train_dan_for_long.sh

#!/bin/bash

# Run this command and analyse the tensorboard logs.
python train.py main \
    data/imdb_sentiment_train_5k.jsonl \
    data/imdb_sentiment_dev.jsonl \
    --seq2vec-choice dan \
    --embedding-dim 50 \
    --num-layers 4 \
    --num-epochs 50 \
    --suffix-name _dan_5k_with_emb_for_50k \
    --pretrained-embedding-file data/glove.6B.50d.txt

# tensorboard --logdir serialization_dirs/main_dan_5k_with_emb_for_50k

# Note: If you run it multiple times, make sure to clean-up tensorboard directory
# Or else, it will have multiple plots in same chart.