Because the V1 version is purely conceptual, but in the process of measuring the style I see many people out of the picture is not very good, and the V1 version will give you a better creative experience if you know how to control the "depth map" and "line drawing".
If you know how to control the "depth map" and "line drawing", then the V1 version will give you a better creative experience.
The previous update of V2 was a bit rushed, so please forgive me for bringing you a bad experience. ......
This time, we are going to re-upload the newly practiced V2, and open the download at the same time, if you have downloaded the V2 version before, please download it again.
So these days have been busy optimizing the Vulcan version of the update ......
1. Optimized the straight out real flame effect
2. Character and flame combination more natural
3. LORA and the major models with good configuration
(* here there is a nonsense, in fact, as long as you will be out of the picture, I personally feel that the V1 version is still better, I personally think so)
zhurongshi, 1girl, solo, hanfu,chinese clothes,blue dress,long dress,fire,Cast magic, cast flame magic,flame ball on the hand,Mythical beasts, glowing,embers,fire hair glowing,embers,fire hair.
Notes:
1. When using a scene, try to weight the scene as much as possible
[Special attention, because it is a special effects class LORA, if you use the size of the raw map is too small, the picture quality is bound to be not too good, so try to make the map larger, in order to get a better picture effect!
2. If you use 512x768, upper body and above try to open the high-resolution fix (Hires. fix), it is recommended to be more than 1.5 times!
Zhurongshi_v2.0
Because the V1 version is purely conceptual, but in the process of measuring the style I see many people out of the picture is not very good, and the V1 version will give you a better creative experience if you know how to control the "depth map" and "line drawing".
If you know how to control the "depth map" and "line drawing", then the V1 version will give you a better creative experience.
The previous update of V2 was a bit rushed, so please forgive me for bringing you a bad experience. ......
This time, we are going to re-upload the newly practiced V2, and open the download at the same time, if you have downloaded the V2 version before, please download it again.
So these days have been busy optimizing the Vulcan version of the update ......
1. Optimized the straight out real flame effect
2. Character and flame combination more natural
3. LORA and the major models with good configuration
(* here there is a nonsense, in fact, as long as you will be out of the picture, I personally feel that the V1 version is still better, I personally think so)
(absurdres, highres, ultra detailed)(masterpiece), (extremely intricate:1.3), (realistic),Cinematic Lighting,4k,(masterpiece),Dreamy. cinematic,Ukiyoe,((divine fantasy,Ethereal Fantasy,action lens,atmospheric,Cloisonne,Gardens,xxTemplexx)),loong
zhurongshi, 1girl, solo, hanfu,chinese clothes,blue dress,long dress,fire,Cast magic, cast flame magic,flame ball on the hand,Mythical beasts, glowing,embers,fire hair glowing,embers,fire hair.
Notes:
1. When using a scene, try to weight the scene as much as possible
[Special attention, because it is a special effects class LORA, if you use the size of the raw map is too small, the picture quality is bound to be not too good, so try to make the map larger, in order to get a better picture effect!
2. If you use 512x768, upper body and above try to open the high-resolution fix (Hires. fix), it is recommended to be more than 1.5 times!
3. Algorithm R-esrgan 4x + anime6B, iteration step 20-40, repainting range 0.5-0.7
4. It is recommended not to open the face repair, use Adetailer>face_yolov8n.pt, the prompt word can not be used.
5. LORA weights are recommended to open about 0.6-0.8
祝融氏Zhurongshi_v2.0
因为V1版练制的纯是概念，然而在测式的过程中我看好多人出图效果并不是很好，
如果您懂得如何控制“深度图”以及“线稿”，那么V1版本给您的创意体验会更好，
之前更新过V2版略为仓促，给大家带来不好的体验有负众望，敬请谅解……
这次打算把新练制的V2重新上传，同时并开放下载，之前下载过V2版本的还请重新下载一次。
所以这几天一直在忙着优化火神版版本的更新……
1.优化了直出真实火焰效果
2.人物与火焰结合更自然
3.LORA与各大模型配置具佳
（*这里有句废话，其实你只要会出图，我个人感觉V1版还是比较好用的，我个人是这么认为）
(absurdres, highres, ultra detailed)(masterpiece), (extremely intricate:1.3), (realistic),Cinematic Lighting,4k,(masterpiece),Dreamy,cinematic,Ukiyoe,((divine fantasy,Ethereal Fantasy,action lens,atmospheric,Cloisonne,Gardens,xxTemplexx)),loong
zhurongshi, 1girl, solo, hanfu,chinese clothes,blue dress,long dress,fire,Cast magic, cast flame magic,flame ball on the hand,Mythical beasts,glowing,embers,fire hair,
注意事项：
1.使用场景时，尽量给场景加权重
【特别注意，因为是特效类的LORA，如果您使用的生图尺寸太小，画质必然不会太好，所以尽量把图生大点，以获更好的出图效果！】
2.如果使用512x768，upper body及以上尽量开高分辨率修复 (Hires. fix)，建议是1.5倍以上
3.算法R-esrgan 4x+ anime6B,迭代步数20-40，重绘幅度0.5-0.7
4.建议不要开脸部修复，使用Adetailer>face_yolov8n.pt，提示词可以不用。
5.LORA权重建议开0.6-0.8左右