aigc-apps / sd-webui-EasyPhoto

📷 EasyPhoto | Your Smart AI Photo Generator.
Apache License 2.0
4.98k stars 399 forks source link

[Bug]: Image with two faces. Template 1 error: Error info is index 0 is out of bounds for axis 0 with size 0, skip it. #431

Open yankai-victor opened 4 months ago

yankai-victor commented 4 months ago

Is there an existing issue for this?

Is EasyPhoto the latest version?

What happened?

u=1994847390,4108431023 fm=253 fmt=auto app=138 f=JPEG I use this image to change face. If I set the Num of Faceid =1, it's fine. But if i set the Num of Faceid =2, there is always wrong.

Steps to reproduce the problem

  1. upload image
  2. set num of faceid, user_0 id and user_1 id
  3. Start Generation Single Image Upload

What should have happened?

The two number of face in the image will be changed

Commit where the problem happens

webui: EastPhoto: Template 1 error: Error info is index 0 is out of bounds for axis 0 with size 0, skip it.

What browsers do you use to access the UI ?

Google Chrome

Command Line Arguments

no

List of enabled extensions

扩展插件/Extension 网址/URL 分支/Branch 版本/Version 日期/Date 更新/Update
adetailer https://github.com/Bing-su/adetailer.git main a7d96113 2024-06-16 08:32:49 latest
sd-webui-EasyPhoto https://github.com/aigc-apps/sd-webui-EasyPhoto main c9a8c799 2024-06-06 09:15:18 latest
sd-webui-controlnet https://github.com/Mikubill/sd-webui-controlnet.git main ee96dc92 2024-07-04 01:31:50 latest
stable-diffusion-webui-chinese https://github.com/VinsonLaro/stable-diffusion-webui-chinese main e240eccb 2024-05-14 21:42:27 latest

Console logs

Loading weights [59ffe2243a] from D:\Project\SD1\models\Stable-diffusion\Chilloutmix-Ni-pruned-fp16-fix.safetensors
2024-07-05 19:08:21,272 - EasyPhoto - ControlNet unit number: 4
2024-07-05 19:08:21,272 - EasyPhoto - Found 2 user id(s), but only 0 image prompt(s) for IP-Adapter Control. Use the reference image corresponding to the user instead.
2024-07-05 19:08:21,272 - EasyPhoto - Display score is forced to be true when IP-Adapter Control is enabled.
2024-07-05 19:08:21,565 - EasyPhoto - Start templates and user_ids preprocess.
Reusing loaded model realisticVisionV51_v51VAE.safetensors [ef76aa2332] to load Chilloutmix-Ni-pruned-fp16-fix.safetensors [59ffe2243a]
Loading weights [59ffe2243a] from D:\Project\SD1\models\Stable-diffusion\Chilloutmix-Ni-pruned-fp16-fix.safetensors
Applying attention optimization: Doggettx... done.
Weights loaded in 1.4s (send model to cpu: 0.6s, apply weights to model: 0.3s, move model to device: 0.4s).
Loading VAE weights specified in settings: D:\Project\SD1\models\VAE\vae-ft-mse-840000-ema-pruned.ckpt
Applying attention optimization: Doggettx... done.
VAE weights loaded.
2024-07-05 19:08:24,531 - EasyPhoto -
            Start Generate template                 : 1;
            user_ids                                : ('jay', 'jay', 'none', 'none', 'none');
            sd_model_checkpoint                     : Chilloutmix-Ni-pruned-fp16-fix.safetensors;
            input_prompts                           : ['easyphoto_face, easyphoto, 1person, <lora:jay:0.9>, <lora:FilmVelvia3:0.65>, masterpiece, beauty', 'easyphoto_face, easyphoto, 1person, <lora:jay:0.9>, <lora:FilmVelvia3:0.65>, masterpiece, beauty', 'none', 'none', 'none'];
            before_face_fusion_ratio                : 0.6;
            after_face_fusion_ratio                 : 0.6;
            first_diffusion_steps                   : 50;
            first_denoising_strength                : 0.45;
            second_diffusion_steps                  : 20;
            second_denoising_strength               : 0.3;
            seed                                    : 30486
            crop_face_preprocess                    : True
            apply_face_fusion_before                : True
            apply_face_fusion_after                 : True
            color_shift_middle                      : True
            color_shift_last                        : True
            super_resolution                        : True
            super_resolution_method                 : gpen
            display_score                           : True
            background_restore                      : False
            background_restore_denoising_strength   : 0.35
            makeup_transfer                         : False
            makeup_transfer_ratio                   : 0.5
            skin_retouching_bool                    : False
            face_shape_match                        : False
            id_control                              : True
            id_control_method                       : IP-Adapter Face
            ipa_weight                              : 0.3
            instantid_id_weight                     : 0.5
            instantid_ipa_weight                    : 0.5
            instantid_image_path                    : None
            ipa_image_path                          : None
            ref_mode_choose                         : Infer with User Lora
            no_user_lora_mode                       : IP-Adapter Face
            ipa_only_weight                         : 0.6
            ipa_only_image_path                     : None
            instantid_only_id_weight                : 0.5
            instantid_only_ipa_weight               : 0.5
            instantid_only_image_path               : None

