#🔒 Model misclassifying

6 messages · Page 1 of 1 (latest)

hollow scroll
#

I have attached the model's specification

The problem i am facing is that even though the model was able to achieve 94% accuracy on thesr 9 classes it just misclassifying everything i doodle on screen as animal_migration.

You can look in the confusion_analysis.txt that it is struggling with this particular class.

I dont understand how can i fix this. I have already wasted 4 hours on this and have alredy used-up the free credits on google colab so no more gpu acceleration.

Let me know if you need anything else.

📎 class_names_9.txt 📎 confusion_analysis.txt 📎 model_info_345.txt prediction_samples_345.png training_history_345.png
kindred thunderBOT
kindred thunderBOT
#

@hollow scroll

Python help channel opened

Remember to:

  • Ask your Python question, not if you can ask or if there's an expert who can help.
  • Show a code sample as text (rather than a screenshot) and the error message, if you've got one.
  • Explain what you expect to happen and what actually happens.

:warning: Do not pip install anything that isn't related to your question, especially if asked to over DMs.

Closes after a period of inactivity, or when you send !close.
hollow scroll
#

The model sure was behaving good when i was using it predict on colab itself but i saved the model in 2 formats the common .keras and for better performance on mobile in .tflite format, i am assuming that when i was using the model in colab it must be existing .keras format (unsaved) or something similiar because i had to use .from_keras_() to save it as .tflite

And i am using .tflite on my phone and drawing using pygame and then CORRECTLY preprocessing the image then passing it to model to predict and i cant use the .keras version since the model was saved using newer version of keras on colab but i cant download that version on pydroid3

kindred thunderBOT
#

@hollow scroll

Python help channel closed for inactivity

This help channel has been closed. Feel free to create a new post in #1035199133436354600. To maximize your chances of getting a response, check out this guide on asking good questions.