ASK STYLE MIX – v1.0
| Run Model | 使用GRAVITI Diffus 关于GRAVITI Diffus |
| Model Name | ASK STYLE MIX |
| Model Type | Checkpoint |
| Base Model | SD 1.5 |
| Version Name | v1.0 |
| File Name | askSTYLEMIX_v10.safetensors |
| SHA256 | dec89d4a32 |
| NSFW | False |
| Trigger Words | |
| Tags | anime, style, illustration, ask |
| Metadata | fp = fp16 size = pruned format = SafeTensor |
| Size | 2.03 GB |
| Version Description |
1.0版本，我个人最推荐的VAE是 clear vae，你可以在civitai上下载。 |
| Source | https://civitai.com/models/34578 |
| Author | dherorunner |
我使用dreambooth在anything3.0上用ASK的插画作品进行微调，得到了ASK风格的checkpoint。之后使用MBW用它和cetus进行混合，在保留画风和笔触的基础上，吸纳cetus的构图和一定的人物构图（因为我发现生成的ASK checkpoint在这些方面并不灵活）
例图使用了dynamic threshold这个插件，动态调节CFG，以获得更多样的色彩呈现。
I used dreambooth to fine-tune with ASK’s illustrations on anything3.0, and got an ASK-style checkpoint. Then use MBW to mix it with cetus. On the basis of retaining the painting style and strokes, absorb the composition of cetus and the composition of characters to a certain extent (because I found that the generated ASK checkpoint is not flexible in these aspects)
The example picture uses the dynamic threshold plug-in to dynamically adjust CFG to obtain more diverse color presentation.
相关文章
