ParlAI examples

This directory contains a few particular examples of basic loops.

• base_train.py: very simple example shows the outline of a training/validation loop using the default Agent parent class
• display_data.py: uses agent.repeat_label to display data from a particular task provided on the command-line
• display_model.py: shows the predictions of a provided model on a particular task provided on the command-line
• eval_model.py: uses the named agent to compute evaluation metrics data for a particular task provided on the command-line
• build_dict.py: build a dictionary from a particular task provided on the command-line using core.dict.DictionaryAgent
• memnn_luatorch_cpu: shows a few examples of training an end-to-end memory network on a few datasets
• drqa: shows how to train the attentive LSTM DrQA model of Chen et al. on SQuAD.

Running These Examples

Most of them can be run simply by typing python {example}.py -t {task_name}. Here are some examples:

Display 10 random examples from task 1 of the "1k training examples" bAbI task:

python display_data.py -t babi:task1k:1

Run a train/valid loop with the basic agent (which prints what it receives and then says hello to the teacher, rather than learning anything) on the babi task:

python base_train.py -t babi:task1k:1

Displays 100 random examples from multi-tasking on the bAbI task and the SQuAD dataset at the same time:

python display_data.py -t babi:task1k:1,squad -n 100

Evaluate an IR baseline model on the validation set of the Movies Subreddit dataset:

python eval_model.py -m ir_baseline -t "#moviedd-reddit" -dt valid

Display the predictions of that same IR baseline model:

python display_model.py -m ir_baseline -t "#moviedd-reddit" -dt valid

Build a dictionary on a bAbI "1k training examples" task 1 and save it to /tmp/dict.tsv

python build_dict.py -t babi:task1k:1 --dict-savepath /tmp/dict.tsv

Train a simple cpu-based memory network on the "10k training examples" bAbI task 1 with 8 threads (python processes) using Hogwild (requires zmq and Lua Torch):

python memnn_luatorch_cpu/full_task_train.py -t babi:task10k:1 -nt 8

Trains an attentive LSTM model on the SQuAD dataset with a batch size of 32 examples (pytorch and regex):

python drqa/train.py -t squad -bs 32