homework 6

designing recurrent networks

evaluating classifier performance

Named entity recognition is the process of identifying and classifying entities in unstructured text. Review the homework in this pdf file. We will be borrowing the codebase from DeepLearning.AI. Remember that reading resources can be found in the syllabus.


data and starter kit

  • The code template for homework 6 is located here. Please fill out where you see # <YOUR-CODE-HERE>. There will be None values for those lines that require attention.

  • The dataset has several different sizes for prototyping, evaluation, and training, and can be found here. If you’re on a GCP VM, you can download with the bash command wget. If you wish to download inside of a notebook, you can type in a shell commnad with a !.

  • You will need the data and loading scripts, which are found in this zipfile. The dataset has several different sizes for prototyping, evaluation, and training. You can download it with the wget command. Inside of this zip file, there is a script called load_data. To make the Python code work, you will put load_data.py where you can import it. Line 12 in assignment6.py appears as follows:

    from load_data import load_data

  • The python code you will need to modify and turn in is located here. There are several functions that you will be modifying. They will be of the form

    def some_function(argument):
  '''Description of arguments and return values
  '''
  ### START CODE HERE ###

  return_values = "This is where you will add or edit the code"

  ### END CODE HERE ###
  return return_values

    You will need to edit between START CODE HERE and END CODE HERE.

  • Document templates can be either Overleaf TeX File or DOCX File. When you’ve compiled/finishe d writing, download the PDF from Overleaf/Google and upload it to the submission link.

  • You will be turning in Python files, but feel free to develop in notebooks. To set one up, you can use, there are several options:



submission instructions

Submit your work to Gradescope. You will need to submit the files:

  • assignment6.py - your solutions to the questions
  • assignment6.h5 - parameters to your model