2024-07-05 19:08:24,839 - EasyPhoto - Start Image resize to 512.0.
2024-07-05 19:08:24,844 - EasyPhoto - Start face detect.
D:\Project\SD1\venv\lib\site-packages\modelscope\models\cv\image_face_fusion\facelib\matlab_cp2tform.py:81: FutureWarning: `rcond` parameter will change to the default of machine precision times ``max(M, N)`` where M and N are the input matrix dimensions.
To use the future default and silence this warning we advise to pass `rcond=None`, to keep using the old, explicitly pass `rcond=-1`.
  r, _, _, _ = lstsq(X, U)
2024-07-05 19:08:27,005 - EasyPhoto - Start First diffusion.
100%|██████████████████████████████████████████████████████████████████████████████████| 10/10 [00:02<00:00,  3.97it/s]
2024-07-05 19:08:30,031 - EasyPhoto - Start color shift middle.
2024-07-05 19:08:30,236 - EasyPhoto - Start second face fusion.
2024-07-05 19:08:31,358 - EasyPhoto - Start mouth detect.
2024-07-05 19:08:31,502 - EasyPhoto - Start Second diffusion.
100%|████████████████████████████████████████████████████████████████████████████████████| 7/7 [00:01<00:00,  6.28it/s]
2024-07-05 19:08:33,093 - EasyPhoto - Start color shift last.
2024-07-05 19:08:33,200 - EasyPhoto - Start paste crop image to origin template.
C:\Users\Cinta\.cache\modelscope\modelscope_modules\cv_retinafce_recognition\image_face_recognition\matlab_cp2tform.py:81: FutureWarning: `rcond` parameter will change to the default of machine precision times ``max(M, N)`` where M and N are the input matrix dimensions.
To use the future default and silence this warning we advise to pass `rcond=None`, to keep using the old, explicitly pass `rcond=-1`.
  r, _, _, _ = lstsq(X, U)
2024-07-05 19:08:33,875 - EasyPhoto - Start paste crop image to origin template in multi people.
2024-07-05 19:08:33,975 - EasyPhoto - Start Image resize to 512.0.
2024-07-05 19:08:33,980 - EasyPhoto - Start face detect.
Traceback (most recent call last):
  File "D:\Project\SD1\extensions\sd-webui-EasyPhoto\scripts\easyphoto_infer.py", line 1338, in easyphoto_infer_forward
    input_image_retinaface_box = input_image_retinaface_boxes[0]
IndexError: index 0 is out of bounds for axis 0 with size 0
2024-07-05 19:08:34,149 - EasyPhoto - Template 1 error: Error info is index 0 is out of bounds for axis 0 with size 0, skip it.
Cleanup completed.
Reusing loaded model Chilloutmix-Ni-pruned-fp16-fix.safetensors [59ffe2243a] to load realisticVisionV51_v51VAE.safetensors [ef76aa2332]
Loading weights [ef76aa2332] from D:\Project\SD1\models\Stable-diffusion\realisticVisionV51_v51VAE.safetensors
Loading VAE weights specified in settings: D:\Project\SD1\models\VAE\vae-ft-mse-840000-ema-pruned.ckpt
Applying attention optimization: Doggettx... done.
Weights loaded in 1.7s (send model to cpu: 0.6s, apply weights to model: 0.5s, load VAE: 0.2s, move model to device: 0.5s).
Restoring base VAE
Applying attention optimization: Doggettx... done.
VAE weights loaded.

Additional information

No response