请不要添加衣服等人物tag，构图正方形更稳定！

Please do not add people tags such as clothes. The square composition is more stable!

接角色训练，根据模型难度决定（88r起步包修改，搜集素材20r）。

感谢打赏的朋友，QQ：3250911131，

Donate a cup of coffee: https://ko-fi.com/timoplay

Donate electricity AFDIAN（爱发电）：https://afdian.net/a/timoplay

Donate a certain amount of coffee and I will make a lora for you (need to bring your own materials)。Discord：timoplay

小多啦一个人太孤单了，给小多啦找了个小伙伴。(使用时请补全关键词）

Little Dola was too lonely, so I found a little partner for Little Dola.

v3更新：比v2稳定性进一步提高，主要是降低了尾巴的出现频率，推荐权重0.7-1，除真人模型外的大部分模型可以使用，甚至可以配合lora做出特定动作。(Please complete the keywords when using)

v3 update: The stability is further improved compared to v2, mainly because the frequency of tail appearance is reduced, and the recommended weight is 0.7-1. Most of the models except the real model can be used, and it can even cooperate with lora to make specific actions.

哆啦a梦模型，请使用新版V2模型，使用时搭配84000vae，权重1，可在大部分模型上使用，如果想改变动作等请添加关键词，本模型目前存在生图模糊（需要搭配高清修复）的问题待解决，介意的朋友可以去下载v1版本。v2稳定性相对v1提高很多，可以使用的模型也更多，但是出完美哆啦的概率仍然不是很高。

Doraemon model, please use the new version V2 model, use it with 84000vae, weight 1, it can be used on most models, if you want to change the action, please add keywords, this model currently has blurred images (need to be repaired with HD ) problem to be solved, friends who mind can download the v1 version. Compared with v1, the stability of v2 is much improved, and there are more models that can be used, But the probability of a perfect doraemon is still not very high.