add model

Dockerfile

@@ -0,0 +1,18 @@
+FROM python:3.8.13-bullseye
+
+LABEL author="Bastien GUILLAUME"
+LABEL version="0.0.2"
+
+
+RUN apt-get update \
+    && apt-get upgrade -y \
+    && apt-get install ffmpeg libsm6 libxext6  -y
+
+COPY requirements.txt /opt/app/
+WORKDIR /opt/app
+RUN pip install -r requirements.txt
+
+COPY inference.py /opt/app/
+
+CMD ["python", "inference.py"]
+EXPOSE 80

app.py

@@ -0,0 +1,78 @@
+# pylint: disable=no-name-in-module
+# pylint: disable=no-member
+
+"""
+Author : Bastien GUILLAUME
+Version : 0.0.1
+Date : 2023-03-16
+
+Title : Inference With Gradio running an onnxruntime backend
+"""
+
+import os
+
+import gradio as gr
+from config_parser import *
+from inferencer import *
+
+gr.close_all()
+
+
+def infertace_builder():
+    interface_list = []
+    logging.log(level=logging.INFO, msg=f"Building Interfaces")
+    for task_number in range(0, len(tasks)):
+        logging.log(level=logging.INFO, msg=f"Building Interface n° {task_number}")
+        interface_list.append(
+            gr.Interface(
+                title=tasks[task_number],
+                description=tasks_description[task_number],
+                fn=analysis,
+                # if tasks_name[task_number] == "Étiquetage"
+                # else corck_screwing_analysis,
+                allow_flagging="never",
+                inputs=[
+                    gr.Textbox(
+                        value=f"task{task_number+1}", visible=False, interactive=False
+                    ),
+                    gr.Dropdown(
+                        tasks_products[task_number],
+                        type="value",
+                        value=tasks_products[task_number][0],
+                        label="Choix",
+                        visible=True if len(tasks_products[task_number]) > 1 else False,
+                        info="Sur quel type de produit, voulez vous lancer l'analyse ?",
+                    ),
+                    gr.Image(
+                        label="Image à analyser",
+                        shape=None,
+                        image_mode="RGB",
+                        invert_colors=False,
+                        source="upload",
+                        tool="editor",
+                        type="numpy",
+                    ),
+                ],
+                outputs=gr.Label(
+                    label="Résultats",
+                ),
+            )
+        )
+    logging.log(level=logging.INFO, msg=f"Interfaces ready\n")
+    return interface_list
+
+
+generated_interface = infertace_builder()
+
+iface_generated = gr.TabbedInterface(
+    interface_list=generated_interface,
+    tab_names=tasks_name,
+    title=title,
+)
+
+if __name__ == "__main__":
+    logging.log(level=logging.INFO, msg="Starting the Gradio server...")
+    iface_generated.launch(
+        server_name="0.0.0.0", server_port=int(os.getenv("PORT", "8150"))
+    )
+    logging.log(level=logging.INFO, msg="Stopping the Gradio server...")

build_image.sh

@@ -0,0 +1,8 @@
+#!/bin/bash -e
+image_name=gcr.io/tough-variety-310920/openvino_inference
+image_tag=1.0.1
+full_image_name=${image_name}:${image_tag}
+
+cd "$(dirname "$0")" 
+docker build -t "${full_image_name}" .
+docker push "$full_image_name"

