evaluate_posthoc_zeroout.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

import torch
import torch.nn as nn
import torch.optim as optim
import os, sys
import numpy as np
import pandas as pd
import argparse
import json
import logging
import gc
import datetime
import sys

# torch.cuda.empty_cache()
# # torch.cuda.memory_summary(device=None, abbreviated=False)
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')





date_time = str(datetime.date.today()) + "_" + ":".join(str(datetime.datetime.now()).split()[1].split(":")[:2])

parser = argparse.ArgumentParser()

parser.add_argument(
    "--dataset", 
    type = str, 
    help = "select dataset / task", 
    default = "agnews",
    # choices = ["multirc", "agnews", "sst", "evinf",
)

parser.add_argument(
    '--use_topk', 
    help='for using the component by GChrys and Aletras 2021', 
    action='store_true',
    default=True,
)

parser.add_argument(
    "--data_dir", 
    type = str, 
    help = "directory of saved processed data", 
    default = "datasets/"
)

parser.add_argument(
    "--model_dir",   
    type = str, 
    help = "directory to save models", 
    # default = "full_text_models/",
    default="trained_models/"
)


parser.add_argument(
    "--evaluation_dir",   
    type = str, 
    help = "directory to save faithfulness results", 
    default = "posthoc_results/"
)

parser.add_argument(
    "--extracted_rationale_dir",   
    type = str, 
    help = "directory to save extracted_rationales", 
    default = "extracted_rationales/"
)


parser.add_argument(
    "--normalise", 
    type = int, 
    help = "decide how to normalise importance scores", 
    default = 1, 
)

parser.add_argument(
    "--thresholder", 
    type = str, 
    help = "thresholder for extracting rationales", 
    default = "topk",
    choices = ["contigious", "topk"]
)

parser.add_argument(
    "--inherently_faithful", 
    type = str, 
    help = "select dataset / task", 
    default = None, 
    choices = [None, "kuma", "rl"]
)




parser.add_argument(
    '--interpolation_analysis', 
    help='for using the component by GChrys and Aletras 2021', 
    action='store_true',
    default=False,
)

user_args = vars(parser.parse_args())
user_args["importance_metric"] = None

log_dir = "experiment_logs/evaluate_" + user_args["dataset"] + "_" +  date_time + "/"
config_dir = "experiment_config/evaluate_" + user_args["dataset"] + "_" +  date_time + "/"


os.makedirs(log_dir, exist_ok = True)
os.makedirs(config_dir, exist_ok = True)


import config.cfg
config.cfg.config_directory = config_dir
logging.basicConfig(
                    filename= log_dir + "/out.log", 
                    format='%(asctime)s %(levelname)-8s %(message)s',
                    level=logging.INFO,
                    datefmt='%Y-%m-%d %H:%M:%S'
                  )

device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
CUDA_LAUNCH_BLOCKING=1

logging.info("Running on cuda ? {}".format(torch.cuda.is_available()))

torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
torch.cuda.empty_cache()
CUDA_LAUNCH_BLOCKING=1

from src.common_code.initialiser import initial_preparations

# creating unique config from stage_config.json file and model_config.json file
args = initial_preparations(user_args, stage = "evaluate")

logging.info("config  : \n ----------------------")
[logging.info(k + " : " + str(v)) for k,v in args.items()]
logging.info("\n ----------------------")


import glob
from src.data_functions.dataholder import BERT_HOLDER
from src.evaluation import evaluation_pipeline

model_path = os.path.join(
            os.getcwd(),
            args["model_dir"],
            ""
        )

print(model_path)

data = BERT_HOLDER(
    args["data_dir"], 
    stage = "eval",
    b_size = 2,
    #b_size = args["batch_size"], # TO FIX CUDA OUT OF MEMORY, MAY NOT WORK
)

evaluator = evaluation_pipeline.evaluate_zeroout(
    model_path = args["model_dir"], 
    output_dims = data.nu_of_labels,
    use_topk = args["use_topk"],
    normalise = 1,
)

# will generate
logging.info("*********conducting in-domain flip experiments")
print('"*********conducting flip experiments on in-domain"')
evaluator.faithfulness_experiments_(data)
print('"********* DONE flip experiments on in-domain"')


evaluator = evaluation_pipeline.evaluate_zeroout(
    model_path = args["model_dir"], 
    output_dims = data.nu_of_labels,
    use_topk = args["use_topk"],
    normalise = 5,
)

# will generate
logging.info("*********conducting in-domain flip experiments")
print('"*********conducting flip experiments on in-domain"')
evaluator.faithfulness_experiments_(data)
print('"********* DONE flip experiments on in-domain"')


del data
del evaluator
gc.collect()



torch.cuda.empty_cache()