All the top 10 vectors closest to 5 different words are in Report.pdf
Run the jupyter file q1.ipynb top to bottom.
Print functions are left in understanding the code and its structure.
Embeddings, pre-trained model and graph images for q1 saved: https://drive.google.com/drive/folders/1D7hJuZ1_D6s4t8gjbubbGI03YFin5PC8
The training time is quite low so it is not really practical to import pre-trained model instead the whole thing.
Run the jupyter file q2.ipynb top to bottom.
Embeddings, pre-trained model and graph images for q2 saved: https://drive.google.com/drive/folders/1AOnJC7WR35HOzVfp5KrtA7J5bFF5LrTO
The embeddings for CBOW are also attached in the folder embeddings.
To check the top10 for any word we can make use of the script z.py.
Restoring of pre-trained model is all shown in z.py and can be used.
E.g:
# top10 for `camera`, run:
python3 z.py camera
# top10 for `awesome`, run:
python3 z.py awesome
# similarly for others