config_parser.py

@@ -0,0 +1,86 @@
+# pylint: disable=no-name-in-module
+# pylint: disable=no-member
+
+"""
+Author : Bastien GUILLAUME
+Version : 0.0.1
+Date : 2023-03-16
+
+Title : Inference With Gradio running an onnxruntime backend
+"""
+
+import json
+import logging
+import os
+from functools import reduce
+
+logging.basicConfig(
+    format="%(asctime)s.%(msecs)03d %(levelname)-8s %(message)s",
+    level=logging.INFO,
+    datefmt="%Y%m%d-%H%M%S",
+)
+
+with open("demo_yvesrocher.json") as config_file:
+    config = json.load(config_file)
+
+logging.log(level=logging.DEBUG, msg=f"Loaded config file : {json.dumps(config)}")
+
+
+def deep_get(dictionary, keys, default=None):
+    return reduce(
+        lambda d, key: d.get(key, default) if isinstance(d, dict) else default,
+        keys.split("."),
+        dictionary,
+    )
+
+
+def get_tasks():
+    tasks = []
+    for task in config["tasks"].keys():
+        tasks.append(config["tasks"][task]["name"]["fr"])
+    return tasks
+
+
+def get_tasks_name():
+    tasks_name = []
+    for task in config["tasks"].keys():
+        tasks_name.append(config["tasks"][task]["shortname"])
+    return tasks_name
+
+
+def get_tasks_description():
+    tasks_descripion = []
+    for task in config["tasks"].keys():
+        tasks_descripion.append(config["tasks"][task]["description"]["fr"])
+    return tasks_descripion
+
+
+def get_tasks_products():
+    tasks_products = []
+    for task in config["tasks"].keys():
+        tasks_products.append(config["tasks"][task]["products"])
+    return tasks_products
+
+
+title = (
+    os.getenv("GRADIO_TITLE")
+    if "GRADIO_TITLE" in os.environ
+    else config.get("title", "TITLE neither set in config file nor in ENV")
+)
+description = (
+    os.getenv("GRADIO_DESCRIPTION")
+    if "GRADIO_DESCRIPTION" in os.environ
+    else config.get("description", "DESCRIPTION neither set in config file nor in ENV")
+)
+tasks = get_tasks()
+tasks_name = get_tasks_name()
+tasks_description = get_tasks_description()
+tasks_products = get_tasks_products()
+logging.log(level=logging.INFO, msg=f"Parsed Data :")
+logging.log(level=logging.INFO, msg=f"Title : {title}")
+logging.log(level=logging.INFO, msg=f"Description : {description}")
+logging.log(level=logging.INFO, msg=f"Tasks : {tasks}")
+logging.log(level=logging.INFO, msg=f"Tasks name : {tasks_name}")
+logging.log(level=logging.INFO, msg=f"Tasks descrption : {tasks_description}")
+logging.log(level=logging.INFO, msg=f"Tasks products : {tasks_products}")
+logging.log(level=logging.INFO, msg=f"End of Parsed Data\n")

configtest.json

@@ -0,0 +1,100 @@
+{
+    "title": "YVES ROCHER",
+    "description": "Démonstration des algos de reconnaissance d'étiquetage/bouchage correct",
+    "tasks": {
+        "task1": {
+            "shortname": "Étiquetage",
+            "name": {
+                "en": "Quality Control of Labels",
+                "fr": "Contrôle de l'Étiquetage"
+            },
+            "description": {
+                "en": "",
+                "fr": "Est-ce que l'étiquette est bien positionnée"
+            },
+            "products":[
+                "497 Pure Algue 200ml",
+                "505 Pure Calmille 200ml",
+                "614 Eco Douche 100ml",
+                "648 Hair Care 300ml"
+            ],
+            "models": {
+                "497 Pure Algue 200ml": {
+                    "type": "classification",
+                    "path":  "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/po2xfZzK0KtWZUpXrFjZ/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage incorrect", "Étiquetage correct"]
+                },
+                "505 Pure Calmille 200ml": {
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/po2xfZzK0KtWZUpXrFjZ/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage incorrect", "Étiquetage correct"]
+                },
+                "614 Eco Douche 100ml": {
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/po2xfZzK0KtWZUpXrFjZ/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage incorrect", "Étiquetage correct"]
+                },
+                "648 Hair Care 300ml": {
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/po2xfZzK0KtWZUpXrFjZ/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage incorrect", "Étiquetage correct"]
+                }
+            }
+        },
+        "task2": {
+            "shortname": "Bouchage",
+            "name": {
+                "en": "Quality Control of Corck Screwing",
+                "fr": "Contrôle du Bouchage"
+            },
+            "description": {
+                "en": "",
+                "fr": "Est-ce que le bouchon est bien positionné et entièrement vissé"
+            },
+            "products":[
+                "187 Hamamelis 300ml",
+                "550 SVC 300ml",
+                "600 PN 500 ml"
+            ],
+            "models": {
+                "187 Hamamelis 300ml":{
+                    "type": "classification",
+                    "path": "/Users/bastien/Downloads/model_corck_screwing.onnx",
+                    "input_shape": 256,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Bouchage incorrect", "Bouchage correct"]
+                },
+                "550 SVC 300ml": {
+                    "type": "classification",
+                    "path": "/Users/bastien/Downloads/model_corck_screwing.onnx",
+                    "input_shape": 256,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Bouchage incorrect", "Bouchage correct"]
+                },
+                "600 PN 500 ml": {
+                    "type": "classification",
+                    "path": "/Users/bastien/Downloads/model_corck_screwing.onnx",
+                    "input_shape": 256,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Bouchage incorrect", "Bouchage correct"]
+                }
+            }
+        }
+    }
+}
+

