Smart Spectrum version o1: Using GLM-Zero to add animation effects to images

Winter is here. Is it snowing at home?

It doesn't matter if you didn't, now you do--Click here

How did it happen?

Answer:Via GLM-Zero, which is what Smart Spectrum posted two days ago. It looks like a Smart Spectrum advertisement... I'd like to try the same. DeepSeek Chat The "deep thinking".

I use Prompt:

### Drawing an SVG snow animation

##### 1. Background
- Use the given image as the base image.
- Introduce the background image via the `` tag.

##### 2. Snowflake Effect
- **Snowflake Shape**: White dots (`r=1.5`) indicate snowflakes.
- **Quantity**: 10-15 snowflakes falling at the same time, scattered from left to right.
- **Animation details**.
- Snowflakes enter from outside the SVG (`y=-10~-50`) and scatter evenly (`x=0~400`).
- Fade in and out effect:
- Starting transparency: `opacity: 0`.
- Maximum transparency: `opacity: 0.8`
- final transparency: `opacity: 0`
- Random start time (`0-2.5s`).
- Different drop speeds (`3.5-5.5s` a cycle).
- Random horizontal offset (`25-60px`).
- Use separate `` elements for each group of snowflakes, do not use ``.
- **Examples per group**:
``html




</g
``

##### 3. Animation implementation points
- Use `` to set the motion path.
- Stagger the snowflake appearance by different `begin` times.
- The motion path goes from the top to the bottom (400px distance).

Output <svg

Then he will generate an svg effect with a placeholder image address:

Replace the image address and there you have it: https://www.aisharenet.com/wp-content/uploads/2025/01/4f7a57875d3d62c.png

What's the principle?

Public websites and websites support an amazing format called SVG, which is like a folder where you can package images, effects, etc. together. For example, here is the "Dynamic Demonstration of the Big Model Principle".

And for the following one, also SVG, wrapped in 2 layers

The entire code is as follows:

Use it like this:

• First copy the code above
• Open the public edit page and type a line 1111
• Open the browser console (press F12 for Windows, Option+Command+I for Mac) and select the line 1111
• Paste the code in and replace the image with the

Dynamic presentations such as "Our Solar System" can also be generated in one click with GLM-Zero.

The prompt used is this

Please help me create an elegant SVG animation of a solar system in orbit with the following points in mind:
1. the canvas should use a deep space background so that the planets are orbiting right in the center
2. the handling of orbits and distances:
- From the sun outward, the orbital spacing should increase naturally, not too dense and not too sparse
- The orbits are represented by thin lines, which are understated and elegant.
- The innermost circle of planets is about 60 units away from the Sun and the outermost circle about 250 units.
3. The planets are represented in a vivid manner:
- Use emoji (☀️🌍🪐) for the Sun, Earth, and Saturn, so that they are the centerpiece.
- For the other planets, use simple dots in colors that match the planet's characteristics (e.g. #bc6038 for Mars)
- Planet sizes should be to scale, with Jupiter being the largest and Mercury the smallest
4. The animation should be delicate and smooth:
- Running speed should be in accordance with the laws of physics, the farther away from the sun, the slower it rotates
- All animations must be infinite and smooth, without any jumps.
- All animations must be infinite and smooth, without any jumps, with an approximate cycle of 3 to 15 seconds from inside to outside.

Original cue word source:Generation of planetary trajectory maps using the generic "code model" as a substitute for Artifacts/canva

Unlock more with the API

Currently through the web version of GLM-Zero, there are still a little bit of constraints, such as a maximum 1000-word input and no support for multiple consecutive deep thoughts.Much more flexible if you use an API. Documentation is here:

https://bigmodel.cn/dev/api/normal-model/glm-zero-preview

Why GLM-Zero

Unlike GLM, which is officially advertised as being good for mathematical problems, I think this model is better suited for information visualization.

There are 2 main points here:For one, the output text is long enough - capable of more than 30,000 characters, enough to carry more information.

Two:During the thinking process, it will output the entire code in its entirety and then iterate (o1 does not). Such as the following

(dialect) remarry

It must also be admitted that the SVG demos above are still quite scrappy, and can only be considered as an exploration, as is the case with the MidJourney The first edition of the.

However, I am confident that as the modeling capabilities continue to improve, and as we continue to explore, more amazing visualizations will emerge in the future. In this regard, I will also continue to explore, document and share.