How to compute FID score for different checkpoints

I have a hyper spectral image of 170 bands. I have used auto-encoder to reconstruct the image. Now I want to plot the FID score for 100 epochs like we can do for MSE plot.

like we do for MSE polt:

model.compile(optimizer=‘adam’, loss=‘mean_absolute_error’, metrics=[‘accuracy’])

history = model.fit(img, img,
epochs=100, batch_size=1, verbose=1,
validation_split=0.33, shuffle=True)

list all data in history

print(history.history.keys())

summarize history for accuracy

plt.plot(history.history[‘accuracy’])
plt.plot(history.history[‘val_accuracy’])
plt.title(‘model accuracy’)
plt.ylabel(‘accuracy’)
plt.xlabel(‘epoch’)
plt.legend([‘train’, ‘test’], loc=‘upper left’)
plt.show()

Is there a simple way to do that? If so can anyone assist me with a demo code is possible…

Thanks in advance…

@shuvro ,

I’d be glad to look into this for you. Although I can’t give provide specific examples I will provide online references on FID scores and plotting.

  • In terms of how to setup an FID evaluation pipeline, please check out this wandbreport for potential tips
  • In regards to plotting FID scores, I found this colab reference online that may include examples of what you are looking for.

If after you are successful in producing your plots and require assistance when logging them using wandb, please do reach out again.

Regards,

Mohammad

Hi @shuvro ,

Since we have not heard back from you we are going to close this request. If you would like to re-open the conversation, please let us know!

Regards,

Mohammad