In this repo, the face patch preprocessing step is same for all the models.

img = cv2.resize(img, target_size)
img_pixels = image.img_to_array(img)
img_pixels = np.expand_dims(img_pixels, axis = 0)
img_pixels /= 255 #normalize input in [0, 1]

Please look into it if this is valid for all, as the inference preprocessing step must match the model's training pipeline for accurate result. For example, the original Facenet repo has done it differently.

Hey i have nvidia 1070 GPU can you guide me to run this on gpu i Tried Deepface.allocateMemory() but it saying it will run on CPU. let me know the necessary tools as well as settings required to run this on GPU.

Thanks

When using the deepface, I use conda the create a virsual envrionment, and using pip install deepface to install the packdge, but when I run the code, I met the error: AttributeError: module 'keras.utils.generic_utils' has no attribute 'populate_dict_with_module_objects' deepface==0.5.1 tensorflow==2.5.0 keras==2.4.3

Hey there,

I just started with DeepFace and it seemed all to work. But something weirds happen when I put the exact same code (copied from the official page) in a function.

This is the code I try to run:

` try:

            faceinfo = DeepFace.analyze("images/happy.jpeg")
            print('[INFO] Emotion:', faceinfo["dominant_emotion"])
            print('[INFO] Age:', faceinfo["age"])
            print('[INFO] Gender:', faceinfo["gender"])
            print('[INFO] Race:', faceinfo["dominant_race"])


        except Exception as e:
            print(str(e) + ' [FACE INFO]')`

The error i'm getting: '_thread._local' object has no attribute 'value'

Could anyone help me out with this?

I'm using same code as you did in Elasticsearch. Im also using euclidean distance in query. I want to get score in % so I multiplied candidate_score * 100 but still results are always coming between 6% to 12%. can you tell me the process to get candidate score between 0 to 100%..

Thanks

api.py not support tensorflow version 2.2.0+, when i run it, report error function get_default_graph not found, and it's can't load some models, report ValueError: rate must be a scalar tensor or a float in the range [0, 1)

I have this error using debian on the docker, I would like to disable cuda support, is there how?

Below is the output:

Using TensorFlow backend. ^[[AUsing VGG-Face model backend and cosine distance. 2020-05-20 14:35:39.426369: E tensorflow/stream_executor/cuda/cuda_driver.cc:313] failed call to cuInit: UNKNOWN ERROR (303) Verification: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 1.46it/s] {'pair_1': {'distance': 0.23476946353912354, 'max_threshold_to_verify': 0.4, 'model': 'VGG-Face', 'similarity_metric': 'cosine', 'verified': True}}

Dear Mr.Sefik! I like the combination between facenet and mtcnn for face recognition, not analyze But I have 2 troubles: 1/ When I execute DeepFace.find, this function generates a file in folder (representations_facenet.pkl), I want to append for new value, new image, I failed because I don't know format statement in file
df = pd.DataFrame(representations, columns = ["identity", "%s_representation" % (model_name)]). How to do that? I tried to write file [a+] or [wb] and that thing, it doesn't work with next time detection | recognition 2/ the reason I do the number 1 thing is - Loading speed for 1 vector file takes a lot of seconds each time ~43 seconds of Facenet model, I agree that the second time is very fast, only 8 seconds [time import, time find function], but the recognition is not correct. Finally, I have to delete the old pkl file and do again. There is no way to improve adding new to the pkl . file. I want to add or combine file, and speed more faster, about 3-5 seconds.

 Thats all | Help me, thanks

I have installed tensorflow-gpu instead of tensortflow after doing pip install deepface. How can I run deepface on a gpu? or be sure that deepface is running on a GPU? DeepFace.verify with 'VGG-Face' takes close to 6 seconds. I would like to reduce this by using a GPU.

I am doing experiment on large scale datasets with extensive memory and GPU. Actually, when i started an experiments on Ensemble-Face-Recognition my process has been killed. Is it possible to consume my GPU for these experiments.

(236167, 3)
Process finished with exit code 137 (interrupted by signal 9: SIGKILL)

When I want to use Ensemble model, like the tutorial"

resp_obj = DeepFace.verify("img1.jpg", "img2.jpg", model_name = "Ensemble")

I get the following error: There is an error with this image and modelCould not open /usr/local/lib/python3.6/dist-packages/deepface/models/face-recognition-ensemble-model.txt

How we can solve that?

the library version is 0.0.31

Description

This PR adds the ability to use png, webp and jpg|jpeg images in local image database.

The commons module now includes an utils.py file that contains:

• A list of default supported image extensions
• Helper to extract the file extension from a given filename
• Helper to check if a given filename ends with any of the supported file extensions
• Helper to generate a label for a canidate from a given path

Fixes #612

Type of change

• [x] New feature (non-breaking change which adds functionality)
• [x] This change requires a documentation update

Current behavior

• Only allows images in jpg format

New behavior

• In addition to the current behavior, also allows using png and webp images
• Supports both JPEG extensions - jpeg | jpg

Tests

• [x] Simple tests included in test/utils.py
• [x] Run tests/unit_test.py

# in colab
from deepface.commons import functions
import numpy as np
url = "https://upload.wikimedia.org/wikipedia/commons/thumb/4/45/MontreGousset001.jpg/440px-MontreGousset001.jpg"
file = "440px-MontreGousset001.jpg"
!wget {url} -O {file} -q

img_from_url = functions.load_image(url)
img_from_file = functions.load_image(file)

print(np.abs(img_from_url-img_from_file).mean())
# 98.4184350027692

https://github.com/serengil/deepface/blob/ba9f56755abdc9d1d80777d7680c4267d5d5e0e9/deepface/commons/functions.py#L85-L92 The reason is because you convert the image to RBG with url And for file, you use cv2.imread, the default setting is rgb

Please align those two methods

Hi! I wanted to detect multiple faces in an image but the library didn't offer any support for multiple face detection. Detecting multiple faces in the images might be a use case for other developers too!

I've run unit_test.py and attached the output : unit_tests.txt

PS: I am not a pro. Please suggest any design changes if it can be done in a better way!