Kopitiam

[Updated Jul 2023]: Here is the updated notebook on how to fine-tune a Seq2Seq model with Huggingface Transformers, (not presented publicly yet), https://www.kaggle.com/code/alvations/how-to-fine-tune-an-opus-mt-model/

This was the repository that host the notebook for the Seq2Seq with PyTorch session in Data Science SG meetup in January 2018. The original accompanying slides can be found on https://goo.gl/Lu6CxB and the 2018 original notebook is on https://github.com/alvations/kopitiam/blob/master/Kopitiam.ipynb

Problem

Ordering coffee in Singapore hawker centre and food court could result in some surprises:

The dataset used in this exercise is hosted on https://www.kaggle.com/alvations/sg-kopi

Name		Name	Last commit message	Last commit date
Latest commit History 30 Commits
old-models		old-models
How to Fine-Tune an OPUS Machine Translation Model_.ipynb		How to Fine-Tune an OPUS Machine Translation Model_.ipynb
Kopitiam - PyTorch RNN Seq2Seq.ipynb		Kopitiam - PyTorch RNN Seq2Seq.ipynb
Kopitiam Emptied.ipynb		Kopitiam Emptied.ipynb
Kopitiam mit Attention.ipynb		Kopitiam mit Attention.ipynb
Kopitiam mit Teacher Forcing.ipynb		Kopitiam mit Teacher Forcing.ipynb
Kopitiam.ipynb		Kopitiam.ipynb
LICENSE		LICENSE
README.md		README.md
decoder_attention_100_100000_0.5.pkl		decoder_attention_100_100000_0.5.pkl
decoder_vanilla_100_100000.pkl		decoder_vanilla_100_100000.pkl
decoder_vanilla_100_100000_0.5.pkl		decoder_vanilla_100_100000_0.5.pkl
encoder_attention_100_100000_0.5.pkl		encoder_attention_100_100000_0.5.pkl
encoder_vanilla_100_100000.pkl		encoder_vanilla_100_100000.pkl
encoder_vanilla_100_100000_0.5.pkl		encoder_vanilla_100_100000_0.5.pkl
english_vocab.Dictionary.pkl		english_vocab.Dictionary.pkl
kopitiam.tsv		kopitiam.tsv
singlish_vocab.Dictionary.pkl		singlish_vocab.Dictionary.pkl