Zhen Li comments

Results 197 comments of


                                            Zhen Li

crop了几张迪丽热巴的照片，生成的结果不太像

![image](https://github.com/TencentARC/PhotoMaker/assets/21050959/ff5d55e1-4799-4d60-858f-dfe6cd3e8cea)

We really can't guarantee perfect ID consistency in some examples, especially on Asian faces. We are still working hard to improve. Please see this [discussion](https://github.com/TencentARC/PhotoMaker/discussions/14) as a solution.

crop了几张迪丽热巴的照片，生成的结果不太像

Hi all, we update the PhotoMaker V2 (with improved ID fiedilty). Read [the report](https://github.com/TencentARC/PhotoMaker/blob/main/README_pmv2.md) for details.

Update replicate code and readme

Hi @yorickvP, there are a lot of changes. We will check it carefully.

Update replicate code and readme

Hi, @yorickvP. Thanks for your efforts. We are happy to accept such an excellent demo from you. But there is a question, how should we do it?

Update replicate code and readme

Thank you @yorickvP!. We've merged your PR.

Update replicate code and readme

Thanks for your awesome work!

建议搞个百度网盘

You can also choose to download manually from this [url](https://huggingface.co/TencentARC/PhotoMaker). And then replace the following lines: ```python from huggingface_hub import hf_hub_download photomaker_path = hf_hub_download(repo_id="TencentARC/PhotoMaker", filename="photomaker-v1.bin", repo_type="model") ``` to ```python photomaker_path...

建议搞个百度网盘

链接：https://pan.baidu.com/s/10sVIr1MnKiw4Et91fL6AJA?pwd=ptmk 提取码：ptmk

Who can tell me the difference between this and faceChain, whose effect is better?

Rapid customization within seconds, with no additional LoRA training.