We will find precision, recall and F1 results using CNN-based deep learning model with leaf images dataset for identification of Bangladeshi medicinal plants. We will show these with a table. We will also apply this process for VGG and DenseNet. And for these we will show the Confusion matrix, training/validation loss, and training/validation accuracy values. Note: I have the data set, but I have no idea how to use it and how to load it. I would appreciate it if you could inform me about it.