Submission Guidelines
1. Overview
Your submission will consist of three of four files that need to be
uploaded to MOLE. These are
• system.py (mandatory)
• model.json.gz (mandatory)
• requirements.txt (optional)
• report.md (mandatory)
No other files will be accepted. The files listed above are detailed
below.
2. The files to submit
system.py
This file stores your python code. All the python code that you have
written must be in this file. It must work correctly when called from
the train.py and evaluate.py provided.
model.json.gz
This is the file that is generated when you call train.py in conjunction
with your own system.py. As stated in the assignment instructions it
must not exceed the size limit of 3 MB.
requirements.txt
If you have developed your system outside of CoCalc or if you have
installed any modules using pip then I will need to know which modules
and module versions you have used. The easiest way to tell me this is
to include a requirements file. This is generated by typing ‘pip freeze’
in your python environment and saving the output into a file called
requirements.txt. (If you have tested your code in CoCalc then there
is no need to include this file.)
1
report.md
Your report must be a concisely written text file called ‘report.md’
strictly following the format below. Cut and paste the text below and
then replace the text in square brackets with your answers. The word
limits are strict.
# OCR assignment report
## Feature Extraction (Max 200 Words)
[Describe and justify the design of your feature extraction. Note,
feature extraction includes all the steps that you have taken to
produce the 10 dimensional feature vector starting from the initial
image of the page.]
## Classifier (Max 200 Words)
[Describe and justify the design of your classifier and any
associated classifier training stage.]
## Error Correction (Max 200 Words)
[Describe and justify the design of any post classification error
correction that you may have attempted.]
## Performance
The percentage errors (to 1 decimal place) for the development data are
as follows:
– Page 1: [Insert percentage here, e.g. 98.1%]
– Page 2: [Insert percentage here, e.g. 98.0%]
– Page 3: [Insert percentage here, e.g. 83.3%]
– Page 4: [Insert percentage here, e.g. 58.1%]
– Page 5: [Insert percentage here, e.g. 38.7%]
– Page 6: [Insert percentage here, e.g. 28.9%]
## Other information (Optional, Max 100 words)
[Optional: highlight any significant aspects of your system that are
NOT covered in the sections above]
2
Submission Guidelines
1. Overview
2. The files to submit
system.py
model.json.gz
requirements.txt
report.md