"KeyError: 'embedding'" in "InstantID Pipeline"

[ScottishFold007]

Describe the bug

I use the code of "InstantID Pipeline":

import diffusers
from diffusers.utils import load_image
from diffusers.models import ControlNetModel

import cv2
import torch
import numpy as np
from PIL import Image

from insightface.app import FaceAnalysis
from pipeline_stable_diffusion_xl_instantid import StableDiffusionXLInstantIDPipeline, draw_kps

# prepare 'antelopev2' under ./models
# https://github.com/deepinsight/insightface/issues/1896#issuecomment-1023867304
app = FaceAnalysis(name='antelopev2', root='./', providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])
app.prepare(ctx_id=0, det_size=(640, 640))

# prepare models under ./checkpoints
# https://huggingface.co/InstantX/InstantID
from huggingface_hub import hf_hub_download
hf_hub_download(repo_id="InstantX/InstantID", filename="ControlNetModel/config.json", local_dir="./checkpoints")
hf_hub_download(repo_id="InstantX/InstantID", filename="ControlNetModel/diffusion_pytorch_model.safetensors", local_dir="./checkpoints")
hf_hub_download(repo_id="InstantX/InstantID", filename="ip-adapter.bin", local_dir="./checkpoints")

face_adapter = f'./checkpoints/ip-adapter.bin'
controlnet_path = f'./checkpoints/ControlNetModel'

# load IdentityNet
controlnet = ControlNetModel.from_pretrained(controlnet_path, torch_dtype=torch.float16)

base_model = 'wangqixun/YamerMIX_v8'
pipe = StableDiffusionXLInstantIDPipeline.from_pretrained(
    base_model,
    controlnet=controlnet,
    torch_dtype=torch.float16
)
pipe.cuda()

# load adapter
pipe.load_ip_adapter_instantid(face_adapter)

# load an image
face_image = load_image("https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/ai_face2.png")

# prepare face emb
face_info = app.get(cv2.cvtColor(np.array(face_image), cv2.COLOR_RGB2BGR))
face_info = sorted(face_info, key=lambda x:(x['bbox'][2]-x['bbox'][0])*x['bbox'][3]-x['bbox'][1])[-1]  # only use the maximum face
face_emb = face_info['embedding']
face_kps = draw_kps(face_image, face_info['kps'])

# prompt
prompt = "film noir style, ink sketch|vector, male man, highly detailed, sharp focus, ultra sharpness, monochrome, high contrast, dramatic shadows, 1940s style, mysterious, cinematic"
negative_prompt = "ugly, deformed, noisy, blurry, low contrast, realism, photorealistic, vibrant, colorful"

# generate image
pipe.set_ip_adapter_scale(0.8)
image = pipe(
    prompt,
    image_embeds=face_emb,
    image=face_kps,
    controlnet_conditioning_scale=0.8,
).images[0]```




### Reproduction

I use the code of "InstantID Pipeline":
```# !pip install opencv-python transformers accelerate insightface
import diffusers
from diffusers.utils import load_image
from diffusers.models import ControlNetModel

import cv2
import torch
import numpy as np
from PIL import Image

from insightface.app import FaceAnalysis
from pipeline_stable_diffusion_xl_instantid import StableDiffusionXLInstantIDPipeline, draw_kps

# prepare 'antelopev2' under ./models
# https://github.com/deepinsight/insightface/issues/1896#issuecomment-1023867304
app = FaceAnalysis(name='antelopev2', root='./', providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])
app.prepare(ctx_id=0, det_size=(640, 640))

# prepare models under ./checkpoints
# https://huggingface.co/InstantX/InstantID
from huggingface_hub import hf_hub_download
hf_hub_download(repo_id="InstantX/InstantID", filename="ControlNetModel/config.json", local_dir="./checkpoints")
hf_hub_download(repo_id="InstantX/InstantID", filename="ControlNetModel/diffusion_pytorch_model.safetensors", local_dir="./checkpoints")
hf_hub_download(repo_id="InstantX/InstantID", filename="ip-adapter.bin", local_dir="./checkpoints")

face_adapter = f'./checkpoints/ip-adapter.bin'
controlnet_path = f'./checkpoints/ControlNetModel'

# load IdentityNet
controlnet = ControlNetModel.from_pretrained(controlnet_path, torch_dtype=torch.float16)

base_model = 'wangqixun/YamerMIX_v8'
pipe = StableDiffusionXLInstantIDPipeline.from_pretrained(
    base_model,
    controlnet=controlnet,
    torch_dtype=torch.float16
)
pipe.cuda()

# load adapter
pipe.load_ip_adapter_instantid(face_adapter)

# load an image
face_image = load_image("https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/ai_face2.png")

# prepare face emb
face_info = app.get(cv2.cvtColor(np.array(face_image), cv2.COLOR_RGB2BGR))
face_info = sorted(face_info, key=lambda x:(x['bbox'][2]-x['bbox'][0])*x['bbox'][3]-x['bbox'][1])[-1]  # only use the maximum face
face_emb = face_info['embedding']
face_kps = draw_kps(face_image, face_info['kps'])

# prompt
prompt = "film noir style, ink sketch|vector, male man, highly detailed, sharp focus, ultra sharpness, monochrome, high contrast, dramatic shadows, 1940s style, mysterious, cinematic"
negative_prompt = "ugly, deformed, noisy, blurry, low contrast, realism, photorealistic, vibrant, colorful"

# generate image
pipe.set_ip_adapter_scale(0.8)
image = pipe(
    prompt,
    image_embeds=face_emb,
    image=face_kps,
    controlnet_conditioning_scale=0.8,
).images[0]```



### Logs

```shell
KeyError                                  Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_36812\3787166824.py in <module>
      5 face_info = app.get(cv2.cvtColor(np.array(face_image), cv2.COLOR_RGB2BGR))
      6 face_info = sorted(face_info, key=lambda x:(x['bbox'][2]-x['bbox'][0])*x['bbox'][3]-x['bbox'][1])[-1]  # only use the maximum face
----> 7 face_emb = face_info['embedding']
      8 face_kps = draw_kps(face_image, face_info['kps'])
      9 

KeyError: 'embedding'

System Info

diffusers==0.26.3
insightface==0.7.3

Who can help?

No response

[tolgacangoz]

Hi @ScottishFold007!
I found this discussion at InstantID's repo. Did/Could you look at that?

[haofanwang]

What's your image?

[yiyixuxu]

cc @fabiorigano here too

[yiyixuxu]

is this an issue with diffusers?

[fabiorigano]

hi, thanks Yiyi for adding me here

I made a colab to test the feature extraction, but it works for me and I am not able to reproduce the issue
https://colab.research.google.com/drive/1ccuqhRGLSkWcr8itQFmpXzi0HAH-Yi_u?usp=sharing

[sayakpaul]

@ScottishFold007 feel free to test the Colab @fabiorigano shared and let us know!

[ScottishFold007]

@ScottishFold007 feel free to test the Colab @fabiorigano shared and let us know!

I found that the problem was that 2 files were missing from the facenet model file, and now it works fine!

[Dominux]

@ScottishFold007 feel free to test the Colab @fabiorigano shared and let us know!

I found that the problem was that 2 files were missing from the facenet model file, and now it works fine!

what files?