LandscapeBing v1.0/Glacial Acetic Acid Versatile Landscape v1.0

After listening to the suggestions of Brother zzz, I gave the model a name that is easy to search and identify. The model content is the same as the previous version.

This model can draw concepts on multiple landscapes, so I hope everyone can enjoy it.

Prompt:

1. Flowers: Garden, Flowers can be used with anything else; with Gravel Road, you can draw a gray gravel road 2. Station: yz_station,

service station, with others

Architecture-related terms, can be used for post station-related

3. Classical gardens: SZYL, East Asian architecture, architecture, classical gardens, with other terms, groups of corresponding styles, stones, etc. 4. Autumn scenery: autumn, autumn leaf trees, matching

Other words can appear in autumn leaf tree species and colors

5. Night lighting: night, light. When used with other words, late-night night lighting effects can appear 6. Rainy scene:

rain, fog, enter a few more times to increase the weight, and use other words to increase the weight. 7. Randomly

You can enter it under the conditions: ZHIGAN, as a trigger word to improve the rendering style of the picture

8. In terms of materials, if you have training on wood (wood) and bricks (bricks), it is recommended to use long sentences to describe natural language. Enter the suggestions a few more times for things that

can't come out

Figure, 1024*576 is good, cfg7 and below. If it's too serious after adding Lora, you can try adjusting the clip skip layer to 4.6, and you can choose to turn on the dynamic cfg coefficient (a plug-in) to reduce image overfitting. You can use it with bad-picture-chill-75v, and other negative embedding

You can use GPT to automatically input prompts, and the results are good, and the results are quite satisfactory when used with controlnet. As a domestic student work, I think it's quite enough. As a large-scale landscape model created and mixed independently, this model has the following

characteristics:

This model is a slightly overfitted model with a targeted nature (a few words can output a good-looking image). It's not very general enough, and the style of painting is often perfect. It's not of high quality when judged by big models, but it still comes out after thinking about it, so that more people can see the landscape and draw a good picture using SD, giving more people some thought.

This model is suitable for the following people: people who are new to SD and want to make good pictures right away, people who can't train Lora on their own, people who want to practice using prompts and drawing good-looking cards, and people who want to use a big model to quickly draw a few good pictures.

This model is not suitable for the following people: people who want to train the Lora model themselves after exploring many models, people who want to go through their own training to plot with a large model for a long time after going through a novice period (I still recommend chill here), and people who think that the effects of SD are mainly determined by the big model after downloading a model that is close to fitting.

This model will be used in some future domestic courses.

Author's qq name: glacial acetic acid (if any friends know me, hehe)