You Could Check what I wrote details about this project HERE
2️⃣ This project was motivated by Amenity Detection led by Airbnb Data Science Team as I completed somewhat it as clone project.
3️⃣ This performs amenity detection using Tensorflow Object Detection API and uses CenterNet Model when Training and was practiced in Google Colab Labatoray, Clould Based Service.
4️⃣ Moreover, 🎇 means I could not be able to upload here as their file size is toooo big but you would be able to download them in jupyter notebook where there are links.
Directory
├── 🎇 Original_open_image_dataset (from Downloading)
│
├── 🎇 filtered_train_images
│ ├── aggregated
│ ├── class_names
│ ├── .....
│
├── 🎇 filtered_validation_images
│ ├── aggregated
│ ├── class_names
│ ├── .....
│
├── 🎇 filtered_test_images
│ ├── aggregated
│ ├── class_names
│ ├── .....
│
├── annotations
│ ├── 🎇 train-annotations-bbox.csv
│ ├── validation-annotations-bbox.csv
│ ├── 🎇test-annotations-bbox.csv
│
├── 🎇saved_model
│ ├── checkpoint
│ ├── train
│ ├── train_log_file (Used for Tensorboard)
│ ├── eval
│ ├── eval_log_file (Used for Tensorboard)
│
├── 🎇 train_tfrecords
│ ├── oid_30_class_train.record-0000x-of-00005
│
├── 🎇 valid_tfrecords
│ ├── oid_30_class_val.record-0000x-of-00005
│
├── py
│ ├── open_images_dataset_30_class_parsing.py (Parse 30 classes out of entire classes and save images)
│ ├── create_oid_v4_tf_record.py (Generate TFRecord files for Train & Validation)
│
├── centernet_hourglass104_512x512_amenity_30class.config (Model Configuration File)
│
├── oid_v4_label_map_amenity_30_class.pbtxt (label id-name mapping files)
│
├── class-descriptions-boxable.csv (Consists of all 'ids' for classes and 'names' {human-readable string format} of Open Image Dataset)
│
├── notebooks
│ ├── Amenity_Detection.ipynb ( Download arguments & Train)
│ ├── Evaluation.ipynb (Validation & Tensorboard)
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent