The plot to show the training and validation processes

[quangchien99]

Can you show me the code the visualize the result after training like the loss, val_loss or accuracy?

[ptran1203]

Hi, currently the code for loss curve visualization is unavailable. The metric for this problem is mAP but I did not implement it yet.

For loss curve visualization you can check the following:

H = model.fit(...)
loss_values = H.history["loss"]
plt.plot(epochs, loss_values, label='Training Loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.legend()

[quangchien99]

So how do i get the accuracy value plot?

[ptran1203]

As I mentioned above, object detection problem is usually evaluated by mAP metric not accuracy, I will try to implement it in keras

[quangchien99]

Thank u, however, i run your code in google colab, when training - the loss is so high - in epochs 1 is 1.7 , and epochs 30 is 0,7.
Is it ok?

[ptran1203]

If the loss is reducing when training, it's ok. You can try to train more epochs or increase the learning rate, eg. 1e-3

[quangchien99]

Oke thank u so muchhhh. You're amazing