For those who appreciate the "imminent facesitting" low-angle shot, this is for you. At weaker settings, and in between ideal outputs, it will generate other nice low angles.
Trained on NAI from 91 custom images, generated from a small number of photo sources.
Sweet spot is around 0.7-0.8, consistency may be improved by the "from below" tag.
Known Issues: Often causes distortion of the nose. Inpainting works as a fix, but I am investigating if this can be avoided in a v1.1
Please post your best work, or share it with me, and I may include the best stuff in a future revision to improve this model's versatility and range. As of right now, it's reasonably flexible, but there's always room for improvement.
---------------
推荐给那些欣赏 "当面 "低角度拍摄的人。它在理想的输出之间或在较弱的设置下产生其他不错的低角度。
NAI由91张自定义图像训练而成，并由少量照片来源生成。
甜点在0.7-0.8左右，一致性可以通过 "from below "的标签来提高。
已知问题：经常出现鼻子失真。 可以通过内画来纠正，但正在调查V1.1版中是否有解决的办法。
为了提高这个模型的通用性和范围，请张贴或与我分享你最好的作品，因为我可能会在未来的修订中包括最好的作品。 目前，它是相当灵活的，但总是有改进的余地。
---------------
イマドキの顔出し」ローアングル撮影を評価する方におすすめです。理想的な出力の合間や、弱めの設定で、他の素敵なローアングルを生成してくれます。
91枚のカスタム画像からNAIをトレーニングし、少数の写真ソースから生成。
スウィートスポットは0.7-0.8程度で、「from below」タグによって一貫性が改善されるかもしれません。
既知の問題点： 鼻の歪みが発生することが多い。インペイントで修正できるが、v1.1で回避できるか調査中。
このモデルの汎用性と範囲を向上させるために、将来の改訂にベストなものを含めるかもしれませんので、あなたのベスト作品を投稿するか、私と共有してください。今のところ、それなりに柔軟性はありますが、常に改善の余地があります。
Description
FAQ
Comments (24)
extremely based post
Fantastic job! May I expect for more POV looking from below, like ass shot or POV while walking?
Thanks - I'm considering a version of this from behind, maybe.
@NeuralEros please do so, it would be very useful!
Been wanting a lora of this pose.
This is awesome! I'm also trying to train a LoRA on a pose. Did you have to use any different settings for the training? I'd love as much info as possible on your process.
I just followed some guide I found to train character loras and hoped it would work out - it seems like it more or less did. YMMV
I think it would be interesting to add "closed leg" pose, and "underboob through bottom of the shirt".
I am having hard time trying both with prompts, but not quite getting there.
not working for me. even using the sample images. i get nothing
Try changing the name of the call for the lora in the prompt to "<lora:povsitting:0.5>" or vice-versa; name it "<lora:povFacesitting:0.5>" if it's the other. I first used and it was lora:povFacesitting then it stopped working. downloaded it again and it would only work with lora:povsitting. Know they both work with a lil dif styles to them. Point is they are fabulous lora (hat tip to NeuralEros)
Worth it to give this a shot in my opinion - one of my fav loras
@Cosmic_Willis curious, are you on A1111?
@NeuralEros I am indeed. Been having probs lately with LORAs and lycoris and xformers (I struggle lol) but it's working right at the moment :D
got it working. i added it to AddtionalNetworks extension instead
This needs a futa version xD
God, I wish. Not enough source material.
Sooo....this works even better than 'upskirt' Loras around here!
I'm havving this issue with the engine. Any way to fix it?
*** Error completing request *** Arguments: ('task(msl2qv3068jqgua)', 'girl', '', [], 20, 0, False, False, 1, 1, 7, -1.0, -1.0, 0, 0, 0, False, 512, 512, False, 0.7, 2, 'Latent', 0, 0, 0, 0, '', '', [], <gradio.routes.Request object at 0x000001D9549A5960>, 0, False, False, 'positive', 'comma', 0, False, False, '', 1, '', [], 0, '', [], 0, '', [], True, False, False, False, 0) {} Traceback (most recent call last): File "D:\Stable_Diffusion\stable-diffusion-webui\modules\call_queue.py", line 58, in f res = list(func(*args, **kwargs)) File "D:\Stable_Diffusion\stable-diffusion-webui\modules\call_queue.py", line 37, in f res = func(*args, **kwargs) File "D:\Stable_Diffusion\stable-diffusion-webui\modules\txt2img.py", line 62, in txt2img processed = processing.process_images(p) File "D:\Stable_Diffusion\stable-diffusion-webui\modules\processing.py", line 677, in process_images res = process_images_inner(p) File "D:\Stable_Diffusion\stable-diffusion-webui\modules\processing.py", line 794, in process_images_inner samples_ddim = p.sample(conditioning=p.c, unconditional_conditioning=p.uc, seeds=p.seeds, subseeds=p.subseeds, subseed_strength=p.subseed_strength, prompts=p.prompts) File "D:\Stable_Diffusion\stable-diffusion-webui\modules\processing.py", line 1054, in sample samples = self.sampler.sample(self, x, conditioning, unconditional_conditioning, image_conditioning=self.txt2img_image_conditioning(x)) File "D:\Stable_Diffusion\stable-diffusion-webui\modules\sd_samplers_kdiffusion.py", line 464, in sample samples = self.launch_sampling(steps, lambda: self.func(self.model_wrap_cfg, x, extra_args={ File "D:\Stable_Diffusion\stable-diffusion-webui\modules\sd_samplers_kdiffusion.py", line 303, in launch_sampling return func() File "D:\Stable_Diffusion\stable-diffusion-webui\modules\sd_samplers_kdiffusion.py", line 464, in <lambda> samples = self.launch_sampling(steps, lambda: self.func(self.model_wrap_cfg, x, extra_args={ File "D:\Stable_Diffusion\stable-diffusion-webui\venv\lib\site-packages\torch\utils\_contextlib.py", line 115, in decorate_context return func(*args, **kwargs) File "D:\Stable_Diffusion\stable-diffusion-webui\repositories\k-diffusion\k_diffusion\sampling.py", line 145, in sample_euler_ancestral denoised = model(x, sigmas[i] * s_in, **extra_args) File "D:\Stable_Diffusion\stable-diffusion-webui\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "D:\Stable_Diffusion\stable-diffusion-webui\modules\sd_samplers_kdiffusion.py", line 215, in forward devices.test_for_nans(x_out, "unet") File "D:\Stable_Diffusion\stable-diffusion-webui\modules\devices.py", line 155, in test_for_nans raise NansException(message) modules.devices.NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check.Did you
"Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this." like the error says, he asked out of morbid curiosity.
Hey bro, can you post this LoRA on tensorart?
SDXL please
What are you suppose to do with this? I'm new
Jerk off?
its ok
Are you using A1111? The model does not appear for me to select.
Details
Files
povFacesitting.safetensors
Mirrors
povFacesitting_v10.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
pov-facesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting_v10.safetensors
povFacesitting.safetensors
povFacesitting_v10.safetensors
povFacesitting.safetensors
povFacesitting_v10.safetensors
povFacesitting_v10.safetensors
povFacesitting_v10.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
povFacesitting.safetensors
facesitting.safetensors
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.