demo_full_yvesrocher.json

@@ -0,0 +1,100 @@
+{
+    "title": "YVES ROCHER",
+    "description": "Démonstration des algos de reconnaissance d'étiquetage/bouchage correct",
+    "tasks": {
+        "task1": {
+            "shortname": "Étiquetage",
+            "name": {
+                "en": "Quality Control of Labels",
+                "fr": "Contrôle de l'Étiquetage"
+            },
+            "description": {
+                "en": "Is the label in the right position ?",
+                "fr": "Est-ce que l'étiquette est bien positionnée ?"
+            },
+            "products":[
+                "497 Pure Algue 200ml",
+                "505 Pure Calmille 200ml",
+                "614 Eco Douche 100ml",
+                "648 Hair Care 300ml"
+            ],
+            "models": {
+                "497 Pure Algue 200ml": {
+                    "type": "classification",
+                    "path":  "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/po2xfZzK0KtWZUpXrFjZ/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage correct", "Étiquetage incorrect"]
+                },
+                "505 Pure Calmille 200ml": {
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/po2xfZzK0KtWZUpXrFjZ/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage correct", "Étiquetage incorrect"]
+                },
+                "614 Eco Douche 100ml": {
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/po2xfZzK0KtWZUpXrFjZ/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage correct", "Étiquetage incorrect"]
+                },
+                "648 Hair Care 300ml": {
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/po2xfZzK0KtWZUpXrFjZ/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage correct", "Étiquetage incorrect"]
+                }
+            }
+        },
+        "task2": {
+            "shortname": "Bouchage",
+            "name": {
+                "en": "Quality Control of Corck Screwing",
+                "fr": "Contrôle du Bouchage"
+            },
+            "description": {
+                "en": "Is the corck in the right position ?",
+                "fr": "Est-ce que le bouchon est bien positionné et entièrement vissé ?"
+            },
+            "products":[
+                "187 Hamamelis 300ml",
+                "550 SVC 300ml",
+                "600 PN 500 ml"
+            ],
+            "models": {
+                "187 Hamamelis 300ml":{
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/Qp6BRHBcLq7KKxqCWqmV/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Bouchage correct", "Bouchage incorrect"]
+                },
+                "550 SVC 300ml": {
+                    "type": "classification",
+                    "path": "/Users/bastien/Downloads/model_corck_screwing.onnx",
+                    "input_shape": 256,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Bouchage correct", "Bouchage incorrect"]
+                },
+                "600 PN 500 ml": {
+                    "type": "classification",
+                    "path": "/Users/bastien/Downloads/model_corck_screwing.onnx",
+                    "input_shape": 256,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Bouchage correct", "Bouchage incorrect"]
+                }
+            }
+        }
+    }
+}
+

demo_yvesrocher.json

@@ -0,0 +1,55 @@
+{
+    "title": "YVES ROCHER",
+    "description": "Démonstration des algos de reconnaissance d'étiquetage/bouchage correct",
+    "tasks": {
+        "task1": {
+            "shortname": "Étiquetage",
+            "name": {
+                "en": "Quality Control of Labels",
+                "fr": "Contrôle de l'Étiquetage"
+            },
+            "description": {
+                "en": "Is the label in the right position ?",
+                "fr": "Est-ce que l'étiquette est bien positionnée ?"
+            },
+            "products":[
+                "505 Pure Calmille 200ml"
+            ],
+            "models": {
+                "505 Pure Calmille 200ml": {
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/dmeQ6Mae0HKDwkHBTdh2/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Étiquetage correct", "Étiquetage incorrect"]
+                }
+            }
+        },
+        "task2": {
+            "shortname": "Bouchage",
+            "name": {
+                "en": "Quality Control of Corck Screwing",
+                "fr": "Contrôle du Bouchage"
+            },
+            "description": {
+                "en": "Is the corck in the right position ?",
+                "fr": "Est-ce que le bouchon est bien positionné et entièrement vissé ?"
+            },
+            "products":[
+                "187 Hamamelis 300ml"
+            ],
+            "models": {
+                "187 Hamamelis 300ml":{
+                    "type": "classification",
+                    "path": "https://automi-test-models.s3.eu-west-3.amazonaws.com/inference-pipeline/3nkoTiBxDNBzqyGtE5w8/model.onnx",
+                    "input_shape": 224,
+                    "mean": [0.485, 0.456, 0.406],
+                    "std": [0.229, 0.224, 0.225],
+                    "class_names": ["Bouchage correct", "Bouchage incorrect"]
+                }
+            }
+        }
+    }
+}
+

