task_setup.yaml

# Memory used by Spark driver (recommend > 6g)
spark_driver_memory: '6g'

# Path for data directory. Note that all other paths in this setup 
# file will be relative to this one (e.g. if data_dir='some/path' 
# and emb_file='to/emb.parquet', the embeddings will be loaded from 
# 'some/path/to/emb.parquet')    
data_dir: 'data'

# Relative path for embedding file
emb_file: 'embeddings/train_emb_256_dataperf.parquet'

# Paths to training and test files for each tasks. Task names must match names
# in eval_tasks above. See example below:
# task_name: ['relative/path/to/training_set.csv', 'relative/path/to/test_set.csv']
Cupcake: ['train_sets/random_500.csv', 'test_sets/alpha_test_set_Cupcake_256.parquet']
Hawk: ['train_sets/random_500.csv', 'test_sets/alpha_test_set_Hawk_256.parquet']
Sushi: ['train_sets/random_500.csv', 'test_sets/alpha_test_set_Sushi_256.parquet']

# Relative path for results file
results_dir: 'results'

# Parameters specific to alpha below (likely not useful to modify)

# Embedding dimensionality
dim: 256

# List the names of tasks to evaluate
eval_tasks: ['Cupcake','Hawk','Sushi']

# Path for data directory in docker image.
# DO NOT MODIFY
docker_data_dir: 'data'