicate.com/blog)[Sign in](https://replicate.com/signin)[Try for free](https://replicate.com/signin)

Menu

[Explore](https://replicate.com/explore)[Pricing](https://replicate.com/pricing)[Enterprise](https://replicate.com/enterprise)[Docs](https://replicate.com/docs)[Blog](https://replicate.com/blog)[Sign in](https://replicate.com/signin)[Try for free](https://replicate.com/signin)

* * *

[Compare models in the Playground](https://replicate.com/playground)

* Replicate
* [Blog](https://replicate.com/blog)

# Wan2.1 parameter sweep

Posted March 5, 2025 by
* [zsxkib](https://replicate.com/zsxkib)

We’ve been playing with [Alibaba’s WAN2.1 text-to-video model](https://replicate.com/blog/wan) lately. Like most image and video generation models, Wan has a lot of input parameters, and each of them can have a profound impact on the quality of the generated output.

What happens when you tweak those mysterious inputs? Let’s find out.

## [](https://replicate.com/blog/wan-21-parameter-sweep#the-experiment)The experiment

We wanted to see how the `guidance scale` and `shift` input parameters affect the output. For our experiment, we used the [WAN2.1 14b text-to-video model](https://replicate.com/wavespeedai/wan-2.1-t2v-720p) with 720p resolution.

To do this, we did what’s called a “parameter sweep”, systematically testing different combinations of input values to understand how they affect the output. We generated videos for each combination of guidance scale and shift values, keeping all other parameters constant.

We kept the following inputs consistent across all the videos:

* `prompt`: `"A smiling woman walking in London at night"`
* `seed`: `42`
* `frames`: `81`
* `sample_steps`: `30`

We then varied just these two inputs, testing against a range of values:

* `sample_guide_scale`: from 0 to 10
* `sample_shift`: from 1 to 9

If you’d like to run similar experiments yourself, we’ve shared [the code on GitHub](https://github.com/zsxkib/blog-wan-2.1-param-sweep) that we used to generate these parameter sweeps.

## [](https://replicate.com/blog/wan-21-parameter-sweep#what-is-guide-scale)What is guide scale?

You can think of the guide scale as the “creativity vs obedience” knob.

At `guide_scale=0`, the model ignores your prompt. As you increase the value, the model tries harder to match your prompt.

* Lower values: more creative freedom.
* Higher values: more literal interpretation.

Here’s what happens when you dial it from 0 to 10:

## [](https://replicate.com/blog/wan-21-parameter-sweep#what-is-shift)What is shift?

Shift controls how the model moves through the denoising process, affecting motion and time flow in your video.

It’s basically controlling the “flow of time” in your generated video.

* Lower values: smoother, more predictable movement.
* Higher values: more dynamic but sometimes chaotic motion.

Here’s what happens when you change shift from 1 to 9:

## [](https://replicate.com/blog/wan-21-parameter-sweep#what-were-seeing-in-these-videos)What we’re seeing in these videos

For guide scale:

* `guide_scale=0`: Really weird but cool outputs. Creative but barely related to the prompt.
* `guide_scale=1-2`: Strange artifacts, especially around the woman’s mouth.
* `guide_scale=3-7`: 👈 The sweet spot. Natural looking with minimal issues.
* `guide_scale=8+`: The dreaded “AI look” creeps in - that overcooked, shiny skin that screams “I was made by AI.”

Recommendation: Use 0 for weird creative stuff, 3-7 for realistic results, and avoid 8+ unless you want that AI shine.

For shift values (all with `guide_scale=5`):

* `shift=1`: Creates a cool “dolly

5 Levels of AI Agents - From Simple LLM Calls to Multi-Agent Systems

Dave Ebbelaar (LLM Eng)

Chapter 3: Looking Inside Large Language Models | Hands-On Large Language Models Book

onepagecode

Hands-On Large Language Models | Chapter 7: Advanced Text Generation Techniques

onepagecode

Hands-On LLMs - Chapter 1: An Introduction to Large Language Models

onepagecode

Chapter 2: Tokens and Embeddings | Hands-On Large Language Models Book

onepagecode

Hands-On Large Language Models | Chapter 5: Text Clustering and Topic Modeling

onepagecode