docker-compose.yaml

@@ -0,0 +1,13 @@
+version: '3.8'
+
+services:
+  magnetoscopie:
+    image: openvino_inference:1.0.1
+    container_name: gradio_showcase_onnx_model
+    volumes:
+      - ./models:/opt/app/models
+    ports:
+      - "8150:80"
+    env_file:
+      - .env
+    restart: unless-stopped

inferencer.py

@@ -0,0 +1,136 @@
+# pylint: disable=no-name-in-module
+# pylint: disable=no-member
+
+"""
+Author : Bastien GUILLAUME
+Version : 0.0.1
+Date : 2023-03-16
+
+Title : Inference With Gradio running an onnxruntime backend
+"""
+
+import numpy as np
+import onnxruntime as ort
+import requests
+from config_parser import *
+from torchvision import transforms
+
+corck_screwing_metadata = {
+    "image_threshold": 0.9247307181358337,
+    "pixel_threshold": 0.9247307181358337,
+    "min": 4.034907666260021e-26,
+    "max": 0.998478353023529,
+}
+
+inferencer_arr = {}
+logging.log(level=logging.INFO, msg="Loading models...")
+for task in config["tasks"].keys():
+    inferencer_arr[task] = {}
+    r = None
+    for model in config["tasks"][task]["models"]:
+        product_name = model
+        model_path = config["tasks"][task]["models"][product_name]["path"]
+        logging.log(
+            level=logging.INFO,
+            msg=f"Loading labeling model for product {product_name}",
+        )
+        inferencer_arr[task][product_name] = {}
+        if model_path.startswith("http"):
+            r = requests.get(model_path, stream=True).content
+        inferencer_arr[task][product_name][product_name] = ort.InferenceSession(
+            r if model_path.startswith("http") else model_path
+        )
+        inferencer_arr[task][product_name]["input_name"] = (
+            inferencer_arr[task][product_name][product_name].get_inputs()[0].name
+        )
+        inferencer_arr[task][product_name]["output_name"] = (
+            inferencer_arr[task][product_name][product_name].get_outputs()[0].name
+        )
+        logging.log(level=logging.INFO, msg=f"Model for {product_name} loaded\n")
+logging.log(level=logging.INFO, msg="All models loaded...\n\n")
+logging.log(level=logging.DEBUG, msg=f"Inferencer Array : {inferencer_arr}")
+
+
+def softmax(x):
+    e_x = np.exp(x - np.max(x))
+    return e_x / e_x.sum(axis=0)
+
+
+def is_anomalous_classification(prediction, meta_data):
+    pred_label = None
+    pred_score = prediction.reshape(-1).max()
+    if "image_threshold" in meta_data:
+        pred_label = (
+            config["tasks"][task]["models"][model]["class_names"][0]
+            if (pred_score >= meta_data["image_threshold"])
+            else config["tasks"][task]["models"][model]["class_names"][1]
+        )
+    return pred_label, pred_score
+
+
+def analysis(task, use_case, image):
+    """
+    Main function that process inference and return results strings
+    Args:
+        - task
+        - use case
+        - image
+    Returns:
+        - String including label and confidence of the model
+    """
+
+    input_image = pre_process_all(task, use_case, image)
+    result = inference(task, use_case, input_image)
+    logging.log(level=logging.DEBUG, msg=result)
+    return result
+
+
+def pre_process_all(task, use_case, image):
+    preprocessed_image = []
+    input_shape = config["tasks"][task]["models"][use_case]["input_shape"]
+    mean = config["tasks"][task]["models"][use_case]["mean"]
+    std = config["tasks"][task]["models"][use_case]["std"]
+    logging.log(level=logging.DEBUG, msg=f"Shape {input_shape}")
+    logging.log(level=logging.DEBUG, msg=f"Mean {mean}")
+    logging.log(level=logging.DEBUG, msg=f"Std {std}")
+    data_transforms = transforms.Compose(
+        [
+            transforms.ToPILImage(),
+            transforms.Resize(input_shape),
+            transforms.CenterCrop(input_shape),
+            transforms.ToTensor(),
+            transforms.Normalize(mean, std),
+        ]
+    )
+    preprocessed_image = data_transforms(image).detach().numpy()
+    preprocessed_image = np.expand_dims(preprocessed_image, axis=0)
+    logging.log(level=logging.DEBUG, msg=preprocessed_image)
+    return preprocessed_image
+
+
+def inference(task, model, input_image):
+    """
+    Process inference for bottle labels
+    Args:
+        - task
+        - use_case (product)
+        - image
+    Returns:
+        - String including label and confidence of the model
+    """
+
+    prediction = inferencer_arr[task][model][model].run(
+        [inferencer_arr[task][model]["output_name"]],
+        {inferencer_arr[task][model]["input_name"]: input_image},
+    )
+    prediction = prediction[0].squeeze()
+    logging.log(level=logging.INFO, msg=f"Softmaxed prediction {softmax(prediction)}")
+    if config["tasks"][task]["models"][model]["type"] == "classification":
+        result = f"{config['tasks'][task]['models'][model]['class_names'][np.argmax(prediction)]} avec une confiance de {str(round(softmax(prediction)[np.argmax(prediction)]*100))} %"
+    elif task == "task2":
+        label, score = is_anomalous_classification(prediction, corck_screwing_metadata)
+        result = f"{label} avec une confiance de {str(round(score*100))} %"
+    else:
+        result = "Houston, we've got a problem"
+    logging.log(level=logging.DEBUG, msg=result)
+    return result

