A Deep Learning Model to Identify Wild Orchids of Taiwan
Dataset can be downloaded from:
https://drive.google.com/drive/folders/1tfILR7tSIv_5PV8HaUa7yx7F8jUPkYv8?usp=sharing
The dataset file including train(Including 935 pictures), test(Including 166 pictures), and
label_versus_Orchid_name.xlsx, the label and the corresponding Orchid name.
train.py & train.ipynb: The python code for training.
The training result folder includes the training result of the project.
The csv file includes image filename, true label and predicted label.
The accuracy and loss are shown below:
Epoch | Train_accuracy | Validation_accuracy | Train_loss | Validation_loss |
---|---|---|---|---|
001 | 0.43061 | 0.75565 | 2.00532 | 0.84623 |
002 | 0.75104 | 0.85119 | 0.93089 | 0.37949 |
003 | 0.88098 | 0.90476 | 0.48639 | 0.31409 |
004 | 0.92256 | 0.95238 | 0.31368 | 0.15387 |
005 | 0.95478 | 0.93452 | 0.20603 | 0.19232 |
006 | 0.96206 | 0.94048 | 0.15989 | 0.29069 |
007 | 0.96544 | 0.94048 | 0.13352 | 0.23897 |
report.pdf: The report for this project.