dont know why but hands are large in general

It is more realistic, but there are many cases that are more ugly. It is changing from really pretty normal female models to just pretty normal models.

V5 is hard to use lora

Honestly, I guess the V5 version is worse than V4, at least in case of hands and feet.

为什么我出的图都很丑啊，不用lora出的人脸不漂亮，用了lora，出的人基本大概率脸会崩坏，就算不坏，也怪怪的，不好看

Yo this model is lit.🔥 Thank you @pleasebankai

Whether to consider increasing the training set for men

I love it thank you, it is so much fun to use.

https://www.instagram.com/ai4hana/

I would highly appreciate it if you made Pickletensor available as well.

too many logos + watermarks

bra......nice name

The model name changed. Is the new check point only generate one person?

Amazing! If you start a Patreon account, I'll be glad to support you there (I can't use PayPal).

face too long

great work. Tried to do it myself, but can't find the "godpussy1" lora, would you kindly post a link,plz?

U meant your title of the model just changed without you knowing?

bravo

-nods- braaaa

このモデルで生成した画像を販売してもいいですか？ライセンス読みましたが、念の為確認です。

首先謝謝大神的無私分享，但还是想給大神一個小小小的建議，這個大模型弱光晚上的展示欠佳。人物臉部容易過於明亮，期待你的更新！！

Does this come with a VAE? The 512x512 pictures I've generated have all looked pixelated and prompting doesn't fix this. I'm using vae-ft-mse-840000 with this model but it doesn't seem to be helping the end result.

1 girl, 48years old, best quality, photorealistic, 8k, high res, photorealistic, (bokeh), (portait:0.6), (dynamic pose:1.2), masterpiece, intricate, realistic, sharp focus, award-winning photograph, (Smile:1), lora:aiGRAVNAGIXx_v1:0.7

Great work! THANKS

sorry newbie here, 😭 Can i ask something? How to remove ribs / rib cage? i put it in neg. Prompt, but it still appear, and make my model looks so skinny. maybe there is prompt / negative prompt that can make my model's body looks normal (not too skinny nor too fat) thank you very much 🙏

Love you~

can this model only create asian woman, or is it possible to create other ethnicities??

Does this checkpoint do asian men well? Many asian char checkpoint does many tranny when dealing male realted prompts.

Excuse me, how many pictures are there in the dataset? How is the proportion of faces in various Asian countries composed?

Found a problem: When generating the whole body, it is very easy to have too many legs, and it is difficult to remove them with reverse prompts

我创建了一个服务器我们可以交流然后相互帮助学习lora什么的分享图片什么请大家多多支持https://discord.gg/VFxs3dUsa5

非常糟糕的一款模型，脸型的适配度极差，无论怎么叠LORA都改变不了大模型的丑脸，而且及其容易毁脸，不推荐

How can you have a distinctive faces on this base model?

I love it，easy to make。

Can you upload Version2 again? :)

Loved version 4! Version 5 makes everyone chubby, I can't even use it anymore. Any chance of bringing back the amazing version 4????

Why is there always texts/logo at the right bottom ? Can you remove it in brav6?

感谢你提供的模型，整体来说还是不错的，训练不容易，加油。

这个模型真好，请问怎样才能联系到您，我们平台诚心邀请您入驻

awesome!

Which version produces the best results?

看到实例有人用<lora:su6 bra1:0.6>，这个lora在哪下载，没找到。

If I were to put it bluntly, I believe even the creators of this V5 model are not entirely satisfied with its outcome. The unique prompt structure can be overly peculiar. When using the same prompt structure with other models, the results can be completely different from V5's composition. It's just too distinctive. While it may suffice for simple character illustrations, it becomes challenging to give more complex instructions. You can easily notice that this model predominantly produces upper-body portraits and standing poses. However, I hope that efforts will be made to improve it. I expect a mass-produced, realistic model with individualized features, rather than everyone having the same face.