liccheck.ini

@@ -0,0 +1,43 @@
+# Authorized and unauthorized licenses in LOWER CASE
+[Licenses]
+authorized_licenses:
+    bsd
+    new bsd
+    bsd license
+    new bsd license
+    simplified bsd
+    apache
+    apache 2.0
+    apache software license
+    gnu lgpl
+    lgpl with exceptions or zpl
+    isc license
+    isc license (iscl)
+    mit
+    mit license
+    python software foundation
+    python software foundation license
+    zpl 2.1
+    historical permission notice and disclaimer
+
+unauthorized_licenses:
+    gpl v3
+
+[Authorized Packages]
+aiohttp: 3.8.1
+aiosignal: 1.2.0
+async-timeout: 4.0.2
+bcrypt: 3.2.0
+certifi: 2021.10.8
+frozenlist: 1.3.0
+gradio: 2.9.1
+monotonic: 1.6
+multidict: 6.0.2
+openvino: 2022.1.0
+paramiko: 2.10.3
+Pillow: 9.1.0
+PyNaCl: 1.5.0
+python-multipart: 0.0.5
+requests: 2.27.1
+yarl: 1.7.2
+

pyproject.toml

@@ -0,0 +1,66 @@
+[tool.poetry]
+name = "Model Demo with Gradio"
+version = "0.0.1"
+description = ""
+authors = ["bastien@automi.ai"]
+
+[tool.poetry.dependencies]
+python = "^3.8, <3.11"
+gradio = "^3.23.0"
+logging = "^0.4.9.6"
+numpy = "^1.24.2"
+onnxruntime = "^1.14.1"
+requests = "^2.28.2"
+torchvision = "^0.15.1"
+
+[tool.poetry.group.dev.dependencies]
+black = "^23.1.0"
+liccheck = "^0.8.3"
+pylint = "^2.17.0"
+pytest = "^7.2.2"
+
+[build-system]
+requires = ["poetry-core>=1.0.0"]
+build-backend = "poetry.core.masonry.api"
+
+#[tool.liccheck]
+#authorized_licenses = [
+#    "bsd",
+#    "new bsd",
+#    "bsd license",
+#    "new bsd license",
+#    "simplified bsd",
+#    "apache",
+#    "apache 2.0",
+#    "apache software license",
+#    "gnu lgpl",
+#    "lgpl with exceptions or zpl",
+#    "isc license",
+#    "isc license (iscl)",
+#    "mit",
+#    "mit license",
+#    "python software foundation",
+#    "python software foundation license",
+#    "zpl 2.1",
+#    "historical permission notice and disclaimer",
+#    "gpl v3"
+#]
+#unauthorized_licenses = [""]
+#authorized_packages = [
+#    "aiohttp: *",
+#    "aiosignal: *",
+#    "async-timeout: *",
+#    "bcrypt: *",
+#    "certifi: *",
+#    "frozenlist: *",
+#    "gradio: *",
+#    "monotonic: *",
+#    "multidict: *",
+#    "openvino: *",
+#    "paramiko: *",
+#    "Pillow: *",
+#    "PyNaCl: *",
+#    "python-multipart: *",
+#    "requests: *",
+#    "yarl: *"
+#]